reduce_ptx.h

Go to the documentation of this file.

static const char* reduce_ptx = 
"       .version 1.4\n"
"       .target sm_10, map_f64_to_f32\n"
"       // compiled with /usr/local/cuda/open64/lib//be\n"
"       // nvopencc 3.1 built on 2010-06-07\n"
"\n"
"       //-----------------------------------------------------------\n"
"       // Compiling /tmp/tmpxft_00007884_00000000-7_reduce.cpp3.i (/tmp/ccBI#.LV1fMO)\n"
"       //-----------------------------------------------------------\n"
"\n"
"       //-----------------------------------------------------------\n"
"       // Options:\n"
"       //-----------------------------------------------------------\n"
"       //  Target:ptx, ISA:sm_10, Endian:little, Pointer Size:64\n"
"       //  -O3 (Optimization level)\n"
"       //  -g0 (Debug level)\n"
"       //  -m2 (Report advisories)\n"
"       //-----------------------------------------------------------\n"
"\n"
"       .file   1       \"<command-line>\"\n"
"       .file   2       \"/tmp/tmpxft_00007884_00000000-6_reduce.cudafe2.gpu\"\n"
"       .file   3       \"reduce.cu\"\n"
"       .file   4       \"/usr/lib/gcc/x86_64-linux-gnu/4.4.3/include/stddef.h\"\n"
"       .file   5       \"/usr/local/cuda/bin/../include/crt/device_runtime.h\"\n"
"       .file   6       \"/usr/local/cuda/bin/../include/host_defines.h\"\n"
"       .file   7       \"/usr/local/cuda/bin/../include/builtin_types.h\"\n"
"       .file   8       \"/usr/local/cuda/bin/../include/device_types.h\"\n"
"       .file   9       \"/usr/local/cuda/bin/../include/driver_types.h\"\n"
"       .file   10      \"/usr/local/cuda/bin/../include/surface_types.h\"\n"
"       .file   11      \"/usr/local/cuda/bin/../include/texture_types.h\"\n"
"       .file   12      \"/usr/local/cuda/bin/../include/vector_types.h\"\n"
"       .file   13      \"/usr/local/cuda/bin/../include/device_launch_parameters.h\"\n"
"       .file   14      \"/usr/local/cuda/bin/../include/crt/storage_class.h\"\n"
"       .file   15      \"/usr/include/bits/types.h\"\n"
"       .file   16      \"/usr/include/time.h\"\n"
"       .file   17      \"/usr/local/cuda/bin/../include/common_functions.h\"\n"
"       .file   18      \"/usr/local/cuda/bin/../include/math_functions.h\"\n"
"       .file   19      \"/usr/local/cuda/bin/../include/math_constants.h\"\n"
"       .file   20      \"/usr/local/cuda/bin/../include/device_functions.h\"\n"
"       .file   21      \"/usr/local/cuda/bin/../include/sm_11_atomic_functions.h\"\n"
"       .file   22      \"/usr/local/cuda/bin/../include/sm_12_atomic_functions.h\"\n"
"       .file   23      \"/usr/local/cuda/bin/../include/sm_13_double_functions.h\"\n"
"       .file   24      \"/usr/local/cuda/bin/../include/sm_20_atomic_functions.h\"\n"
"       .file   25      \"/usr/local/cuda/bin/../include/sm_20_intrinsics.h\"\n"
"       .file   26      \"/usr/local/cuda/bin/../include/surface_functions.h\"\n"
"       .file   27      \"/usr/local/cuda/bin/../include/texture_fetch_functions.h\"\n"
"       .file   28      \"/usr/local/cuda/bin/../include/math_functions_dbl_ptx1.h\"\n"
"\n"
"       .extern .shared .align 4 .b8 __smem[];\n"
"       .tex .u64 tex_ref_1;\n"
"       .tex .u64 tex_ref_2;\n"
"\n"
"       .entry chamfer_and_reduce (\n"
"               .param .u64 __cudaparm_chamfer_and_reduce_g_idata_1,\n"
"               .param .u64 __cudaparm_chamfer_and_reduce_g_idata_2,\n"
"               .param .u64 __cudaparm_chamfer_and_reduce_g_odata,\n"
"               .param .u32 __cudaparm_chamfer_and_reduce_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<18>;\n"
"       .reg .f32 %f<23>;\n"
"       .reg .pred %p<9>;\n"
"       .loc    3       148     0\n"
"$LDWbegin_chamfer_and_reduce:\n"
"       .loc    3       105     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 512;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_chamfer_and_reduce_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_0_18178;\n"
"       add.u32         %r6, %r4, 256;\n"
"       ld.param.u32    %r5, [__cudaparm_chamfer_and_reduce_n];\n"
"       add.u32         %r7, %r5, 256;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 512;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       mul.wide.u32    %rd2, %r4, 4;\n"
"       cvt.s64.u32     %rd3, %r8;\n"
"       ld.param.u64    %rd4, [__cudaparm_chamfer_and_reduce_g_idata_1];\n"
"       add.u64         %rd5, %rd4, %rd2;\n"
"       mul.wide.u32    %rd6, %r8, 4;\n"
"       ld.param.u64    %rd7, [__cudaparm_chamfer_and_reduce_g_idata_2];\n"
"       add.u64         %rd8, %rd7, %rd2;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_0_15106:\n"
" //<loop> Loop body line 105, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       119     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       ld.global.f32   %f3, [%rd8+0];\n"
"       mad.f32         %f1, %f2, %f3, %f1;\n"
"       .loc    3       105     0\n"
"       ld.param.u32    %r5, [__cudaparm_chamfer_and_reduce_n];\n"
"       .loc    3       119     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_0_15362;\n"
" //<loop> Part of loop body line 105, head labeled $Lt_0_15106\n"
"       .loc    3       133     0\n"
"       ld.global.f32   %f4, [%rd5+1024];\n"
"       ld.global.f32   %f5, [%rd8+1024];\n"
"       mad.f32         %f1, %f4, %f5, %f1;\n"
"$Lt_0_15362:\n"
" //<loop> Part of loop body line 105, head labeled $Lt_0_15106\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd8, %rd8, %rd6;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_0_15106;\n"
"       bra.uni         $Lt_0_14594;\n"
"$Lt_0_18178:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_0_14594:\n"
"       .loc    3       139     0\n"
"       mov.f32         %f6, %f1;\n"
"       mov.f32         %f7, %f6;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd9, __smem;\n"
"       cvt.u64.u32     %rd10, %r3;\n"
"       mul.wide.u32    %rd11, %r3, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.volatile.shared.f32  [%rd12+0], %f6;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 127;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_0_16130;\n"
"       .loc    3       76      0\n"
"       ld.volatile.shared.f32  %f8, [%rd12+512];\n"
"       add.f32         %f7, %f8, %f6;\n"
"       st.volatile.shared.f32  [%rd12+0], %f7;\n"
"$Lt_0_16130:\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 63;\n"
"       setp.gt.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_0_16642;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f9, [%rd12+256];\n"
"       add.f32         %f7, %f9, %f7;\n"
"       st.volatile.shared.f32  [%rd12+0], %f7;\n"
"$Lt_0_16642:\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 31;\n"
"       setp.gt.u32     %p6, %r3, %r11;\n"
"       @%p6 bra        $Lt_0_17154;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f10, [%rd12+128];\n"
"       add.f32         %f11, %f10, %f7;\n"
"       st.volatile.shared.f32  [%rd12+0], %f11;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f12, [%rd12+64];\n"
"       add.f32         %f13, %f12, %f11;\n"
"       st.volatile.shared.f32  [%rd12+0], %f13;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f14, [%rd12+32];\n"
"       add.f32         %f15, %f14, %f13;\n"
"       st.volatile.shared.f32  [%rd12+0], %f15;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f16, [%rd12+16];\n"
"       add.f32         %f17, %f16, %f15;\n"
"       st.volatile.shared.f32  [%rd12+0], %f17;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f18, [%rd12+8];\n"
"       add.f32         %f19, %f18, %f17;\n"
"       st.volatile.shared.f32  [%rd12+0], %f19;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f20, [%rd12+4];\n"
"       add.f32         %f7, %f20, %f19;\n"
"       st.volatile.shared.f32  [%rd12+0], %f7;\n"
"$Lt_0_17154:\n"
"       .loc    3       139     0\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p7, %r3, %r12;\n"
"       @%p7 bra        $Lt_0_17666;\n"
"       .loc    3       143     0\n"
"       ld.shared.f32   %f21, [__smem+0];\n"
"       ld.param.u64    %rd13, [__cudaparm_chamfer_and_reduce_g_odata];\n"
"       cvt.u64.u32     %rd14, %r1;\n"
"       mul.wide.u32    %rd15, %r1, 4;\n"
"       add.u64         %rd16, %rd13, %rd15;\n"
"       st.global.f32   [%rd16+0], %f21;\n"
"$Lt_0_17666:\n"
"       .loc    3       151     0\n"
"       exit;\n"
"$LDWend_chamfer_and_reduce:\n"
"       } // chamfer_and_reduce\n"
"\n"
"       .entry squared_chamfer_and_reduce (\n"
"               .param .u64 __cudaparm_squared_chamfer_and_reduce_g_idata_1,\n"
"               .param .u64 __cudaparm_squared_chamfer_and_reduce_g_idata_2,\n"
"               .param .u64 __cudaparm_squared_chamfer_and_reduce_g_odata,\n"
"               .param .u32 __cudaparm_squared_chamfer_and_reduce_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<18>;\n"
"       .reg .f32 %f<25>;\n"
"       .reg .pred %p<9>;\n"
"       .loc    3       154     0\n"
"$LDWbegin_squared_chamfer_and_reduce:\n"
"       .loc    3       105     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 512;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_squared_chamfer_and_reduce_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_1_18178;\n"
"       add.u32         %r6, %r4, 256;\n"
"       ld.param.u32    %r5, [__cudaparm_squared_chamfer_and_reduce_n];\n"
"       add.u32         %r7, %r5, 256;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 512;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       mul.wide.u32    %rd2, %r4, 4;\n"
"       cvt.s64.u32     %rd3, %r8;\n"
"       ld.param.u64    %rd4, [__cudaparm_squared_chamfer_and_reduce_g_idata_1];\n"
"       add.u64         %rd5, %rd4, %rd2;\n"
"       mul.wide.u32    %rd6, %r8, 4;\n"
"       ld.param.u64    %rd7, [__cudaparm_squared_chamfer_and_reduce_g_idata_2];\n"
"       add.u64         %rd8, %rd7, %rd2;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_1_15106:\n"
" //<loop> Loop body line 105, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       114     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       ld.global.f32   %f3, [%rd8+0];\n"
"       mul.f32         %f4, %f2, %f3;\n"
"       .loc    3       115     0\n"
"       mad.f32         %f1, %f4, %f4, %f1;\n"
"       .loc    3       105     0\n"
"       ld.param.u32    %r5, [__cudaparm_squared_chamfer_and_reduce_n];\n"
"       .loc    3       115     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_1_15362;\n"
" //<loop> Part of loop body line 105, head labeled $Lt_1_15106\n"
"       .loc    3       127     0\n"
"       ld.global.f32   %f5, [%rd5+1024];\n"
"       ld.global.f32   %f6, [%rd8+1024];\n"
"       mul.f32         %f7, %f5, %f6;\n"
"       .loc    3       129     0\n"
"       mad.f32         %f1, %f7, %f7, %f1;\n"
"$Lt_1_15362:\n"
" //<loop> Part of loop body line 105, head labeled $Lt_1_15106\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd8, %rd8, %rd6;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_1_15106;\n"
"       bra.uni         $Lt_1_14594;\n"
"$Lt_1_18178:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_1_14594:\n"
"       .loc    3       139     0\n"
"       mov.f32         %f8, %f1;\n"
"       mov.f32         %f9, %f8;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd9, __smem;\n"
"       cvt.u64.u32     %rd10, %r3;\n"
"       mul.wide.u32    %rd11, %r3, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.volatile.shared.f32  [%rd12+0], %f8;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 127;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_1_16130;\n"
"       .loc    3       76      0\n"
"       ld.volatile.shared.f32  %f10, [%rd12+512];\n"
"       add.f32         %f9, %f10, %f8;\n"
"       st.volatile.shared.f32  [%rd12+0], %f9;\n"
"$Lt_1_16130:\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 63;\n"
"       setp.gt.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_1_16642;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f11, [%rd12+256];\n"
"       add.f32         %f9, %f11, %f9;\n"
"       st.volatile.shared.f32  [%rd12+0], %f9;\n"
"$Lt_1_16642:\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 31;\n"
"       setp.gt.u32     %p6, %r3, %r11;\n"
"       @%p6 bra        $Lt_1_17154;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f12, [%rd12+128];\n"
"       add.f32         %f13, %f12, %f9;\n"
"       st.volatile.shared.f32  [%rd12+0], %f13;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f14, [%rd12+64];\n"
"       add.f32         %f15, %f14, %f13;\n"
"       st.volatile.shared.f32  [%rd12+0], %f15;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f16, [%rd12+32];\n"
"       add.f32         %f17, %f16, %f15;\n"
"       st.volatile.shared.f32  [%rd12+0], %f17;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f18, [%rd12+16];\n"
"       add.f32         %f19, %f18, %f17;\n"
"       st.volatile.shared.f32  [%rd12+0], %f19;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f20, [%rd12+8];\n"
"       add.f32         %f21, %f20, %f19;\n"
"       st.volatile.shared.f32  [%rd12+0], %f21;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f22, [%rd12+4];\n"
"       add.f32         %f9, %f22, %f21;\n"
"       st.volatile.shared.f32  [%rd12+0], %f9;\n"
"$Lt_1_17154:\n"
"       .loc    3       139     0\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p7, %r3, %r12;\n"
"       @%p7 bra        $Lt_1_17666;\n"
"       .loc    3       143     0\n"
"       ld.shared.f32   %f23, [__smem+0];\n"
"       ld.param.u64    %rd13, [__cudaparm_squared_chamfer_and_reduce_g_odata];\n"
"       cvt.u64.u32     %rd14, %r1;\n"
"       mul.wide.u32    %rd15, %r1, 4;\n"
"       add.u64         %rd16, %rd13, %rd15;\n"
"       st.global.f32   [%rd16+0], %f23;\n"
"$Lt_1_17666:\n"
"       .loc    3       157     0\n"
"       exit;\n"
"$LDWend_squared_chamfer_and_reduce:\n"
"       } // squared_chamfer_and_reduce\n"
"\n"
"       .entry reduce_float_1_true (\n"
"               .param .u64 __cudaparm_reduce_float_1_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_1_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_1_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<10>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<7>;\n"
"       .reg .pred %p<5>;\n"
"       .loc    3       372     0\n"
"$LDWbegin_reduce_float_1_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 2;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_1_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_2_16642;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 2;\n"
"       cvt.s64.u32     %rd1, %r7;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_1_true_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r7, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_2_15618:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f4, [%rd5+4];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_1_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_2_15618;\n"
"       bra.uni         $Lt_2_15106;\n"
"$Lt_2_16642:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_2_15106:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       .loc    3       195     0\n"
"       mov.u32         %r8, 0;\n"
"       setp.ne.u32     %p3, %r3, %r8;\n"
"       @%p3 bra        $Lt_2_16130;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f5, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_1_true_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f5;\n"
"$Lt_2_16130:\n"
"       .loc    3       375     0\n"
"       exit;\n"
"$LDWend_reduce_float_1_true:\n"
"       } // reduce_float_1_true\n"
"\n"
"       .entry reduce_float_2_true (\n"
"               .param .u64 __cudaparm_reduce_float_2_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_2_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_2_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<11>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<9>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       377     0\n"
"$LDWbegin_reduce_float_2_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 4;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_2_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_3_16898;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 4;\n"
"       cvt.s64.u32     %rd1, %r7;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_2_true_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r7, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_3_15362:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f4, [%rd5+8];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_2_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_3_15362;\n"
"       bra.uni         $Lt_3_14850;\n"
"$Lt_3_16898:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_3_14850:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r8, 31;\n"
"       setp.gt.u32     %p3, %r3, %r8;\n"
"       @%p3 bra        $Lt_3_15874;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f5, [%rd10+4];\n"
"       add.f32         %f6, %f5, %f1;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"$Lt_3_15874:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r9, 0;\n"
"       setp.ne.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_3_16386;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f7, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_2_true_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f7;\n"
"$Lt_3_16386:\n"
"       .loc    3       380     0\n"
"       exit;\n"
"$LDWend_reduce_float_2_true:\n"
"       } // reduce_float_2_true\n"
"\n"
"       .entry reduce_float_4_true (\n"
"               .param .u64 __cudaparm_reduce_float_4_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_4_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_4_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<11>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<11>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       382     0\n"
"$LDWbegin_reduce_float_4_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 8;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_4_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_4_16642;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 8;\n"
"       cvt.s64.u32     %rd1, %r7;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_4_true_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r7, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_4_15106:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f4, [%rd5+16];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_4_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_4_15106;\n"
"       bra.uni         $Lt_4_14594;\n"
"$Lt_4_16642:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_4_14594:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r8, 31;\n"
"       setp.gt.u32     %p3, %r3, %r8;\n"
"       @%p3 bra        $Lt_4_15618;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f5, [%rd10+8];\n"
"       add.f32         %f6, %f5, %f1;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f7, [%rd10+4];\n"
"       add.f32         %f8, %f7, %f6;\n"
"       st.volatile.shared.f32  [%rd10+0], %f8;\n"
"$Lt_4_15618:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r9, 0;\n"
"       setp.ne.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_4_16130;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f9, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_4_true_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f9;\n"
"$Lt_4_16130:\n"
"       .loc    3       385     0\n"
"       exit;\n"
"$LDWend_reduce_float_4_true:\n"
"       } // reduce_float_4_true\n"
"\n"
"       .entry reduce_float_8_true (\n"
"               .param .u64 __cudaparm_reduce_float_8_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_8_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_8_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<11>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<13>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       387     0\n"
"$LDWbegin_reduce_float_8_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 16;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_8_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_5_16386;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 16;\n"
"       cvt.s64.u32     %rd1, %r7;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_8_true_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r7, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_5_14850:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f4, [%rd5+32];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_8_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_5_14850;\n"
"       bra.uni         $Lt_5_14338;\n"
"$Lt_5_16386:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_5_14338:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r8, 31;\n"
"       setp.gt.u32     %p3, %r3, %r8;\n"
"       @%p3 bra        $Lt_5_15362;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f5, [%rd10+16];\n"
"       add.f32         %f6, %f5, %f1;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f7, [%rd10+8];\n"
"       add.f32         %f8, %f7, %f6;\n"
"       st.volatile.shared.f32  [%rd10+0], %f8;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f9, [%rd10+4];\n"
"       add.f32         %f10, %f9, %f8;\n"
"       st.volatile.shared.f32  [%rd10+0], %f10;\n"
"$Lt_5_15362:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r9, 0;\n"
"       setp.ne.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_5_15874;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f11, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_8_true_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f11;\n"
"$Lt_5_15874:\n"
"       .loc    3       390     0\n"
"       exit;\n"
"$LDWend_reduce_float_8_true:\n"
"       } // reduce_float_8_true\n"
"\n"
"       .entry reduce_float_16_true (\n"
"               .param .u64 __cudaparm_reduce_float_16_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_16_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_16_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<11>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<15>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       392     0\n"
"$LDWbegin_reduce_float_16_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 32;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_16_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_6_16130;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 32;\n"
"       cvt.s64.u32     %rd1, %r7;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_16_true_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r7, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_6_14594:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f4, [%rd5+64];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_16_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_6_14594;\n"
"       bra.uni         $Lt_6_14082;\n"
"$Lt_6_16130:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_6_14082:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r8, 31;\n"
"       setp.gt.u32     %p3, %r3, %r8;\n"
"       @%p3 bra        $Lt_6_15106;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f5, [%rd10+32];\n"
"       add.f32         %f6, %f5, %f1;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f7, [%rd10+16];\n"
"       add.f32         %f8, %f7, %f6;\n"
"       st.volatile.shared.f32  [%rd10+0], %f8;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f9, [%rd10+8];\n"
"       add.f32         %f10, %f9, %f8;\n"
"       st.volatile.shared.f32  [%rd10+0], %f10;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f11, [%rd10+4];\n"
"       add.f32         %f12, %f11, %f10;\n"
"       st.volatile.shared.f32  [%rd10+0], %f12;\n"
"$Lt_6_15106:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r9, 0;\n"
"       setp.ne.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_6_15618;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f13, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_16_true_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f13;\n"
"$Lt_6_15618:\n"
"       .loc    3       395     0\n"
"       exit;\n"
"$LDWend_reduce_float_16_true:\n"
"       } // reduce_float_16_true\n"
"\n"
"       .entry reduce_float_32_true (\n"
"               .param .u64 __cudaparm_reduce_float_32_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_32_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_32_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<11>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<17>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       397     0\n"
"$LDWbegin_reduce_float_32_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 64;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_32_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_7_15874;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 64;\n"
"       cvt.s64.u32     %rd1, %r7;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_32_true_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r7, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_7_14338:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f4, [%rd5+128];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_32_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_7_14338;\n"
"       bra.uni         $Lt_7_13826;\n"
"$Lt_7_15874:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_7_13826:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r8, 31;\n"
"       setp.gt.u32     %p3, %r3, %r8;\n"
"       @%p3 bra        $Lt_7_14850;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f5, [%rd10+64];\n"
"       add.f32         %f6, %f5, %f1;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f7, [%rd10+32];\n"
"       add.f32         %f8, %f7, %f6;\n"
"       st.volatile.shared.f32  [%rd10+0], %f8;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f9, [%rd10+16];\n"
"       add.f32         %f10, %f9, %f8;\n"
"       st.volatile.shared.f32  [%rd10+0], %f10;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f11, [%rd10+8];\n"
"       add.f32         %f12, %f11, %f10;\n"
"       st.volatile.shared.f32  [%rd10+0], %f12;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f13, [%rd10+4];\n"
"       add.f32         %f14, %f13, %f12;\n"
"       st.volatile.shared.f32  [%rd10+0], %f14;\n"
"$Lt_7_14850:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r9, 0;\n"
"       setp.ne.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_7_15362;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f15, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_32_true_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f15;\n"
"$Lt_7_15362:\n"
"       .loc    3       400     0\n"
"       exit;\n"
"$LDWend_reduce_float_32_true:\n"
"       } // reduce_float_32_true\n"
"\n"
"       .entry reduce_float_64_true (\n"
"               .param .u64 __cudaparm_reduce_float_64_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_64_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_64_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<11>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<19>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       402     0\n"
"$LDWbegin_reduce_float_64_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 128;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_64_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_8_15618;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 128;\n"
"       cvt.s64.u32     %rd1, %r7;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_64_true_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r7, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_8_14082:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f4, [%rd5+256];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_64_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_8_14082;\n"
"       bra.uni         $Lt_8_13570;\n"
"$Lt_8_15618:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_8_13570:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r8, 31;\n"
"       setp.gt.u32     %p3, %r3, %r8;\n"
"       @%p3 bra        $Lt_8_14594;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f5, [%rd10+128];\n"
"       add.f32         %f6, %f5, %f1;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f7, [%rd10+64];\n"
"       add.f32         %f8, %f7, %f6;\n"
"       st.volatile.shared.f32  [%rd10+0], %f8;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f9, [%rd10+32];\n"
"       add.f32         %f10, %f9, %f8;\n"
"       st.volatile.shared.f32  [%rd10+0], %f10;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f11, [%rd10+16];\n"
"       add.f32         %f12, %f11, %f10;\n"
"       st.volatile.shared.f32  [%rd10+0], %f12;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f13, [%rd10+8];\n"
"       add.f32         %f14, %f13, %f12;\n"
"       st.volatile.shared.f32  [%rd10+0], %f14;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f15, [%rd10+4];\n"
"       add.f32         %f16, %f15, %f14;\n"
"       st.volatile.shared.f32  [%rd10+0], %f16;\n"
"$Lt_8_14594:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r9, 0;\n"
"       setp.ne.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_8_15106;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f17, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_64_true_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f17;\n"
"$Lt_8_15106:\n"
"       .loc    3       405     0\n"
"       exit;\n"
"$LDWend_reduce_float_64_true:\n"
"       } // reduce_float_64_true\n"
"\n"
"       .entry reduce_float_128_true (\n"
"               .param .u64 __cudaparm_reduce_float_128_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_128_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_128_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<12>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<21>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       407     0\n"
"$LDWbegin_reduce_float_128_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 256;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_128_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_9_15874;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 256;\n"
"       cvt.s64.u32     %rd1, %r7;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_128_true_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r7, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_9_13826:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f4, [%rd5+512];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_128_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_9_13826;\n"
"       bra.uni         $Lt_9_13314;\n"
"$Lt_9_15874:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_9_13314:\n"
"       .loc    3       195     0\n"
"       mov.f32         %f5, %f1;\n"
"       mov.f32         %f6, %f5;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r8, 63;\n"
"       setp.gt.u32     %p3, %r3, %r8;\n"
"       @%p3 bra        $Lt_9_14338;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f7, [%rd10+256];\n"
"       add.f32         %f6, %f7, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"$Lt_9_14338:\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 31;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_9_14850;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f8, [%rd10+128];\n"
"       add.f32         %f9, %f8, %f6;\n"
"       st.volatile.shared.f32  [%rd10+0], %f9;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f10, [%rd10+64];\n"
"       add.f32         %f11, %f10, %f9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f11;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f12, [%rd10+32];\n"
"       add.f32         %f13, %f12, %f11;\n"
"       st.volatile.shared.f32  [%rd10+0], %f13;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f14, [%rd10+16];\n"
"       add.f32         %f15, %f14, %f13;\n"
"       st.volatile.shared.f32  [%rd10+0], %f15;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f16, [%rd10+8];\n"
"       add.f32         %f17, %f16, %f15;\n"
"       st.volatile.shared.f32  [%rd10+0], %f17;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f18, [%rd10+4];\n"
"       add.f32         %f6, %f18, %f17;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"$Lt_9_14850:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r10, 0;\n"
"       setp.ne.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_9_15362;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f19, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_128_true_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f19;\n"
"$Lt_9_15362:\n"
"       .loc    3       410     0\n"
"       exit;\n"
"$LDWend_reduce_float_128_true:\n"
"       } // reduce_float_128_true\n"
"\n"
"       .entry reduce_float_256_true (\n"
"               .param .u64 __cudaparm_reduce_float_256_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_256_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_256_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<13>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<22>;\n"
"       .reg .pred %p<8>;\n"
"       .loc    3       412     0\n"
"$LDWbegin_reduce_float_256_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 512;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_256_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_10_16130;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 512;\n"
"       cvt.s64.u32     %rd1, %r7;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_256_true_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r7, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_10_13570:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f4, [%rd5+1024];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_256_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_10_13570;\n"
"       bra.uni         $Lt_10_13058;\n"
"$Lt_10_16130:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_10_13058:\n"
"       .loc    3       195     0\n"
"       mov.f32         %f5, %f1;\n"
"       mov.f32         %f6, %f5;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r8, 127;\n"
"       setp.gt.u32     %p3, %r3, %r8;\n"
"       @%p3 bra        $Lt_10_14082;\n"
"       .loc    3       76      0\n"
"       ld.volatile.shared.f32  %f7, [%rd10+512];\n"
"       add.f32         %f6, %f7, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"$Lt_10_14082:\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 63;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_10_14594;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f8, [%rd10+256];\n"
"       add.f32         %f6, %f8, %f6;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"$Lt_10_14594:\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 31;\n"
"       setp.gt.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_10_15106;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f9, [%rd10+128];\n"
"       add.f32         %f10, %f9, %f6;\n"
"       st.volatile.shared.f32  [%rd10+0], %f10;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f11, [%rd10+64];\n"
"       add.f32         %f12, %f11, %f10;\n"
"       st.volatile.shared.f32  [%rd10+0], %f12;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f13, [%rd10+32];\n"
"       add.f32         %f14, %f13, %f12;\n"
"       st.volatile.shared.f32  [%rd10+0], %f14;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f15, [%rd10+16];\n"
"       add.f32         %f16, %f15, %f14;\n"
"       st.volatile.shared.f32  [%rd10+0], %f16;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f17, [%rd10+8];\n"
"       add.f32         %f18, %f17, %f16;\n"
"       st.volatile.shared.f32  [%rd10+0], %f18;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f19, [%rd10+4];\n"
"       add.f32         %f6, %f19, %f18;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"$Lt_10_15106:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r11, 0;\n"
"       setp.ne.u32     %p6, %r3, %r11;\n"
"       @%p6 bra        $Lt_10_15618;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f20, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_256_true_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f20;\n"
"$Lt_10_15618:\n"
"       .loc    3       415     0\n"
"       exit;\n"
"$LDWend_reduce_float_256_true:\n"
"       } // reduce_float_256_true\n"
"\n"
"       .entry reduce_float_512_true (\n"
"               .param .u64 __cudaparm_reduce_float_512_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_512_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_512_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<23>;\n"
"       .reg .pred %p<9>;\n"
"       .loc    3       417     0\n"
"$LDWbegin_reduce_float_512_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 1024;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_512_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_11_16386;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 1024;\n"
"       cvt.s64.u32     %rd1, %r7;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_512_true_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r7, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_11_13314:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f4, [%rd5+2048];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_float_512_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_11_13314;\n"
"       bra.uni         $Lt_11_12802;\n"
"$Lt_11_16386:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_11_12802:\n"
"       .loc    3       195     0\n"
"       mov.f32         %f5, %f1;\n"
"       mov.f32         %f6, %f5;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r8, 255;\n"
"       setp.gt.u32     %p3, %r3, %r8;\n"
"       @%p3 bra        $Lt_11_13826;\n"
"       .loc    3       75      0\n"
"       ld.volatile.shared.f32  %f7, [%rd10+1024];\n"
"       add.f32         %f6, %f7, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"$Lt_11_13826:\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 127;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_11_14338;\n"
"       .loc    3       76      0\n"
"       ld.volatile.shared.f32  %f8, [%rd10+512];\n"
"       add.f32         %f6, %f8, %f6;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"$Lt_11_14338:\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 63;\n"
"       setp.gt.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_11_14850;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f9, [%rd10+256];\n"
"       add.f32         %f6, %f9, %f6;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"$Lt_11_14850:\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 31;\n"
"       setp.gt.u32     %p6, %r3, %r11;\n"
"       @%p6 bra        $Lt_11_15362;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f10, [%rd10+128];\n"
"       add.f32         %f11, %f10, %f6;\n"
"       st.volatile.shared.f32  [%rd10+0], %f11;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f12, [%rd10+64];\n"
"       add.f32         %f13, %f12, %f11;\n"
"       st.volatile.shared.f32  [%rd10+0], %f13;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f14, [%rd10+32];\n"
"       add.f32         %f15, %f14, %f13;\n"
"       st.volatile.shared.f32  [%rd10+0], %f15;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f16, [%rd10+16];\n"
"       add.f32         %f17, %f16, %f15;\n"
"       st.volatile.shared.f32  [%rd10+0], %f17;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f18, [%rd10+8];\n"
"       add.f32         %f19, %f18, %f17;\n"
"       st.volatile.shared.f32  [%rd10+0], %f19;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f20, [%rd10+4];\n"
"       add.f32         %f6, %f20, %f19;\n"
"       st.volatile.shared.f32  [%rd10+0], %f6;\n"
"$Lt_11_15362:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p7, %r3, %r12;\n"
"       @%p7 bra        $Lt_11_15874;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f21, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_512_true_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f21;\n"
"$Lt_11_15874:\n"
"       .loc    3       420     0\n"
"       exit;\n"
"$LDWend_reduce_float_512_true:\n"
"       } // reduce_float_512_true\n"
"\n"
"       .entry reduce_float_1_false (\n"
"               .param .u64 __cudaparm_reduce_float_1_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_1_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_1_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<11>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<6>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       423     0\n"
"$LDWbegin_reduce_float_1_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 2;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_1_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_12_17154;\n"
"       add.u32         %r6, %r4, 1;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_1_false_n];\n"
"       add.u32         %r7, %r5, 1;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 2;\n"
"       cvt.s64.u32     %rd1, %r8;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_1_false_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r8, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_12_15618:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_1_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_12_15874;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_12_15618\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f3, [%rd5+4];\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_12_15874:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_12_15618\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_12_15618;\n"
"       bra.uni         $Lt_12_15106;\n"
"$Lt_12_17154:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_12_15106:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       .loc    3       195     0\n"
"       mov.u32         %r9, 0;\n"
"       setp.ne.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_12_16642;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f4, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_1_false_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f4;\n"
"$Lt_12_16642:\n"
"       .loc    3       426     0\n"
"       exit;\n"
"$LDWend_reduce_float_1_false:\n"
"       } // reduce_float_1_false\n"
"\n"
"       .entry reduce_float_2_false (\n"
"               .param .u64 __cudaparm_reduce_float_2_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_2_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_2_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<12>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<8>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       428     0\n"
"$LDWbegin_reduce_float_2_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 4;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_2_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_13_17410;\n"
"       add.u32         %r6, %r4, 2;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_2_false_n];\n"
"       add.u32         %r7, %r5, 2;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 4;\n"
"       cvt.s64.u32     %rd1, %r8;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_2_false_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r8, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_13_15362:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_2_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_13_15618;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_13_15362\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f3, [%rd5+8];\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_13_15618:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_13_15362\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_13_15362;\n"
"       bra.uni         $Lt_13_14850;\n"
"$Lt_13_17410:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_13_14850:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 31;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_13_16386;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f4, [%rd10+4];\n"
"       add.f32         %f5, %f4, %f1;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"$Lt_13_16386:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r10, 0;\n"
"       setp.ne.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_13_16898;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f6, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_2_false_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f6;\n"
"$Lt_13_16898:\n"
"       .loc    3       431     0\n"
"       exit;\n"
"$LDWend_reduce_float_2_false:\n"
"       } // reduce_float_2_false\n"
"\n"
"       .entry reduce_float_4_false (\n"
"               .param .u64 __cudaparm_reduce_float_4_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_4_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_4_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<12>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<10>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       433     0\n"
"$LDWbegin_reduce_float_4_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 8;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_4_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_14_17154;\n"
"       add.u32         %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_4_false_n];\n"
"       add.u32         %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 8;\n"
"       cvt.s64.u32     %rd1, %r8;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_4_false_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r8, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_14_15106:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_4_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_14_15362;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_14_15106\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f3, [%rd5+16];\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_14_15362:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_14_15106\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_14_15106;\n"
"       bra.uni         $Lt_14_14594;\n"
"$Lt_14_17154:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_14_14594:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 31;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_14_16130;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f4, [%rd10+8];\n"
"       add.f32         %f5, %f4, %f1;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f6, [%rd10+4];\n"
"       add.f32         %f7, %f6, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f7;\n"
"$Lt_14_16130:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r10, 0;\n"
"       setp.ne.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_14_16642;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f8, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_4_false_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f8;\n"
"$Lt_14_16642:\n"
"       .loc    3       436     0\n"
"       exit;\n"
"$LDWend_reduce_float_4_false:\n"
"       } // reduce_float_4_false\n"
"\n"
"       .entry reduce_float_8_false (\n"
"               .param .u64 __cudaparm_reduce_float_8_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_8_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_8_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<12>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<12>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       438     0\n"
"$LDWbegin_reduce_float_8_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 16;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_8_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_15_16898;\n"
"       add.u32         %r6, %r4, 8;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_8_false_n];\n"
"       add.u32         %r7, %r5, 8;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 16;\n"
"       cvt.s64.u32     %rd1, %r8;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_8_false_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r8, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_15_14850:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_8_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_15_15106;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_15_14850\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f3, [%rd5+32];\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_15_15106:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_15_14850\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_15_14850;\n"
"       bra.uni         $Lt_15_14338;\n"
"$Lt_15_16898:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_15_14338:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 31;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_15_15874;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f4, [%rd10+16];\n"
"       add.f32         %f5, %f4, %f1;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f6, [%rd10+8];\n"
"       add.f32         %f7, %f6, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f7;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f8, [%rd10+4];\n"
"       add.f32         %f9, %f8, %f7;\n"
"       st.volatile.shared.f32  [%rd10+0], %f9;\n"
"$Lt_15_15874:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r10, 0;\n"
"       setp.ne.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_15_16386;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f10, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_8_false_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f10;\n"
"$Lt_15_16386:\n"
"       .loc    3       441     0\n"
"       exit;\n"
"$LDWend_reduce_float_8_false:\n"
"       } // reduce_float_8_false\n"
"\n"
"       .entry reduce_float_16_false (\n"
"               .param .u64 __cudaparm_reduce_float_16_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_16_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_16_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<12>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<14>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       443     0\n"
"$LDWbegin_reduce_float_16_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 32;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_16_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_16_16642;\n"
"       add.u32         %r6, %r4, 16;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_16_false_n];\n"
"       add.u32         %r7, %r5, 16;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 32;\n"
"       cvt.s64.u32     %rd1, %r8;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_16_false_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r8, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_16_14594:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_16_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_16_14850;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_16_14594\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f3, [%rd5+64];\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_16_14850:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_16_14594\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_16_14594;\n"
"       bra.uni         $Lt_16_14082;\n"
"$Lt_16_16642:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_16_14082:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 31;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_16_15618;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f4, [%rd10+32];\n"
"       add.f32         %f5, %f4, %f1;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f6, [%rd10+16];\n"
"       add.f32         %f7, %f6, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f7;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f8, [%rd10+8];\n"
"       add.f32         %f9, %f8, %f7;\n"
"       st.volatile.shared.f32  [%rd10+0], %f9;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f10, [%rd10+4];\n"
"       add.f32         %f11, %f10, %f9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f11;\n"
"$Lt_16_15618:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r10, 0;\n"
"       setp.ne.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_16_16130;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f12, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_16_false_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f12;\n"
"$Lt_16_16130:\n"
"       .loc    3       446     0\n"
"       exit;\n"
"$LDWend_reduce_float_16_false:\n"
"       } // reduce_float_16_false\n"
"\n"
"       .entry reduce_float_32_false (\n"
"               .param .u64 __cudaparm_reduce_float_32_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_32_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_32_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<12>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<16>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       448     0\n"
"$LDWbegin_reduce_float_32_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 64;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_32_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_17_16386;\n"
"       add.u32         %r6, %r4, 32;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_32_false_n];\n"
"       add.u32         %r7, %r5, 32;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 64;\n"
"       cvt.s64.u32     %rd1, %r8;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_32_false_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r8, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_17_14338:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_32_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_17_14594;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_17_14338\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f3, [%rd5+128];\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_17_14594:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_17_14338\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_17_14338;\n"
"       bra.uni         $Lt_17_13826;\n"
"$Lt_17_16386:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_17_13826:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 31;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_17_15362;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f4, [%rd10+64];\n"
"       add.f32         %f5, %f4, %f1;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f6, [%rd10+32];\n"
"       add.f32         %f7, %f6, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f7;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f8, [%rd10+16];\n"
"       add.f32         %f9, %f8, %f7;\n"
"       st.volatile.shared.f32  [%rd10+0], %f9;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f10, [%rd10+8];\n"
"       add.f32         %f11, %f10, %f9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f11;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f12, [%rd10+4];\n"
"       add.f32         %f13, %f12, %f11;\n"
"       st.volatile.shared.f32  [%rd10+0], %f13;\n"
"$Lt_17_15362:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r10, 0;\n"
"       setp.ne.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_17_15874;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f14, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_32_false_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f14;\n"
"$Lt_17_15874:\n"
"       .loc    3       451     0\n"
"       exit;\n"
"$LDWend_reduce_float_32_false:\n"
"       } // reduce_float_32_false\n"
"\n"
"       .entry reduce_float_64_false (\n"
"               .param .u64 __cudaparm_reduce_float_64_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_64_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_64_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<12>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<18>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       453     0\n"
"$LDWbegin_reduce_float_64_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 128;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_64_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_18_16130;\n"
"       add.u32         %r6, %r4, 64;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_64_false_n];\n"
"       add.u32         %r7, %r5, 64;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 128;\n"
"       cvt.s64.u32     %rd1, %r8;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_64_false_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r8, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_18_14082:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_64_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_18_14338;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_18_14082\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f3, [%rd5+256];\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_18_14338:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_18_14082\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_18_14082;\n"
"       bra.uni         $Lt_18_13570;\n"
"$Lt_18_16130:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_18_13570:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 31;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_18_15106;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f4, [%rd10+128];\n"
"       add.f32         %f5, %f4, %f1;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f6, [%rd10+64];\n"
"       add.f32         %f7, %f6, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f7;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f8, [%rd10+32];\n"
"       add.f32         %f9, %f8, %f7;\n"
"       st.volatile.shared.f32  [%rd10+0], %f9;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f10, [%rd10+16];\n"
"       add.f32         %f11, %f10, %f9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f11;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f12, [%rd10+8];\n"
"       add.f32         %f13, %f12, %f11;\n"
"       st.volatile.shared.f32  [%rd10+0], %f13;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f14, [%rd10+4];\n"
"       add.f32         %f15, %f14, %f13;\n"
"       st.volatile.shared.f32  [%rd10+0], %f15;\n"
"$Lt_18_15106:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r10, 0;\n"
"       setp.ne.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_18_15618;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f16, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_64_false_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f16;\n"
"$Lt_18_15618:\n"
"       .loc    3       456     0\n"
"       exit;\n"
"$LDWend_reduce_float_64_false:\n"
"       } // reduce_float_64_false\n"
"\n"
"       .entry reduce_float_128_false (\n"
"               .param .u64 __cudaparm_reduce_float_128_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_128_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_128_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<13>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<20>;\n"
"       .reg .pred %p<8>;\n"
"       .loc    3       458     0\n"
"$LDWbegin_reduce_float_128_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 256;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_128_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_19_16386;\n"
"       add.u32         %r6, %r4, 128;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_128_false_n];\n"
"       add.u32         %r7, %r5, 128;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 256;\n"
"       cvt.s64.u32     %rd1, %r8;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_128_false_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r8, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_19_13826:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_128_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_19_14082;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_19_13826\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f3, [%rd5+512];\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_19_14082:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_19_13826\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_19_13826;\n"
"       bra.uni         $Lt_19_13314;\n"
"$Lt_19_16386:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_19_13314:\n"
"       .loc    3       195     0\n"
"       mov.f32         %f4, %f1;\n"
"       mov.f32         %f5, %f4;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f4;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 63;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_19_14850;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f6, [%rd10+256];\n"
"       add.f32         %f5, %f6, %f4;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"$Lt_19_14850:\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 31;\n"
"       setp.gt.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_19_15362;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f7, [%rd10+128];\n"
"       add.f32         %f8, %f7, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f8;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f9, [%rd10+64];\n"
"       add.f32         %f10, %f9, %f8;\n"
"       st.volatile.shared.f32  [%rd10+0], %f10;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f11, [%rd10+32];\n"
"       add.f32         %f12, %f11, %f10;\n"
"       st.volatile.shared.f32  [%rd10+0], %f12;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f13, [%rd10+16];\n"
"       add.f32         %f14, %f13, %f12;\n"
"       st.volatile.shared.f32  [%rd10+0], %f14;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f15, [%rd10+8];\n"
"       add.f32         %f16, %f15, %f14;\n"
"       st.volatile.shared.f32  [%rd10+0], %f16;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f17, [%rd10+4];\n"
"       add.f32         %f5, %f17, %f16;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"$Lt_19_15362:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r11, 0;\n"
"       setp.ne.u32     %p6, %r3, %r11;\n"
"       @%p6 bra        $Lt_19_15874;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f18, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_128_false_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f18;\n"
"$Lt_19_15874:\n"
"       .loc    3       461     0\n"
"       exit;\n"
"$LDWend_reduce_float_128_false:\n"
"       } // reduce_float_128_false\n"
"\n"
"       .entry reduce_float_256_false (\n"
"               .param .u64 __cudaparm_reduce_float_256_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_256_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_256_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<21>;\n"
"       .reg .pred %p<9>;\n"
"       .loc    3       463     0\n"
"$LDWbegin_reduce_float_256_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 512;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_256_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_20_16642;\n"
"       add.u32         %r6, %r4, 256;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_256_false_n];\n"
"       add.u32         %r7, %r5, 256;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 512;\n"
"       cvt.s64.u32     %rd1, %r8;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_256_false_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r8, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_20_13570:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_256_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_20_13826;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_20_13570\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f3, [%rd5+1024];\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_20_13826:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_20_13570\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_20_13570;\n"
"       bra.uni         $Lt_20_13058;\n"
"$Lt_20_16642:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_20_13058:\n"
"       .loc    3       195     0\n"
"       mov.f32         %f4, %f1;\n"
"       mov.f32         %f5, %f4;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f4;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 127;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_20_14594;\n"
"       .loc    3       76      0\n"
"       ld.volatile.shared.f32  %f6, [%rd10+512];\n"
"       add.f32         %f5, %f6, %f4;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"$Lt_20_14594:\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 63;\n"
"       setp.gt.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_20_15106;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f7, [%rd10+256];\n"
"       add.f32         %f5, %f7, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"$Lt_20_15106:\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 31;\n"
"       setp.gt.u32     %p6, %r3, %r11;\n"
"       @%p6 bra        $Lt_20_15618;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f8, [%rd10+128];\n"
"       add.f32         %f9, %f8, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f9;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f10, [%rd10+64];\n"
"       add.f32         %f11, %f10, %f9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f11;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f12, [%rd10+32];\n"
"       add.f32         %f13, %f12, %f11;\n"
"       st.volatile.shared.f32  [%rd10+0], %f13;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f14, [%rd10+16];\n"
"       add.f32         %f15, %f14, %f13;\n"
"       st.volatile.shared.f32  [%rd10+0], %f15;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f16, [%rd10+8];\n"
"       add.f32         %f17, %f16, %f15;\n"
"       st.volatile.shared.f32  [%rd10+0], %f17;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f18, [%rd10+4];\n"
"       add.f32         %f5, %f18, %f17;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"$Lt_20_15618:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p7, %r3, %r12;\n"
"       @%p7 bra        $Lt_20_16130;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f19, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_256_false_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f19;\n"
"$Lt_20_16130:\n"
"       .loc    3       466     0\n"
"       exit;\n"
"$LDWend_reduce_float_256_false:\n"
"       } // reduce_float_256_false\n"
"\n"
"       .entry reduce_float_512_false (\n"
"               .param .u64 __cudaparm_reduce_float_512_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_float_512_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_float_512_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<15>;\n"
"       .reg .u64 %rd<16>;\n"
"       .reg .f32 %f<22>;\n"
"       .reg .pred %p<10>;\n"
"       .loc    3       468     0\n"
"$LDWbegin_reduce_float_512_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 1024;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_512_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_21_16898;\n"
"       add.u32         %r6, %r4, 512;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_512_false_n];\n"
"       add.u32         %r7, %r5, 512;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 1024;\n"
"       cvt.s64.u32     %rd1, %r8;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_float_512_false_g_idata];\n"
"       cvt.u64.u32     %rd3, %r4;\n"
"       mul.wide.u32    %rd4, %r4, 4;\n"
"       add.u64         %rd5, %rd2, %rd4;\n"
"       mul.wide.u32    %rd6, %r8, 4;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_21_13314:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.f32   %f2, [%rd5+0];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_float_512_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_21_13570;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_21_13314\n"
"       .loc    3       191     0\n"
"       ld.global.f32   %f3, [%rd5+2048];\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_21_13570:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_21_13314\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd5, %rd5, %rd6;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_21_13314;\n"
"       bra.uni         $Lt_21_12802;\n"
"$Lt_21_16898:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_21_12802:\n"
"       .loc    3       195     0\n"
"       mov.f32         %f4, %f1;\n"
"       mov.f32         %f5, %f4;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd7, __smem;\n"
"       cvt.u64.u32     %rd8, %r3;\n"
"       mul.wide.u32    %rd9, %r3, 4;\n"
"       add.u64         %rd10, %rd7, %rd9;\n"
"       st.volatile.shared.f32  [%rd10+0], %f4;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r9, 255;\n"
"       setp.gt.u32     %p4, %r3, %r9;\n"
"       @%p4 bra        $Lt_21_14338;\n"
"       .loc    3       75      0\n"
"       ld.volatile.shared.f32  %f6, [%rd10+1024];\n"
"       add.f32         %f5, %f6, %f4;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"$Lt_21_14338:\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 127;\n"
"       setp.gt.u32     %p5, %r3, %r10;\n"
"       @%p5 bra        $Lt_21_14850;\n"
"       .loc    3       76      0\n"
"       ld.volatile.shared.f32  %f7, [%rd10+512];\n"
"       add.f32         %f5, %f7, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"$Lt_21_14850:\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 63;\n"
"       setp.gt.u32     %p6, %r3, %r11;\n"
"       @%p6 bra        $Lt_21_15362;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f8, [%rd10+256];\n"
"       add.f32         %f5, %f8, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"$Lt_21_15362:\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 31;\n"
"       setp.gt.u32     %p7, %r3, %r12;\n"
"       @%p7 bra        $Lt_21_15874;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f9, [%rd10+128];\n"
"       add.f32         %f10, %f9, %f5;\n"
"       st.volatile.shared.f32  [%rd10+0], %f10;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f11, [%rd10+64];\n"
"       add.f32         %f12, %f11, %f10;\n"
"       st.volatile.shared.f32  [%rd10+0], %f12;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f13, [%rd10+32];\n"
"       add.f32         %f14, %f13, %f12;\n"
"       st.volatile.shared.f32  [%rd10+0], %f14;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f15, [%rd10+16];\n"
"       add.f32         %f16, %f15, %f14;\n"
"       st.volatile.shared.f32  [%rd10+0], %f16;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f17, [%rd10+8];\n"
"       add.f32         %f18, %f17, %f16;\n"
"       st.volatile.shared.f32  [%rd10+0], %f18;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f19, [%rd10+4];\n"
"       add.f32         %f5, %f19, %f18;\n"
"       st.volatile.shared.f32  [%rd10+0], %f5;\n"
"$Lt_21_15874:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r13, 0;\n"
"       setp.ne.u32     %p8, %r3, %r13;\n"
"       @%p8 bra        $Lt_21_16386;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f20, [__smem+0];\n"
"       ld.param.u64    %rd11, [__cudaparm_reduce_float_512_false_g_odata];\n"
"       cvt.u64.u32     %rd12, %r1;\n"
"       mul.wide.u32    %rd13, %r1, 4;\n"
"       add.u64         %rd14, %rd11, %rd13;\n"
"       st.global.f32   [%rd14+0], %f20;\n"
"$Lt_21_16386:\n"
"       .loc    3       471     0\n"
"       exit;\n"
"$LDWend_reduce_float_512_false:\n"
"       } // reduce_float_512_false\n"
"\n"
"       .entry tex_reduce_256_false (\n"
"               .param .u64 __cudaparm_tex_reduce_256_false_g_odata,\n"
"               .param .u32 __cudaparm_tex_reduce_256_false_n,\n"
"               .param .u32 __cudaparm_tex_reduce_256_false_stride)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<19>;\n"
"       .reg .u64 %rd<10>;\n"
"       .reg .f32 %f<37>;\n"
"       .reg .pred %p<9>;\n"
"       .loc    3       477     0\n"
"$LDWbegin_tex_reduce_256_false:\n"
"       .loc    3       240     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 512;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_tex_reduce_256_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_22_16642;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 512;\n"
"       ld.param.u32    %r8, [__cudaparm_tex_reduce_256_false_stride];\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_22_13570:\n"
" //<loop> Loop body line 240, nesting depth: 1, estimated iterations: unknown\n"
"       ld.param.u32    %r8, [__cudaparm_tex_reduce_256_false_stride];\n"
"       rem.u32         %r9, %r5, %r8;\n"
"       cvt.rn.f32.u32  %f2, %r9;\n"
"       div.u32         %r10, %r5, %r8;\n"
"       cvt.rn.f32.u32  %f3, %r10;\n"
"       mov.f32         %f4, 0f00000000;        // 0\n"
"       mov.f32         %f5, 0f00000000;        // 0\n"
"       tex.2d.v4.f32.f32 {%f6,%f7,%f8,%f9},[tex_ref_1,{%f2,%f3,%f4,%f5}];\n"
" //<loop> Part of loop body line 240, head labeled $Lt_22_13570\n"
"       .loc    3       248     0\n"
"       mov.f32         %f10, %f6;\n"
"       add.f32         %f1, %f10, %f1;\n"
"       add.u32         %r11, %r5, 256;\n"
"       .loc    3       240     0\n"
"       ld.param.u32    %r6, [__cudaparm_tex_reduce_256_false_n];\n"
"       .loc    3       248     0\n"
"       setp.ge.u32     %p2, %r11, %r6;\n"
"       @%p2 bra        $Lt_22_13826;\n"
" //<loop> Part of loop body line 240, head labeled $Lt_22_13570\n"
"       .loc    3       240     0\n"
"       ld.param.u32    %r8, [__cudaparm_tex_reduce_256_false_stride];\n"
"       .loc    3       248     0\n"
"       rem.u32         %r12, %r11, %r8;\n"
"       cvt.rn.f32.u32  %f11, %r12;\n"
"       div.u32         %r13, %r11, %r8;\n"
"       cvt.rn.f32.u32  %f12, %r13;\n"
"       mov.f32         %f13, 0f00000000;       // 0\n"
"       mov.f32         %f14, 0f00000000;       // 0\n"
"       tex.2d.v4.f32.f32 {%f15,%f16,%f17,%f18},[tex_ref_1,{%f11,%f12,%f13,%f14}];\n"
" //<loop> Part of loop body line 240, head labeled $Lt_22_13570\n"
"       .loc    3       253     0\n"
"       mov.f32         %f19, %f15;\n"
"       add.f32         %f1, %f19, %f1;\n"
"$Lt_22_13826:\n"
" //<loop> Part of loop body line 240, head labeled $Lt_22_13570\n"
"       add.u32         %r5, %r7, %r5;\n"
"       .loc    3       240     0\n"
"       ld.param.u32    %r6, [__cudaparm_tex_reduce_256_false_n];\n"
"       .loc    3       253     0\n"
"       setp.lt.u32     %p3, %r5, %r6;\n"
"       @%p3 bra        $Lt_22_13570;\n"
"       bra.uni         $Lt_22_13058;\n"
"$Lt_22_16642:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_22_13058:\n"
"       .loc    3       258     0\n"
"       mov.f32         %f20, %f1;\n"
"       mov.f32         %f21, %f20;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd1, __smem;\n"
"       cvt.u64.u32     %rd2, %r3;\n"
"       mul.wide.u32    %rd3, %r3, 4;\n"
"       add.u64         %rd4, %rd1, %rd3;\n"
"       st.volatile.shared.f32  [%rd4+0], %f20;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r14, 127;\n"
"       setp.gt.u32     %p4, %r3, %r14;\n"
"       @%p4 bra        $Lt_22_14594;\n"
"       .loc    3       76      0\n"
"       ld.volatile.shared.f32  %f22, [%rd4+512];\n"
"       add.f32         %f21, %f22, %f20;\n"
"       st.volatile.shared.f32  [%rd4+0], %f21;\n"
"$Lt_22_14594:\n"
"       bar.sync        0;\n"
"       mov.u32         %r15, 63;\n"
"       setp.gt.u32     %p5, %r3, %r15;\n"
"       @%p5 bra        $Lt_22_15106;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f23, [%rd4+256];\n"
"       add.f32         %f21, %f23, %f21;\n"
"       st.volatile.shared.f32  [%rd4+0], %f21;\n"
"$Lt_22_15106:\n"
"       bar.sync        0;\n"
"       mov.u32         %r16, 31;\n"
"       setp.gt.u32     %p6, %r3, %r16;\n"
"       @%p6 bra        $Lt_22_15618;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f24, [%rd4+128];\n"
"       add.f32         %f25, %f24, %f21;\n"
"       st.volatile.shared.f32  [%rd4+0], %f25;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f26, [%rd4+64];\n"
"       add.f32         %f27, %f26, %f25;\n"
"       st.volatile.shared.f32  [%rd4+0], %f27;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f28, [%rd4+32];\n"
"       add.f32         %f29, %f28, %f27;\n"
"       st.volatile.shared.f32  [%rd4+0], %f29;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f30, [%rd4+16];\n"
"       add.f32         %f31, %f30, %f29;\n"
"       st.volatile.shared.f32  [%rd4+0], %f31;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f32, [%rd4+8];\n"
"       add.f32         %f33, %f32, %f31;\n"
"       st.volatile.shared.f32  [%rd4+0], %f33;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f34, [%rd4+4];\n"
"       add.f32         %f21, %f34, %f33;\n"
"       st.volatile.shared.f32  [%rd4+0], %f21;\n"
"$Lt_22_15618:\n"
"       .loc    3       258     0\n"
"       mov.u32         %r17, 0;\n"
"       setp.ne.u32     %p7, %r3, %r17;\n"
"       @%p7 bra        $Lt_22_16130;\n"
"       .loc    3       262     0\n"
"       ld.shared.f32   %f35, [__smem+0];\n"
"       ld.param.u64    %rd5, [__cudaparm_tex_reduce_256_false_g_odata];\n"
"       cvt.u64.u32     %rd6, %r1;\n"
"       mul.wide.u32    %rd7, %r1, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.global.f32   [%rd8+0], %f35;\n"
"$Lt_22_16130:\n"
"       .loc    3       480     0\n"
"       exit;\n"
"$LDWend_tex_reduce_256_false:\n"
"       } // tex_reduce_256_false\n"
"\n"
"       .entry tex_count_256_false (\n"
"               .param .u64 __cudaparm_tex_count_256_false_g_odata,\n"
"               .param .u32 __cudaparm_tex_count_256_false_n,\n"
"               .param .u32 __cudaparm_tex_count_256_false_stride)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<19>;\n"
"       .reg .u64 %rd<10>;\n"
"       .reg .f32 %f<41>;\n"
"       .reg .f64 %fd<6>;\n"
"       .reg .pred %p<11>;\n"
"       .loc    3       482     0\n"
"$LDWbegin_tex_count_256_false:\n"
"       .loc    3       333     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 512;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_tex_count_256_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_23_18178;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 512;\n"
"       ld.param.u32    %r8, [__cudaparm_tex_count_256_false_stride];\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_23_15106:\n"
" //<loop> Loop body line 333, nesting depth: 1, estimated iterations: unknown\n"
"       ld.param.u32    %r8, [__cudaparm_tex_count_256_false_stride];\n"
"       rem.u32         %r9, %r5, %r8;\n"
"       cvt.rn.f32.u32  %f2, %r9;\n"
"       div.u32         %r10, %r5, %r8;\n"
"       cvt.rn.f32.u32  %f3, %r10;\n"
"       mov.f32         %f4, 0f00000000;        // 0\n"
"       mov.f32         %f5, 0f00000000;        // 0\n"
"       tex.2d.v4.f32.f32 {%f6,%f7,%f8,%f9},[tex_ref_1,{%f2,%f3,%f4,%f5}];\n"
" //<loop> Part of loop body line 333, head labeled $Lt_23_15106\n"
"       .loc    3       341     0\n"
"       mov.f32         %f10, %f6;\n"
"       .loc    3       340     0\n"
"       mov.f32         %f11, 0f3f800000;       // 1\n"
"       add.f32         %f12, %f1, %f11;\n"
"       cvt.f64.f32     %fd1, %f10;\n"
"       mov.f64         %fd2, 0d0000000000000000;       // 0\n"
"       setp.ne.f64     %p2, %fd1, %fd2;\n"
"       selp.f32        %f1, %f12, %f1, %p2;\n"
"       add.u32         %r11, %r5, 256;\n"
"       .loc    3       333     0\n"
"       ld.param.u32    %r6, [__cudaparm_tex_count_256_false_n];\n"
"       .loc    3       340     0\n"
"       setp.ge.u32     %p3, %r11, %r6;\n"
"       @%p3 bra        $Lt_23_15362;\n"
" //<loop> Part of loop body line 333, head labeled $Lt_23_15106\n"
"       .loc    3       333     0\n"
"       ld.param.u32    %r8, [__cudaparm_tex_count_256_false_stride];\n"
"       .loc    3       340     0\n"
"       rem.u32         %r12, %r11, %r8;\n"
"       cvt.rn.f32.u32  %f13, %r12;\n"
"       div.u32         %r13, %r11, %r8;\n"
"       cvt.rn.f32.u32  %f14, %r13;\n"
"       mov.f32         %f15, 0f00000000;       // 0\n"
"       mov.f32         %f16, 0f00000000;       // 0\n"
"       tex.2d.v4.f32.f32 {%f17,%f18,%f19,%f20},[tex_ref_1,{%f13,%f14,%f15,%f16}];\n"
" //<loop> Part of loop body line 333, head labeled $Lt_23_15106\n"
"       .loc    3       348     0\n"
"       mov.f32         %f21, %f17;\n"
"       .loc    3       347     0\n"
"       mov.f32         %f22, 0f3f800000;       // 1\n"
"       add.f32         %f23, %f1, %f22;\n"
"       cvt.f64.f32     %fd3, %f21;\n"
"       mov.f64         %fd4, 0d0000000000000000;       // 0\n"
"       setp.ne.f64     %p4, %fd3, %fd4;\n"
"       selp.f32        %f1, %f23, %f1, %p4;\n"
"$Lt_23_15362:\n"
" //<loop> Part of loop body line 333, head labeled $Lt_23_15106\n"
"       add.u32         %r5, %r7, %r5;\n"
"       .loc    3       333     0\n"
"       ld.param.u32    %r6, [__cudaparm_tex_count_256_false_n];\n"
"       .loc    3       347     0\n"
"       setp.lt.u32     %p5, %r5, %r6;\n"
"       @%p5 bra        $Lt_23_15106;\n"
"       bra.uni         $Lt_23_14594;\n"
"$Lt_23_18178:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_23_14594:\n"
"       .loc    3       355     0\n"
"       mov.f32         %f24, %f1;\n"
"       mov.f32         %f25, %f24;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd1, __smem;\n"
"       cvt.u64.u32     %rd2, %r3;\n"
"       mul.wide.u32    %rd3, %r3, 4;\n"
"       add.u64         %rd4, %rd1, %rd3;\n"
"       st.volatile.shared.f32  [%rd4+0], %f24;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r14, 127;\n"
"       setp.gt.u32     %p6, %r3, %r14;\n"
"       @%p6 bra        $Lt_23_16130;\n"
"       .loc    3       76      0\n"
"       ld.volatile.shared.f32  %f26, [%rd4+512];\n"
"       add.f32         %f25, %f26, %f24;\n"
"       st.volatile.shared.f32  [%rd4+0], %f25;\n"
"$Lt_23_16130:\n"
"       bar.sync        0;\n"
"       mov.u32         %r15, 63;\n"
"       setp.gt.u32     %p7, %r3, %r15;\n"
"       @%p7 bra        $Lt_23_16642;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f27, [%rd4+256];\n"
"       add.f32         %f25, %f27, %f25;\n"
"       st.volatile.shared.f32  [%rd4+0], %f25;\n"
"$Lt_23_16642:\n"
"       bar.sync        0;\n"
"       mov.u32         %r16, 31;\n"
"       setp.gt.u32     %p8, %r3, %r16;\n"
"       @%p8 bra        $Lt_23_17154;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f28, [%rd4+128];\n"
"       add.f32         %f29, %f28, %f25;\n"
"       st.volatile.shared.f32  [%rd4+0], %f29;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f30, [%rd4+64];\n"
"       add.f32         %f31, %f30, %f29;\n"
"       st.volatile.shared.f32  [%rd4+0], %f31;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f32, [%rd4+32];\n"
"       add.f32         %f33, %f32, %f31;\n"
"       st.volatile.shared.f32  [%rd4+0], %f33;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f34, [%rd4+16];\n"
"       add.f32         %f35, %f34, %f33;\n"
"       st.volatile.shared.f32  [%rd4+0], %f35;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f36, [%rd4+8];\n"
"       add.f32         %f37, %f36, %f35;\n"
"       st.volatile.shared.f32  [%rd4+0], %f37;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f38, [%rd4+4];\n"
"       add.f32         %f25, %f38, %f37;\n"
"       st.volatile.shared.f32  [%rd4+0], %f25;\n"
"$Lt_23_17154:\n"
"       .loc    3       355     0\n"
"       mov.u32         %r17, 0;\n"
"       setp.ne.u32     %p9, %r3, %r17;\n"
"       @%p9 bra        $Lt_23_17666;\n"
"       .loc    3       359     0\n"
"       ld.shared.f32   %f39, [__smem+0];\n"
"       ld.param.u64    %rd5, [__cudaparm_tex_count_256_false_g_odata];\n"
"       cvt.u64.u32     %rd6, %r1;\n"
"       mul.wide.u32    %rd7, %r1, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.global.f32   [%rd8+0], %f39;\n"
"$Lt_23_17666:\n"
"       .loc    3       485     0\n"
"       exit;\n"
"$LDWend_tex_count_256_false:\n"
"       } // tex_count_256_false\n"
"\n"
"       .entry chamfer_reduce_256_false (\n"
"               .param .u64 __cudaparm_chamfer_reduce_256_false_g_odata,\n"
"               .param .u32 __cudaparm_chamfer_reduce_256_false_n,\n"
"               .param .u32 __cudaparm_chamfer_reduce_256_false_stride)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<19>;\n"
"       .reg .u64 %rd<10>;\n"
"       .reg .f32 %f<59>;\n"
"       .reg .pred %p<9>;\n"
"       .loc    3       487     0\n"
"$LDWbegin_chamfer_reduce_256_false:\n"
"       .loc    3       287     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 512;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_chamfer_reduce_256_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_24_16642;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 512;\n"
"       ld.param.u32    %r8, [__cudaparm_chamfer_reduce_256_false_stride];\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_24_13570:\n"
" //<loop> Loop body line 287, nesting depth: 1, estimated iterations: unknown\n"
"       ld.param.u32    %r8, [__cudaparm_chamfer_reduce_256_false_stride];\n"
"       rem.u32         %r9, %r5, %r8;\n"
"       cvt.rn.f32.u32  %f2, %r9;\n"
"       div.u32         %r10, %r5, %r8;\n"
"       cvt.rn.f32.u32  %f3, %r10;\n"
"       mov.f32         %f4, %f2;\n"
"       mov.f32         %f5, %f3;\n"
"       mov.f32         %f6, 0f00000000;        // 0\n"
"       mov.f32         %f7, 0f00000000;        // 0\n"
"       tex.2d.v4.f32.f32 {%f8,%f9,%f10,%f11},[tex_ref_1,{%f4,%f5,%f6,%f7}];\n"
" //<loop> Part of loop body line 287, head labeled $Lt_24_13570\n"
"       .loc    3       295     0\n"
"       mov.f32         %f12, %f8;\n"
"       mov.f32         %f13, %f2;\n"
"       mov.f32         %f14, %f3;\n"
"       mov.f32         %f15, 0f00000000;       // 0\n"
"       mov.f32         %f16, 0f00000000;       // 0\n"
"       tex.2d.v4.f32.f32 {%f17,%f18,%f19,%f20},[tex_ref_2,{%f13,%f14,%f15,%f16}];\n"
" //<loop> Part of loop body line 287, head labeled $Lt_24_13570\n"
"       mov.f32         %f21, %f17;\n"
"       mad.f32         %f1, %f12, %f21, %f1;\n"
"       add.u32         %r11, %r5, 256;\n"
"       .loc    3       287     0\n"
"       ld.param.u32    %r6, [__cudaparm_chamfer_reduce_256_false_n];\n"
"       .loc    3       295     0\n"
"       setp.ge.u32     %p2, %r11, %r6;\n"
"       @%p2 bra        $Lt_24_13826;\n"
" //<loop> Part of loop body line 287, head labeled $Lt_24_13570\n"
"       .loc    3       287     0\n"
"       ld.param.u32    %r8, [__cudaparm_chamfer_reduce_256_false_stride];\n"
"       .loc    3       295     0\n"
"       rem.u32         %r12, %r11, %r8;\n"
"       cvt.rn.f32.u32  %f22, %r12;\n"
"       div.u32         %r13, %r11, %r8;\n"
"       cvt.rn.f32.u32  %f23, %r13;\n"
"       mov.f32         %f24, %f22;\n"
"       mov.f32         %f25, %f23;\n"
"       mov.f32         %f26, 0f00000000;       // 0\n"
"       mov.f32         %f27, 0f00000000;       // 0\n"
"       tex.2d.v4.f32.f32 {%f28,%f29,%f30,%f31},[tex_ref_1,{%f24,%f25,%f26,%f27}];\n"
" //<loop> Part of loop body line 287, head labeled $Lt_24_13570\n"
"       .loc    3       300     0\n"
"       mov.f32         %f32, %f28;\n"
"       mov.f32         %f33, %f22;\n"
"       mov.f32         %f34, %f23;\n"
"       mov.f32         %f35, 0f00000000;       // 0\n"
"       mov.f32         %f36, 0f00000000;       // 0\n"
"       tex.2d.v4.f32.f32 {%f37,%f38,%f39,%f40},[tex_ref_2,{%f33,%f34,%f35,%f36}];\n"
" //<loop> Part of loop body line 287, head labeled $Lt_24_13570\n"
"       mov.f32         %f41, %f37;\n"
"       mad.f32         %f1, %f32, %f41, %f1;\n"
"$Lt_24_13826:\n"
" //<loop> Part of loop body line 287, head labeled $Lt_24_13570\n"
"       add.u32         %r5, %r7, %r5;\n"
"       .loc    3       287     0\n"
"       ld.param.u32    %r6, [__cudaparm_chamfer_reduce_256_false_n];\n"
"       .loc    3       300     0\n"
"       setp.lt.u32     %p3, %r5, %r6;\n"
"       @%p3 bra        $Lt_24_13570;\n"
"       bra.uni         $Lt_24_13058;\n"
"$Lt_24_16642:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_24_13058:\n"
"       .loc    3       305     0\n"
"       mov.f32         %f42, %f1;\n"
"       mov.f32         %f43, %f42;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd1, __smem;\n"
"       cvt.u64.u32     %rd2, %r3;\n"
"       mul.wide.u32    %rd3, %r3, 4;\n"
"       add.u64         %rd4, %rd1, %rd3;\n"
"       st.volatile.shared.f32  [%rd4+0], %f42;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r14, 127;\n"
"       setp.gt.u32     %p4, %r3, %r14;\n"
"       @%p4 bra        $Lt_24_14594;\n"
"       .loc    3       76      0\n"
"       ld.volatile.shared.f32  %f44, [%rd4+512];\n"
"       add.f32         %f43, %f44, %f42;\n"
"       st.volatile.shared.f32  [%rd4+0], %f43;\n"
"$Lt_24_14594:\n"
"       bar.sync        0;\n"
"       mov.u32         %r15, 63;\n"
"       setp.gt.u32     %p5, %r3, %r15;\n"
"       @%p5 bra        $Lt_24_15106;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f45, [%rd4+256];\n"
"       add.f32         %f43, %f45, %f43;\n"
"       st.volatile.shared.f32  [%rd4+0], %f43;\n"
"$Lt_24_15106:\n"
"       bar.sync        0;\n"
"       mov.u32         %r16, 31;\n"
"       setp.gt.u32     %p6, %r3, %r16;\n"
"       @%p6 bra        $Lt_24_15618;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f46, [%rd4+128];\n"
"       add.f32         %f47, %f46, %f43;\n"
"       st.volatile.shared.f32  [%rd4+0], %f47;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f48, [%rd4+64];\n"
"       add.f32         %f49, %f48, %f47;\n"
"       st.volatile.shared.f32  [%rd4+0], %f49;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f50, [%rd4+32];\n"
"       add.f32         %f51, %f50, %f49;\n"
"       st.volatile.shared.f32  [%rd4+0], %f51;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f52, [%rd4+16];\n"
"       add.f32         %f53, %f52, %f51;\n"
"       st.volatile.shared.f32  [%rd4+0], %f53;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f54, [%rd4+8];\n"
"       add.f32         %f55, %f54, %f53;\n"
"       st.volatile.shared.f32  [%rd4+0], %f55;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f56, [%rd4+4];\n"
"       add.f32         %f43, %f56, %f55;\n"
"       st.volatile.shared.f32  [%rd4+0], %f43;\n"
"$Lt_24_15618:\n"
"       .loc    3       305     0\n"
"       mov.u32         %r17, 0;\n"
"       setp.ne.u32     %p7, %r3, %r17;\n"
"       @%p7 bra        $Lt_24_16130;\n"
"       .loc    3       309     0\n"
"       ld.shared.f32   %f57, [__smem+0];\n"
"       ld.param.u64    %rd5, [__cudaparm_chamfer_reduce_256_false_g_odata];\n"
"       cvt.u64.u32     %rd6, %r1;\n"
"       mul.wide.u32    %rd7, %r1, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.global.f32   [%rd8+0], %f57;\n"
"$Lt_24_16130:\n"
"       .loc    3       490     0\n"
"       exit;\n"
"$LDWend_chamfer_reduce_256_false:\n"
"       } // chamfer_reduce_256_false\n"
"\n"
"       .entry reduce_uchar_1_true (\n"
"               .param .u64 __cudaparm_reduce_uchar_1_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_1_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_1_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<12>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<7>;\n"
"       .reg .pred %p<5>;\n"
"       .loc    3       500     0\n"
"$LDWbegin_reduce_uchar_1_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 2;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_1_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_25_16642;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 2;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_1_true_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r7;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_25_15618:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r8, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r8;\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r9, [%rd3+1];\n"
"       cvt.rn.f32.u32  %f4, %r9;\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_1_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_25_15618;\n"
"       bra.uni         $Lt_25_15106;\n"
"$Lt_25_16642:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_25_15106:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       .loc    3       195     0\n"
"       mov.u32         %r10, 0;\n"
"       setp.ne.u32     %p3, %r3, %r10;\n"
"       @%p3 bra        $Lt_25_16130;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f5, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_1_true_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f5;\n"
"$Lt_25_16130:\n"
"       .loc    3       503     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_1_true:\n"
"       } // reduce_uchar_1_true\n"
"\n"
"       .entry reduce_uchar_2_true (\n"
"               .param .u64 __cudaparm_reduce_uchar_2_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_2_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_2_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<13>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<9>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       505     0\n"
"$LDWbegin_reduce_uchar_2_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 4;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_2_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_26_16898;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 4;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_2_true_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r7;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_26_15362:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r8, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r8;\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r9, [%rd3+2];\n"
"       cvt.rn.f32.u32  %f4, %r9;\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_2_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_26_15362;\n"
"       bra.uni         $Lt_26_14850;\n"
"$Lt_26_16898:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_26_14850:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 31;\n"
"       setp.gt.u32     %p3, %r3, %r10;\n"
"       @%p3 bra        $Lt_26_15874;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f5, [%rd8+4];\n"
"       add.f32         %f6, %f5, %f1;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"$Lt_26_15874:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r11, 0;\n"
"       setp.ne.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_26_16386;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f7, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_2_true_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f7;\n"
"$Lt_26_16386:\n"
"       .loc    3       508     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_2_true:\n"
"       } // reduce_uchar_2_true\n"
"\n"
"       .entry reduce_uchar_4_true (\n"
"               .param .u64 __cudaparm_reduce_uchar_4_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_4_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_4_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<13>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<11>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       510     0\n"
"$LDWbegin_reduce_uchar_4_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 8;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_4_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_27_16642;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 8;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_4_true_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r7;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_27_15106:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r8, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r8;\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r9, [%rd3+4];\n"
"       cvt.rn.f32.u32  %f4, %r9;\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_4_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_27_15106;\n"
"       bra.uni         $Lt_27_14594;\n"
"$Lt_27_16642:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_27_14594:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 31;\n"
"       setp.gt.u32     %p3, %r3, %r10;\n"
"       @%p3 bra        $Lt_27_15618;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f5, [%rd8+8];\n"
"       add.f32         %f6, %f5, %f1;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f7, [%rd8+4];\n"
"       add.f32         %f8, %f7, %f6;\n"
"       st.volatile.shared.f32  [%rd8+0], %f8;\n"
"$Lt_27_15618:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r11, 0;\n"
"       setp.ne.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_27_16130;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f9, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_4_true_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f9;\n"
"$Lt_27_16130:\n"
"       .loc    3       513     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_4_true:\n"
"       } // reduce_uchar_4_true\n"
"\n"
"       .entry reduce_uchar_8_true (\n"
"               .param .u64 __cudaparm_reduce_uchar_8_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_8_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_8_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<13>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<13>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       515     0\n"
"$LDWbegin_reduce_uchar_8_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 16;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_8_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_28_16386;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 16;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_8_true_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r7;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_28_14850:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r8, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r8;\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r9, [%rd3+8];\n"
"       cvt.rn.f32.u32  %f4, %r9;\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_8_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_28_14850;\n"
"       bra.uni         $Lt_28_14338;\n"
"$Lt_28_16386:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_28_14338:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 31;\n"
"       setp.gt.u32     %p3, %r3, %r10;\n"
"       @%p3 bra        $Lt_28_15362;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f5, [%rd8+16];\n"
"       add.f32         %f6, %f5, %f1;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f7, [%rd8+8];\n"
"       add.f32         %f8, %f7, %f6;\n"
"       st.volatile.shared.f32  [%rd8+0], %f8;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f9, [%rd8+4];\n"
"       add.f32         %f10, %f9, %f8;\n"
"       st.volatile.shared.f32  [%rd8+0], %f10;\n"
"$Lt_28_15362:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r11, 0;\n"
"       setp.ne.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_28_15874;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f11, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_8_true_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f11;\n"
"$Lt_28_15874:\n"
"       .loc    3       518     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_8_true:\n"
"       } // reduce_uchar_8_true\n"
"\n"
"       .entry reduce_uchar_16_true (\n"
"               .param .u64 __cudaparm_reduce_uchar_16_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_16_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_16_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<13>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<15>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       520     0\n"
"$LDWbegin_reduce_uchar_16_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 32;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_16_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_29_16130;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 32;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_16_true_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r7;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_29_14594:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r8, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r8;\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r9, [%rd3+16];\n"
"       cvt.rn.f32.u32  %f4, %r9;\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_16_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_29_14594;\n"
"       bra.uni         $Lt_29_14082;\n"
"$Lt_29_16130:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_29_14082:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 31;\n"
"       setp.gt.u32     %p3, %r3, %r10;\n"
"       @%p3 bra        $Lt_29_15106;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f5, [%rd8+32];\n"
"       add.f32         %f6, %f5, %f1;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f7, [%rd8+16];\n"
"       add.f32         %f8, %f7, %f6;\n"
"       st.volatile.shared.f32  [%rd8+0], %f8;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f9, [%rd8+8];\n"
"       add.f32         %f10, %f9, %f8;\n"
"       st.volatile.shared.f32  [%rd8+0], %f10;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f11, [%rd8+4];\n"
"       add.f32         %f12, %f11, %f10;\n"
"       st.volatile.shared.f32  [%rd8+0], %f12;\n"
"$Lt_29_15106:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r11, 0;\n"
"       setp.ne.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_29_15618;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f13, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_16_true_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f13;\n"
"$Lt_29_15618:\n"
"       .loc    3       523     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_16_true:\n"
"       } // reduce_uchar_16_true\n"
"\n"
"       .entry reduce_uchar_32_true (\n"
"               .param .u64 __cudaparm_reduce_uchar_32_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_32_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_32_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<13>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<17>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       525     0\n"
"$LDWbegin_reduce_uchar_32_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 64;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_32_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_30_15874;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 64;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_32_true_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r7;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_30_14338:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r8, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r8;\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r9, [%rd3+32];\n"
"       cvt.rn.f32.u32  %f4, %r9;\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_32_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_30_14338;\n"
"       bra.uni         $Lt_30_13826;\n"
"$Lt_30_15874:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_30_13826:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 31;\n"
"       setp.gt.u32     %p3, %r3, %r10;\n"
"       @%p3 bra        $Lt_30_14850;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f5, [%rd8+64];\n"
"       add.f32         %f6, %f5, %f1;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f7, [%rd8+32];\n"
"       add.f32         %f8, %f7, %f6;\n"
"       st.volatile.shared.f32  [%rd8+0], %f8;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f9, [%rd8+16];\n"
"       add.f32         %f10, %f9, %f8;\n"
"       st.volatile.shared.f32  [%rd8+0], %f10;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f11, [%rd8+8];\n"
"       add.f32         %f12, %f11, %f10;\n"
"       st.volatile.shared.f32  [%rd8+0], %f12;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f13, [%rd8+4];\n"
"       add.f32         %f14, %f13, %f12;\n"
"       st.volatile.shared.f32  [%rd8+0], %f14;\n"
"$Lt_30_14850:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r11, 0;\n"
"       setp.ne.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_30_15362;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f15, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_32_true_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f15;\n"
"$Lt_30_15362:\n"
"       .loc    3       528     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_32_true:\n"
"       } // reduce_uchar_32_true\n"
"\n"
"       .entry reduce_uchar_64_true (\n"
"               .param .u64 __cudaparm_reduce_uchar_64_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_64_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_64_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<13>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<19>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       530     0\n"
"$LDWbegin_reduce_uchar_64_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 128;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_64_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_31_15618;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 128;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_64_true_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r7;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_31_14082:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r8, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r8;\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r9, [%rd3+64];\n"
"       cvt.rn.f32.u32  %f4, %r9;\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_64_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_31_14082;\n"
"       bra.uni         $Lt_31_13570;\n"
"$Lt_31_15618:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_31_13570:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 31;\n"
"       setp.gt.u32     %p3, %r3, %r10;\n"
"       @%p3 bra        $Lt_31_14594;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f5, [%rd8+128];\n"
"       add.f32         %f6, %f5, %f1;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f7, [%rd8+64];\n"
"       add.f32         %f8, %f7, %f6;\n"
"       st.volatile.shared.f32  [%rd8+0], %f8;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f9, [%rd8+32];\n"
"       add.f32         %f10, %f9, %f8;\n"
"       st.volatile.shared.f32  [%rd8+0], %f10;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f11, [%rd8+16];\n"
"       add.f32         %f12, %f11, %f10;\n"
"       st.volatile.shared.f32  [%rd8+0], %f12;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f13, [%rd8+8];\n"
"       add.f32         %f14, %f13, %f12;\n"
"       st.volatile.shared.f32  [%rd8+0], %f14;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f15, [%rd8+4];\n"
"       add.f32         %f16, %f15, %f14;\n"
"       st.volatile.shared.f32  [%rd8+0], %f16;\n"
"$Lt_31_14594:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r11, 0;\n"
"       setp.ne.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_31_15106;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f17, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_64_true_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f17;\n"
"$Lt_31_15106:\n"
"       .loc    3       533     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_64_true:\n"
"       } // reduce_uchar_64_true\n"
"\n"
"       .entry reduce_uchar_128_true (\n"
"               .param .u64 __cudaparm_reduce_uchar_128_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_128_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_128_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<21>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       535     0\n"
"$LDWbegin_reduce_uchar_128_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 256;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_128_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_32_15874;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 256;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_128_true_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r7;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_32_13826:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r8, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r8;\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r9, [%rd3+128];\n"
"       cvt.rn.f32.u32  %f4, %r9;\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_128_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_32_13826;\n"
"       bra.uni         $Lt_32_13314;\n"
"$Lt_32_15874:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_32_13314:\n"
"       .loc    3       195     0\n"
"       mov.f32         %f5, %f1;\n"
"       mov.f32         %f6, %f5;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 63;\n"
"       setp.gt.u32     %p3, %r3, %r10;\n"
"       @%p3 bra        $Lt_32_14338;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f7, [%rd8+256];\n"
"       add.f32         %f6, %f7, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"$Lt_32_14338:\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 31;\n"
"       setp.gt.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_32_14850;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f8, [%rd8+128];\n"
"       add.f32         %f9, %f8, %f6;\n"
"       st.volatile.shared.f32  [%rd8+0], %f9;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f10, [%rd8+64];\n"
"       add.f32         %f11, %f10, %f9;\n"
"       st.volatile.shared.f32  [%rd8+0], %f11;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f12, [%rd8+32];\n"
"       add.f32         %f13, %f12, %f11;\n"
"       st.volatile.shared.f32  [%rd8+0], %f13;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f14, [%rd8+16];\n"
"       add.f32         %f15, %f14, %f13;\n"
"       st.volatile.shared.f32  [%rd8+0], %f15;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f16, [%rd8+8];\n"
"       add.f32         %f17, %f16, %f15;\n"
"       st.volatile.shared.f32  [%rd8+0], %f17;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f18, [%rd8+4];\n"
"       add.f32         %f6, %f18, %f17;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"$Lt_32_14850:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p5, %r3, %r12;\n"
"       @%p5 bra        $Lt_32_15362;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f19, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_128_true_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f19;\n"
"$Lt_32_15362:\n"
"       .loc    3       538     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_128_true:\n"
"       } // reduce_uchar_128_true\n"
"\n"
"       .entry reduce_uchar_256_true (\n"
"               .param .u64 __cudaparm_reduce_uchar_256_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_256_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_256_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<15>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<22>;\n"
"       .reg .pred %p<8>;\n"
"       .loc    3       540     0\n"
"$LDWbegin_reduce_uchar_256_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 512;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_256_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_33_16130;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 512;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_256_true_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r7;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_33_13570:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r8, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r8;\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r9, [%rd3+256];\n"
"       cvt.rn.f32.u32  %f4, %r9;\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_256_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_33_13570;\n"
"       bra.uni         $Lt_33_13058;\n"
"$Lt_33_16130:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_33_13058:\n"
"       .loc    3       195     0\n"
"       mov.f32         %f5, %f1;\n"
"       mov.f32         %f6, %f5;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 127;\n"
"       setp.gt.u32     %p3, %r3, %r10;\n"
"       @%p3 bra        $Lt_33_14082;\n"
"       .loc    3       76      0\n"
"       ld.volatile.shared.f32  %f7, [%rd8+512];\n"
"       add.f32         %f6, %f7, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"$Lt_33_14082:\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 63;\n"
"       setp.gt.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_33_14594;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f8, [%rd8+256];\n"
"       add.f32         %f6, %f8, %f6;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"$Lt_33_14594:\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 31;\n"
"       setp.gt.u32     %p5, %r3, %r12;\n"
"       @%p5 bra        $Lt_33_15106;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f9, [%rd8+128];\n"
"       add.f32         %f10, %f9, %f6;\n"
"       st.volatile.shared.f32  [%rd8+0], %f10;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f11, [%rd8+64];\n"
"       add.f32         %f12, %f11, %f10;\n"
"       st.volatile.shared.f32  [%rd8+0], %f12;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f13, [%rd8+32];\n"
"       add.f32         %f14, %f13, %f12;\n"
"       st.volatile.shared.f32  [%rd8+0], %f14;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f15, [%rd8+16];\n"
"       add.f32         %f16, %f15, %f14;\n"
"       st.volatile.shared.f32  [%rd8+0], %f16;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f17, [%rd8+8];\n"
"       add.f32         %f18, %f17, %f16;\n"
"       st.volatile.shared.f32  [%rd8+0], %f18;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f19, [%rd8+4];\n"
"       add.f32         %f6, %f19, %f18;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"$Lt_33_15106:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r13, 0;\n"
"       setp.ne.u32     %p6, %r3, %r13;\n"
"       @%p6 bra        $Lt_33_15618;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f20, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_256_true_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f20;\n"
"$Lt_33_15618:\n"
"       .loc    3       543     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_256_true:\n"
"       } // reduce_uchar_256_true\n"
"\n"
"       .entry reduce_uchar_512_true (\n"
"               .param .u64 __cudaparm_reduce_uchar_512_true_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_512_true_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_512_true_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<16>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<23>;\n"
"       .reg .pred %p<9>;\n"
"       .loc    3       545     0\n"
"$LDWbegin_reduce_uchar_512_true:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 1024;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       mov.s32         %r5, %r4;\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_512_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r6;\n"
"       @%p1 bra        $Lt_34_16386;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r7, %rh1, 1024;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_512_true_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r7;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_34_13314:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r8, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r8;\n"
"       add.f32         %f3, %f2, %f1;\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r9, [%rd3+512];\n"
"       cvt.rn.f32.u32  %f4, %r9;\n"
"       add.f32         %f1, %f4, %f3;\n"
"       add.u32         %r5, %r7, %r5;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r6, [__cudaparm_reduce_uchar_512_true_n];\n"
"       .loc    3       191     0\n"
"       setp.lt.u32     %p2, %r5, %r6;\n"
"       @%p2 bra        $Lt_34_13314;\n"
"       bra.uni         $Lt_34_12802;\n"
"$Lt_34_16386:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_34_12802:\n"
"       .loc    3       195     0\n"
"       mov.f32         %f5, %f1;\n"
"       mov.f32         %f6, %f5;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r10, 255;\n"
"       setp.gt.u32     %p3, %r3, %r10;\n"
"       @%p3 bra        $Lt_34_13826;\n"
"       .loc    3       75      0\n"
"       ld.volatile.shared.f32  %f7, [%rd8+1024];\n"
"       add.f32         %f6, %f7, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"$Lt_34_13826:\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 127;\n"
"       setp.gt.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_34_14338;\n"
"       .loc    3       76      0\n"
"       ld.volatile.shared.f32  %f8, [%rd8+512];\n"
"       add.f32         %f6, %f8, %f6;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"$Lt_34_14338:\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 63;\n"
"       setp.gt.u32     %p5, %r3, %r12;\n"
"       @%p5 bra        $Lt_34_14850;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f9, [%rd8+256];\n"
"       add.f32         %f6, %f9, %f6;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"$Lt_34_14850:\n"
"       bar.sync        0;\n"
"       mov.u32         %r13, 31;\n"
"       setp.gt.u32     %p6, %r3, %r13;\n"
"       @%p6 bra        $Lt_34_15362;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f10, [%rd8+128];\n"
"       add.f32         %f11, %f10, %f6;\n"
"       st.volatile.shared.f32  [%rd8+0], %f11;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f12, [%rd8+64];\n"
"       add.f32         %f13, %f12, %f11;\n"
"       st.volatile.shared.f32  [%rd8+0], %f13;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f14, [%rd8+32];\n"
"       add.f32         %f15, %f14, %f13;\n"
"       st.volatile.shared.f32  [%rd8+0], %f15;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f16, [%rd8+16];\n"
"       add.f32         %f17, %f16, %f15;\n"
"       st.volatile.shared.f32  [%rd8+0], %f17;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f18, [%rd8+8];\n"
"       add.f32         %f19, %f18, %f17;\n"
"       st.volatile.shared.f32  [%rd8+0], %f19;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f20, [%rd8+4];\n"
"       add.f32         %f6, %f20, %f19;\n"
"       st.volatile.shared.f32  [%rd8+0], %f6;\n"
"$Lt_34_15362:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r14, 0;\n"
"       setp.ne.u32     %p7, %r3, %r14;\n"
"       @%p7 bra        $Lt_34_15874;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f21, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_512_true_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f21;\n"
"$Lt_34_15874:\n"
"       .loc    3       548     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_512_true:\n"
"       } // reduce_uchar_512_true\n"
"\n"
"       .entry reduce_uchar_1_false (\n"
"               .param .u64 __cudaparm_reduce_uchar_1_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_1_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_1_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<13>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<6>;\n"
"       .reg .pred %p<6>;\n"
"       .loc    3       551     0\n"
"$LDWbegin_reduce_uchar_1_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 2;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_1_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_35_17154;\n"
"       add.u32         %r6, %r4, 1;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_1_false_n];\n"
"       add.u32         %r7, %r5, 1;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 2;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_1_false_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r8;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_35_15618:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r9, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r9;\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_1_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_35_15874;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_35_15618\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r10, [%rd3+1];\n"
"       cvt.rn.f32.u32  %f3, %r10;\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_35_15874:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_35_15618\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_35_15618;\n"
"       bra.uni         $Lt_35_15106;\n"
"$Lt_35_17154:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_35_15106:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       .loc    3       195     0\n"
"       mov.u32         %r11, 0;\n"
"       setp.ne.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_35_16642;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f4, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_1_false_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f4;\n"
"$Lt_35_16642:\n"
"       .loc    3       554     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_1_false:\n"
"       } // reduce_uchar_1_false\n"
"\n"
"       .entry reduce_uchar_2_false (\n"
"               .param .u64 __cudaparm_reduce_uchar_2_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_2_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_2_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<8>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       556     0\n"
"$LDWbegin_reduce_uchar_2_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 4;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_2_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_36_17410;\n"
"       add.u32         %r6, %r4, 2;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_2_false_n];\n"
"       add.u32         %r7, %r5, 2;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 4;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_2_false_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r8;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_36_15362:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r9, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r9;\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_2_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_36_15618;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_36_15362\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r10, [%rd3+2];\n"
"       cvt.rn.f32.u32  %f3, %r10;\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_36_15618:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_36_15362\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_36_15362;\n"
"       bra.uni         $Lt_36_14850;\n"
"$Lt_36_17410:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_36_14850:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 31;\n"
"       setp.gt.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_36_16386;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f4, [%rd8+4];\n"
"       add.f32         %f5, %f4, %f1;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"$Lt_36_16386:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p5, %r3, %r12;\n"
"       @%p5 bra        $Lt_36_16898;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f6, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_2_false_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f6;\n"
"$Lt_36_16898:\n"
"       .loc    3       559     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_2_false:\n"
"       } // reduce_uchar_2_false\n"
"\n"
"       .entry reduce_uchar_4_false (\n"
"               .param .u64 __cudaparm_reduce_uchar_4_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_4_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_4_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<10>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       561     0\n"
"$LDWbegin_reduce_uchar_4_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 8;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_4_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_37_17154;\n"
"       add.u32         %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_4_false_n];\n"
"       add.u32         %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 8;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_4_false_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r8;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_37_15106:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r9, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r9;\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_4_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_37_15362;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_37_15106\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r10, [%rd3+4];\n"
"       cvt.rn.f32.u32  %f3, %r10;\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_37_15362:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_37_15106\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_37_15106;\n"
"       bra.uni         $Lt_37_14594;\n"
"$Lt_37_17154:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_37_14594:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 31;\n"
"       setp.gt.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_37_16130;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f4, [%rd8+8];\n"
"       add.f32         %f5, %f4, %f1;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f6, [%rd8+4];\n"
"       add.f32         %f7, %f6, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f7;\n"
"$Lt_37_16130:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p5, %r3, %r12;\n"
"       @%p5 bra        $Lt_37_16642;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f8, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_4_false_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f8;\n"
"$Lt_37_16642:\n"
"       .loc    3       564     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_4_false:\n"
"       } // reduce_uchar_4_false\n"
"\n"
"       .entry reduce_uchar_8_false (\n"
"               .param .u64 __cudaparm_reduce_uchar_8_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_8_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_8_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<12>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       566     0\n"
"$LDWbegin_reduce_uchar_8_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 16;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_8_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_38_16898;\n"
"       add.u32         %r6, %r4, 8;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_8_false_n];\n"
"       add.u32         %r7, %r5, 8;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 16;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_8_false_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r8;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_38_14850:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r9, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r9;\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_8_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_38_15106;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_38_14850\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r10, [%rd3+8];\n"
"       cvt.rn.f32.u32  %f3, %r10;\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_38_15106:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_38_14850\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_38_14850;\n"
"       bra.uni         $Lt_38_14338;\n"
"$Lt_38_16898:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_38_14338:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 31;\n"
"       setp.gt.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_38_15874;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f4, [%rd8+16];\n"
"       add.f32         %f5, %f4, %f1;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f6, [%rd8+8];\n"
"       add.f32         %f7, %f6, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f7;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f8, [%rd8+4];\n"
"       add.f32         %f9, %f8, %f7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f9;\n"
"$Lt_38_15874:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p5, %r3, %r12;\n"
"       @%p5 bra        $Lt_38_16386;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f10, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_8_false_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f10;\n"
"$Lt_38_16386:\n"
"       .loc    3       569     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_8_false:\n"
"       } // reduce_uchar_8_false\n"
"\n"
"       .entry reduce_uchar_16_false (\n"
"               .param .u64 __cudaparm_reduce_uchar_16_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_16_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_16_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<14>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       571     0\n"
"$LDWbegin_reduce_uchar_16_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 32;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_16_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_39_16642;\n"
"       add.u32         %r6, %r4, 16;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_16_false_n];\n"
"       add.u32         %r7, %r5, 16;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 32;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_16_false_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r8;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_39_14594:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r9, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r9;\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_16_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_39_14850;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_39_14594\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r10, [%rd3+16];\n"
"       cvt.rn.f32.u32  %f3, %r10;\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_39_14850:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_39_14594\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_39_14594;\n"
"       bra.uni         $Lt_39_14082;\n"
"$Lt_39_16642:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_39_14082:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 31;\n"
"       setp.gt.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_39_15618;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f4, [%rd8+32];\n"
"       add.f32         %f5, %f4, %f1;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f6, [%rd8+16];\n"
"       add.f32         %f7, %f6, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f7;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f8, [%rd8+8];\n"
"       add.f32         %f9, %f8, %f7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f9;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f10, [%rd8+4];\n"
"       add.f32         %f11, %f10, %f9;\n"
"       st.volatile.shared.f32  [%rd8+0], %f11;\n"
"$Lt_39_15618:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p5, %r3, %r12;\n"
"       @%p5 bra        $Lt_39_16130;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f12, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_16_false_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f12;\n"
"$Lt_39_16130:\n"
"       .loc    3       574     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_16_false:\n"
"       } // reduce_uchar_16_false\n"
"\n"
"       .entry reduce_uchar_32_false (\n"
"               .param .u64 __cudaparm_reduce_uchar_32_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_32_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_32_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<16>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       576     0\n"
"$LDWbegin_reduce_uchar_32_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 64;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_32_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_40_16386;\n"
"       add.u32         %r6, %r4, 32;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_32_false_n];\n"
"       add.u32         %r7, %r5, 32;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 64;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_32_false_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r8;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_40_14338:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r9, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r9;\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_32_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_40_14594;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_40_14338\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r10, [%rd3+32];\n"
"       cvt.rn.f32.u32  %f3, %r10;\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_40_14594:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_40_14338\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_40_14338;\n"
"       bra.uni         $Lt_40_13826;\n"
"$Lt_40_16386:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_40_13826:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 31;\n"
"       setp.gt.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_40_15362;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f4, [%rd8+64];\n"
"       add.f32         %f5, %f4, %f1;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f6, [%rd8+32];\n"
"       add.f32         %f7, %f6, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f7;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f8, [%rd8+16];\n"
"       add.f32         %f9, %f8, %f7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f9;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f10, [%rd8+8];\n"
"       add.f32         %f11, %f10, %f9;\n"
"       st.volatile.shared.f32  [%rd8+0], %f11;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f12, [%rd8+4];\n"
"       add.f32         %f13, %f12, %f11;\n"
"       st.volatile.shared.f32  [%rd8+0], %f13;\n"
"$Lt_40_15362:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p5, %r3, %r12;\n"
"       @%p5 bra        $Lt_40_15874;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f14, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_32_false_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f14;\n"
"$Lt_40_15874:\n"
"       .loc    3       579     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_32_false:\n"
"       } // reduce_uchar_32_false\n"
"\n"
"       .entry reduce_uchar_64_false (\n"
"               .param .u64 __cudaparm_reduce_uchar_64_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_64_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_64_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<18>;\n"
"       .reg .pred %p<7>;\n"
"       .loc    3       581     0\n"
"$LDWbegin_reduce_uchar_64_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 128;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_64_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_41_16130;\n"
"       add.u32         %r6, %r4, 64;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_64_false_n];\n"
"       add.u32         %r7, %r5, 64;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 128;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_64_false_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r8;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_41_14082:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r9, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r9;\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_64_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_41_14338;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_41_14082\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r10, [%rd3+64];\n"
"       cvt.rn.f32.u32  %f3, %r10;\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_41_14338:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_41_14082\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_41_14082;\n"
"       bra.uni         $Lt_41_13570;\n"
"$Lt_41_16130:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_41_13570:\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f1;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 31;\n"
"       setp.gt.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_41_15106;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f4, [%rd8+128];\n"
"       add.f32         %f5, %f4, %f1;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f6, [%rd8+64];\n"
"       add.f32         %f7, %f6, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f7;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f8, [%rd8+32];\n"
"       add.f32         %f9, %f8, %f7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f9;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f10, [%rd8+16];\n"
"       add.f32         %f11, %f10, %f9;\n"
"       st.volatile.shared.f32  [%rd8+0], %f11;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f12, [%rd8+8];\n"
"       add.f32         %f13, %f12, %f11;\n"
"       st.volatile.shared.f32  [%rd8+0], %f13;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f14, [%rd8+4];\n"
"       add.f32         %f15, %f14, %f13;\n"
"       st.volatile.shared.f32  [%rd8+0], %f15;\n"
"$Lt_41_15106:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p5, %r3, %r12;\n"
"       @%p5 bra        $Lt_41_15618;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f16, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_64_false_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f16;\n"
"$Lt_41_15618:\n"
"       .loc    3       584     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_64_false:\n"
"       } // reduce_uchar_64_false\n"
"\n"
"       .entry reduce_uchar_128_false (\n"
"               .param .u64 __cudaparm_reduce_uchar_128_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_128_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_128_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<15>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<20>;\n"
"       .reg .pred %p<8>;\n"
"       .loc    3       586     0\n"
"$LDWbegin_reduce_uchar_128_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 256;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_128_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_42_16386;\n"
"       add.u32         %r6, %r4, 128;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_128_false_n];\n"
"       add.u32         %r7, %r5, 128;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 256;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_128_false_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r8;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_42_13826:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r9, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r9;\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_128_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_42_14082;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_42_13826\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r10, [%rd3+128];\n"
"       cvt.rn.f32.u32  %f3, %r10;\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_42_14082:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_42_13826\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_42_13826;\n"
"       bra.uni         $Lt_42_13314;\n"
"$Lt_42_16386:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_42_13314:\n"
"       .loc    3       195     0\n"
"       mov.f32         %f4, %f1;\n"
"       mov.f32         %f5, %f4;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f4;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 63;\n"
"       setp.gt.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_42_14850;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f6, [%rd8+256];\n"
"       add.f32         %f5, %f6, %f4;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"$Lt_42_14850:\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 31;\n"
"       setp.gt.u32     %p5, %r3, %r12;\n"
"       @%p5 bra        $Lt_42_15362;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f7, [%rd8+128];\n"
"       add.f32         %f8, %f7, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f8;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f9, [%rd8+64];\n"
"       add.f32         %f10, %f9, %f8;\n"
"       st.volatile.shared.f32  [%rd8+0], %f10;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f11, [%rd8+32];\n"
"       add.f32         %f12, %f11, %f10;\n"
"       st.volatile.shared.f32  [%rd8+0], %f12;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f13, [%rd8+16];\n"
"       add.f32         %f14, %f13, %f12;\n"
"       st.volatile.shared.f32  [%rd8+0], %f14;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f15, [%rd8+8];\n"
"       add.f32         %f16, %f15, %f14;\n"
"       st.volatile.shared.f32  [%rd8+0], %f16;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f17, [%rd8+4];\n"
"       add.f32         %f5, %f17, %f16;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"$Lt_42_15362:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r13, 0;\n"
"       setp.ne.u32     %p6, %r3, %r13;\n"
"       @%p6 bra        $Lt_42_15874;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f18, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_128_false_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f18;\n"
"$Lt_42_15874:\n"
"       .loc    3       589     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_128_false:\n"
"       } // reduce_uchar_128_false\n"
"\n"
"       .entry reduce_uchar_256_false (\n"
"               .param .u64 __cudaparm_reduce_uchar_256_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_256_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_256_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<16>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<21>;\n"
"       .reg .pred %p<9>;\n"
"       .loc    3       591     0\n"
"$LDWbegin_reduce_uchar_256_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 512;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_256_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_43_16642;\n"
"       add.u32         %r6, %r4, 256;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_256_false_n];\n"
"       add.u32         %r7, %r5, 256;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 512;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_256_false_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r8;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_43_13570:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r9, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r9;\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_256_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_43_13826;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_43_13570\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r10, [%rd3+256];\n"
"       cvt.rn.f32.u32  %f3, %r10;\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_43_13826:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_43_13570\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_43_13570;\n"
"       bra.uni         $Lt_43_13058;\n"
"$Lt_43_16642:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_43_13058:\n"
"       .loc    3       195     0\n"
"       mov.f32         %f4, %f1;\n"
"       mov.f32         %f5, %f4;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f4;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 127;\n"
"       setp.gt.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_43_14594;\n"
"       .loc    3       76      0\n"
"       ld.volatile.shared.f32  %f6, [%rd8+512];\n"
"       add.f32         %f5, %f6, %f4;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"$Lt_43_14594:\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 63;\n"
"       setp.gt.u32     %p5, %r3, %r12;\n"
"       @%p5 bra        $Lt_43_15106;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f7, [%rd8+256];\n"
"       add.f32         %f5, %f7, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"$Lt_43_15106:\n"
"       bar.sync        0;\n"
"       mov.u32         %r13, 31;\n"
"       setp.gt.u32     %p6, %r3, %r13;\n"
"       @%p6 bra        $Lt_43_15618;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f8, [%rd8+128];\n"
"       add.f32         %f9, %f8, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f9;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f10, [%rd8+64];\n"
"       add.f32         %f11, %f10, %f9;\n"
"       st.volatile.shared.f32  [%rd8+0], %f11;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f12, [%rd8+32];\n"
"       add.f32         %f13, %f12, %f11;\n"
"       st.volatile.shared.f32  [%rd8+0], %f13;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f14, [%rd8+16];\n"
"       add.f32         %f15, %f14, %f13;\n"
"       st.volatile.shared.f32  [%rd8+0], %f15;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f16, [%rd8+8];\n"
"       add.f32         %f17, %f16, %f15;\n"
"       st.volatile.shared.f32  [%rd8+0], %f17;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f18, [%rd8+4];\n"
"       add.f32         %f5, %f18, %f17;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"$Lt_43_15618:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r14, 0;\n"
"       setp.ne.u32     %p7, %r3, %r14;\n"
"       @%p7 bra        $Lt_43_16130;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f19, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_256_false_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f19;\n"
"$Lt_43_16130:\n"
"       .loc    3       594     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_256_false:\n"
"       } // reduce_uchar_256_false\n"
"\n"
"       .entry reduce_uchar_512_false (\n"
"               .param .u64 __cudaparm_reduce_uchar_512_false_g_idata,\n"
"               .param .u64 __cudaparm_reduce_uchar_512_false_g_odata,\n"
"               .param .u32 __cudaparm_reduce_uchar_512_false_n)\n"
"       {\n"
"       .reg .u16 %rh<3>;\n"
"       .reg .u32 %r<17>;\n"
"       .reg .u64 %rd<14>;\n"
"       .reg .f32 %f<22>;\n"
"       .reg .pred %p<10>;\n"
"       .loc    3       596     0\n"
"$LDWbegin_reduce_uchar_512_false:\n"
"       .loc    3       181     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul.lo.u32      %r2, %r1, 1024;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_512_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_44_16898;\n"
"       add.u32         %r6, %r4, 512;\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_512_false_n];\n"
"       add.u32         %r7, %r5, 512;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 1024;\n"
"       cvt.u64.u32     %rd1, %r4;\n"
"       ld.param.u64    %rd2, [__cudaparm_reduce_uchar_512_false_g_idata];\n"
"       add.u64         %rd3, %rd1, %rd2;\n"
"       cvt.s64.u32     %rd4, %r8;\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_44_13314:\n"
" //<loop> Loop body line 181, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       188     0\n"
"       ld.global.u8    %r9, [%rd3+0];\n"
"       cvt.rn.f32.u32  %f2, %r9;\n"
"       add.f32         %f1, %f2, %f1;\n"
"       .loc    3       181     0\n"
"       ld.param.u32    %r5, [__cudaparm_reduce_uchar_512_false_n];\n"
"       .loc    3       188     0\n"
"       setp.ge.u32     %p2, %r6, %r5;\n"
"       @%p2 bra        $Lt_44_13570;\n"
" //<loop> Part of loop body line 181, head labeled $Lt_44_13314\n"
"       .loc    3       191     0\n"
"       ld.global.u8    %r10, [%rd3+512];\n"
"       cvt.rn.f32.u32  %f3, %r10;\n"
"       add.f32         %f1, %f3, %f1;\n"
"$Lt_44_13570:\n"
" //<loop> Part of loop body line 181, head labeled $Lt_44_13314\n"
"       add.u32         %r6, %r6, %r8;\n"
"       add.u64         %rd3, %rd4, %rd3;\n"
"       setp.lt.u32     %p3, %r6, %r7;\n"
"       @%p3 bra        $Lt_44_13314;\n"
"       bra.uni         $Lt_44_12802;\n"
"$Lt_44_16898:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_44_12802:\n"
"       .loc    3       195     0\n"
"       mov.f32         %f4, %f1;\n"
"       mov.f32         %f5, %f4;\n"
"       .loc    3       71      0\n"
"       mov.u64         %rd5, __smem;\n"
"       cvt.u64.u32     %rd6, %r3;\n"
"       mul.wide.u32    %rd7, %r3, 4;\n"
"       add.u64         %rd8, %rd5, %rd7;\n"
"       st.volatile.shared.f32  [%rd8+0], %f4;\n"
"       .loc    3       72      0\n"
"       bar.sync        0;\n"
"       mov.u32         %r11, 255;\n"
"       setp.gt.u32     %p4, %r3, %r11;\n"
"       @%p4 bra        $Lt_44_14338;\n"
"       .loc    3       75      0\n"
"       ld.volatile.shared.f32  %f6, [%rd8+1024];\n"
"       add.f32         %f5, %f6, %f4;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"$Lt_44_14338:\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 127;\n"
"       setp.gt.u32     %p5, %r3, %r12;\n"
"       @%p5 bra        $Lt_44_14850;\n"
"       .loc    3       76      0\n"
"       ld.volatile.shared.f32  %f7, [%rd8+512];\n"
"       add.f32         %f5, %f7, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"$Lt_44_14850:\n"
"       bar.sync        0;\n"
"       mov.u32         %r13, 63;\n"
"       setp.gt.u32     %p6, %r3, %r13;\n"
"       @%p6 bra        $Lt_44_15362;\n"
"       .loc    3       77      0\n"
"       ld.volatile.shared.f32  %f8, [%rd8+256];\n"
"       add.f32         %f5, %f8, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"$Lt_44_15362:\n"
"       bar.sync        0;\n"
"       mov.u32         %r14, 31;\n"
"       setp.gt.u32     %p7, %r3, %r14;\n"
"       @%p7 bra        $Lt_44_15874;\n"
"       .loc    3       83      0\n"
"       ld.volatile.shared.f32  %f9, [%rd8+128];\n"
"       add.f32         %f10, %f9, %f5;\n"
"       st.volatile.shared.f32  [%rd8+0], %f10;\n"
"       .loc    3       84      0\n"
"       ld.volatile.shared.f32  %f11, [%rd8+64];\n"
"       add.f32         %f12, %f11, %f10;\n"
"       st.volatile.shared.f32  [%rd8+0], %f12;\n"
"       .loc    3       85      0\n"
"       ld.volatile.shared.f32  %f13, [%rd8+32];\n"
"       add.f32         %f14, %f13, %f12;\n"
"       st.volatile.shared.f32  [%rd8+0], %f14;\n"
"       .loc    3       86      0\n"
"       ld.volatile.shared.f32  %f15, [%rd8+16];\n"
"       add.f32         %f16, %f15, %f14;\n"
"       st.volatile.shared.f32  [%rd8+0], %f16;\n"
"       .loc    3       87      0\n"
"       ld.volatile.shared.f32  %f17, [%rd8+8];\n"
"       add.f32         %f18, %f17, %f16;\n"
"       st.volatile.shared.f32  [%rd8+0], %f18;\n"
"       .loc    3       88      0\n"
"       ld.volatile.shared.f32  %f19, [%rd8+4];\n"
"       add.f32         %f5, %f19, %f18;\n"
"       st.volatile.shared.f32  [%rd8+0], %f5;\n"
"$Lt_44_15874:\n"
"       .loc    3       195     0\n"
"       mov.u32         %r15, 0;\n"
"       setp.ne.u32     %p8, %r3, %r15;\n"
"       @%p8 bra        $Lt_44_16386;\n"
"       .loc    3       199     0\n"
"       ld.shared.f32   %f20, [__smem+0];\n"
"       ld.param.u64    %rd9, [__cudaparm_reduce_uchar_512_false_g_odata];\n"
"       cvt.u64.u32     %rd10, %r1;\n"
"       mul.wide.u32    %rd11, %r1, 4;\n"
"       add.u64         %rd12, %rd9, %rd11;\n"
"       st.global.f32   [%rd12+0], %f20;\n"
"$Lt_44_16386:\n"
"       .loc    3       599     0\n"
"       exit;\n"
"$LDWend_reduce_uchar_512_false:\n"
"       } // reduce_uchar_512_false\n"
"\n"
"       .entry packed_float_reduce_1_false_false (\n"
"               .param .u64 __cudaparm_packed_float_reduce_1_false_false_g_idata,\n"
"               .param .u64 __cudaparm_packed_float_reduce_1_false_false_g_odata,\n"
"               .param .u32 __cudaparm_packed_float_reduce_1_false_false_n)\n"
"       {\n"
"       .reg .u16 %rh<7>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<13>;\n"
"       .reg .f32 %f<4>;\n"
"       .reg .pred %p<5>;\n"
"       .local .align 4 .b8 __cuda_local_var_22761_19_pack_0[4];\n"
"       .loc    3       722     0\n"
"$LDWbegin_packed_float_reduce_1_false_false:\n"
"       .loc    3       637     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 2;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_1_false_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_45_18178;\n"
"       mul.lo.u32      %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_1_false_false_n];\n"
"       mul.lo.u32      %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 8;\n"
"       add.u32         %r9, %r6, 4;\n"
"       add.u32         %r10, %r7, 4;\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_1_false_false_g_idata];\n"
"$Lt_45_17154:\n"
" //<loop> Loop body line 637, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       677     0\n"
"       cvt.u8.u32      %r11, %r9;\n"
"       cvt.u64.u32     %rd2, %r11;\n"
"       .loc    3       637     0\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_1_false_false_g_idata];\n"
"       .loc    3       677     0\n"
"       add.u64         %rd3, %rd2, %rd1;\n"
"       ld.global.u8    %rh2, [%rd3+0];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+3], %rh2;\n"
"       .loc    3       678     0\n"
"       ld.global.u8    %rh3, [%rd3+1];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+2], %rh3;\n"
"       .loc    3       679     0\n"
"       ld.global.u8    %rh4, [%rd3+2];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+1], %rh4;\n"
"       .loc    3       680     0\n"
"       ld.global.u8    %rh5, [%rd3+3];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+0], %rh5;\n"
"       add.u32         %r9, %r8, %r9;\n"
"       setp.lt.u32     %p2, %r9, %r10;\n"
"       @%p2 bra        $Lt_45_17154;\n"
"       ld.local.f32    %f1, [__cuda_local_var_22761_19_pack_0+0];\n"
"       bra.uni         $Lt_45_16642;\n"
"$Lt_45_18178:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_45_16642:\n"
"       .loc    3       692     0\n"
"       mov.u64         %rd4, __smem;\n"
"       cvt.u64.u32     %rd5, %r3;\n"
"       mul.wide.u32    %rd6, %r3, 4;\n"
"       add.u64         %rd7, %rd4, %rd6;\n"
"       st.shared.f32   [%rd7+0], %f1;\n"
"       .loc    3       693     0\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p3, %r3, %r12;\n"
"       @%p3 bra        $Lt_45_17666;\n"
"       .loc    3       719     0\n"
"       ld.shared.f32   %f2, [__smem+0];\n"
"       ld.param.u64    %rd8, [__cudaparm_packed_float_reduce_1_false_false_g_odata];\n"
"       cvt.u64.u32     %rd9, %r1;\n"
"       mul.wide.u32    %rd10, %r1, 4;\n"
"       add.u64         %rd11, %rd8, %rd10;\n"
"       st.global.f32   [%rd11+0], %f2;\n"
"$Lt_45_17666:\n"
"       .loc    3       723     0\n"
"       exit;\n"
"$LDWend_packed_float_reduce_1_false_false:\n"
"       } // packed_float_reduce_1_false_false\n"
"\n"
"       .entry packed_float_reduce_1_false_true (\n"
"               .param .u64 __cudaparm_packed_float_reduce_1_false_true_g_idata,\n"
"               .param .u64 __cudaparm_packed_float_reduce_1_false_true_g_odata,\n"
"               .param .u32 __cudaparm_packed_float_reduce_1_false_true_n)\n"
"       {\n"
"       .reg .u16 %rh<7>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<13>;\n"
"       .reg .f32 %f<4>;\n"
"       .reg .pred %p<5>;\n"
"       .local .align 4 .b8 __cuda_local_var_22761_19_pack_0[4];\n"
"       .loc    3       724     0\n"
"$LDWbegin_packed_float_reduce_1_false_true:\n"
"       .loc    3       637     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 2;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_1_false_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_46_18178;\n"
"       mul.lo.u32      %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_1_false_true_n];\n"
"       mul.lo.u32      %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 8;\n"
"       add.u32         %r9, %r6, 4;\n"
"       add.u32         %r10, %r7, 4;\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_1_false_true_g_idata];\n"
"$Lt_46_17154:\n"
" //<loop> Loop body line 637, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       677     0\n"
"       cvt.u8.u32      %r11, %r9;\n"
"       cvt.u64.u32     %rd2, %r11;\n"
"       .loc    3       637     0\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_1_false_true_g_idata];\n"
"       .loc    3       677     0\n"
"       add.u64         %rd3, %rd2, %rd1;\n"
"       ld.global.u8    %rh2, [%rd3+0];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+3], %rh2;\n"
"       .loc    3       678     0\n"
"       ld.global.u8    %rh3, [%rd3+1];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+2], %rh3;\n"
"       .loc    3       679     0\n"
"       ld.global.u8    %rh4, [%rd3+2];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+1], %rh4;\n"
"       .loc    3       680     0\n"
"       ld.global.u8    %rh5, [%rd3+3];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+0], %rh5;\n"
"       add.u32         %r9, %r8, %r9;\n"
"       setp.lt.u32     %p2, %r9, %r10;\n"
"       @%p2 bra        $Lt_46_17154;\n"
"       ld.local.f32    %f1, [__cuda_local_var_22761_19_pack_0+0];\n"
"       bra.uni         $Lt_46_16642;\n"
"$Lt_46_18178:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_46_16642:\n"
"       .loc    3       692     0\n"
"       mov.u64         %rd4, __smem;\n"
"       cvt.u64.u32     %rd5, %r3;\n"
"       mul.wide.u32    %rd6, %r3, 4;\n"
"       add.u64         %rd7, %rd4, %rd6;\n"
"       st.shared.f32   [%rd7+0], %f1;\n"
"       .loc    3       693     0\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p3, %r3, %r12;\n"
"       @%p3 bra        $Lt_46_17666;\n"
"       .loc    3       719     0\n"
"       ld.shared.f32   %f2, [__smem+0];\n"
"       ld.param.u64    %rd8, [__cudaparm_packed_float_reduce_1_false_true_g_odata];\n"
"       cvt.u64.u32     %rd9, %r1;\n"
"       mul.wide.u32    %rd10, %r1, 4;\n"
"       add.u64         %rd11, %rd8, %rd10;\n"
"       st.global.f32   [%rd11+0], %f2;\n"
"$Lt_46_17666:\n"
"       .loc    3       725     0\n"
"       exit;\n"
"$LDWend_packed_float_reduce_1_false_true:\n"
"       } // packed_float_reduce_1_false_true\n"
"\n"
"       .entry packed_float_reduce_1_true_false (\n"
"               .param .u64 __cudaparm_packed_float_reduce_1_true_false_g_idata,\n"
"               .param .u64 __cudaparm_packed_float_reduce_1_true_false_g_odata,\n"
"               .param .u32 __cudaparm_packed_float_reduce_1_true_false_n)\n"
"       {\n"
"       .reg .u16 %rh<7>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<13>;\n"
"       .reg .f32 %f<4>;\n"
"       .reg .pred %p<5>;\n"
"       .local .align 4 .b8 __cuda_local_var_22761_19_pack_0[4];\n"
"       .loc    3       726     0\n"
"$LDWbegin_packed_float_reduce_1_true_false:\n"
"       .loc    3       637     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 2;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_1_true_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_47_18178;\n"
"       mul.lo.u32      %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_1_true_false_n];\n"
"       mul.lo.u32      %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 8;\n"
"       add.u32         %r9, %r6, 4;\n"
"       add.u32         %r10, %r7, 4;\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_1_true_false_g_idata];\n"
"$Lt_47_17154:\n"
" //<loop> Loop body line 637, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       677     0\n"
"       cvt.u8.u32      %r11, %r9;\n"
"       cvt.u64.u32     %rd2, %r11;\n"
"       .loc    3       637     0\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_1_true_false_g_idata];\n"
"       .loc    3       677     0\n"
"       add.u64         %rd3, %rd2, %rd1;\n"
"       ld.global.u8    %rh2, [%rd3+0];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+3], %rh2;\n"
"       .loc    3       678     0\n"
"       ld.global.u8    %rh3, [%rd3+1];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+2], %rh3;\n"
"       .loc    3       679     0\n"
"       ld.global.u8    %rh4, [%rd3+2];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+1], %rh4;\n"
"       .loc    3       680     0\n"
"       ld.global.u8    %rh5, [%rd3+3];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+0], %rh5;\n"
"       add.u32         %r9, %r8, %r9;\n"
"       setp.lt.u32     %p2, %r9, %r10;\n"
"       @%p2 bra        $Lt_47_17154;\n"
"       ld.local.f32    %f1, [__cuda_local_var_22761_19_pack_0+0];\n"
"       bra.uni         $Lt_47_16642;\n"
"$Lt_47_18178:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_47_16642:\n"
"       .loc    3       692     0\n"
"       mov.u64         %rd4, __smem;\n"
"       cvt.u64.u32     %rd5, %r3;\n"
"       mul.wide.u32    %rd6, %r3, 4;\n"
"       add.u64         %rd7, %rd4, %rd6;\n"
"       st.shared.f32   [%rd7+0], %f1;\n"
"       .loc    3       693     0\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p3, %r3, %r12;\n"
"       @%p3 bra        $Lt_47_17666;\n"
"       .loc    3       719     0\n"
"       ld.shared.f32   %f2, [__smem+0];\n"
"       ld.param.u64    %rd8, [__cudaparm_packed_float_reduce_1_true_false_g_odata];\n"
"       cvt.u64.u32     %rd9, %r1;\n"
"       mul.wide.u32    %rd10, %r1, 4;\n"
"       add.u64         %rd11, %rd8, %rd10;\n"
"       st.global.f32   [%rd11+0], %f2;\n"
"$Lt_47_17666:\n"
"       .loc    3       727     0\n"
"       exit;\n"
"$LDWend_packed_float_reduce_1_true_false:\n"
"       } // packed_float_reduce_1_true_false\n"
"\n"
"       .entry packed_float_reduce_1_true_true (\n"
"               .param .u64 __cudaparm_packed_float_reduce_1_true_true_g_idata,\n"
"               .param .u64 __cudaparm_packed_float_reduce_1_true_true_g_odata,\n"
"               .param .u32 __cudaparm_packed_float_reduce_1_true_true_n)\n"
"       {\n"
"       .reg .u16 %rh<7>;\n"
"       .reg .u32 %r<14>;\n"
"       .reg .u64 %rd<13>;\n"
"       .reg .f32 %f<4>;\n"
"       .reg .pred %p<5>;\n"
"       .local .align 4 .b8 __cuda_local_var_22761_19_pack_0[4];\n"
"       .loc    3       728     0\n"
"$LDWbegin_packed_float_reduce_1_true_true:\n"
"       .loc    3       637     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 2;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_1_true_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_48_18178;\n"
"       mul.lo.u32      %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_1_true_true_n];\n"
"       mul.lo.u32      %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 8;\n"
"       add.u32         %r9, %r6, 4;\n"
"       add.u32         %r10, %r7, 4;\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_1_true_true_g_idata];\n"
"$Lt_48_17154:\n"
" //<loop> Loop body line 637, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       677     0\n"
"       cvt.u8.u32      %r11, %r9;\n"
"       cvt.u64.u32     %rd2, %r11;\n"
"       .loc    3       637     0\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_1_true_true_g_idata];\n"
"       .loc    3       677     0\n"
"       add.u64         %rd3, %rd2, %rd1;\n"
"       ld.global.u8    %rh2, [%rd3+0];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+3], %rh2;\n"
"       .loc    3       678     0\n"
"       ld.global.u8    %rh3, [%rd3+1];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+2], %rh3;\n"
"       .loc    3       679     0\n"
"       ld.global.u8    %rh4, [%rd3+2];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+1], %rh4;\n"
"       .loc    3       680     0\n"
"       ld.global.u8    %rh5, [%rd3+3];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+0], %rh5;\n"
"       add.u32         %r9, %r8, %r9;\n"
"       setp.lt.u32     %p2, %r9, %r10;\n"
"       @%p2 bra        $Lt_48_17154;\n"
"       ld.local.f32    %f1, [__cuda_local_var_22761_19_pack_0+0];\n"
"       bra.uni         $Lt_48_16642;\n"
"$Lt_48_18178:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_48_16642:\n"
"       .loc    3       692     0\n"
"       mov.u64         %rd4, __smem;\n"
"       cvt.u64.u32     %rd5, %r3;\n"
"       mul.wide.u32    %rd6, %r3, 4;\n"
"       add.u64         %rd7, %rd4, %rd6;\n"
"       st.shared.f32   [%rd7+0], %f1;\n"
"       .loc    3       693     0\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 0;\n"
"       setp.ne.u32     %p3, %r3, %r12;\n"
"       @%p3 bra        $Lt_48_17666;\n"
"       .loc    3       719     0\n"
"       ld.shared.f32   %f2, [__smem+0];\n"
"       ld.param.u64    %rd8, [__cudaparm_packed_float_reduce_1_true_true_g_odata];\n"
"       cvt.u64.u32     %rd9, %r1;\n"
"       mul.wide.u32    %rd10, %r1, 4;\n"
"       add.u64         %rd11, %rd8, %rd10;\n"
"       st.global.f32   [%rd11+0], %f2;\n"
"$Lt_48_17666:\n"
"       .loc    3       729     0\n"
"       exit;\n"
"$LDWend_packed_float_reduce_1_true_true:\n"
"       } // packed_float_reduce_1_true_true\n"
"\n"
"       .entry packed_float_reduce_2_false_false (\n"
"               .param .u64 __cudaparm_packed_float_reduce_2_false_false_g_idata,\n"
"               .param .u64 __cudaparm_packed_float_reduce_2_false_false_g_odata,\n"
"               .param .u32 __cudaparm_packed_float_reduce_2_false_false_n)\n"
"       {\n"
"       .reg .u16 %rh<7>;\n"
"       .reg .u32 %r<15>;\n"
"       .reg .u64 %rd<13>;\n"
"       .reg .f32 %f<5>;\n"
"       .reg .pred %p<6>;\n"
"       .local .align 4 .b8 __cuda_local_var_22761_19_pack_0[4];\n"
"       .loc    3       731     0\n"
"$LDWbegin_packed_float_reduce_2_false_false:\n"
"       .loc    3       637     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 4;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_2_false_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_49_18434;\n"
"       mul.lo.u32      %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_2_false_false_n];\n"
"       mul.lo.u32      %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 16;\n"
"       add.u32         %r9, %r6, 8;\n"
"       add.u32         %r10, %r7, 8;\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_2_false_false_g_idata];\n"
"$Lt_49_16898:\n"
" //<loop> Loop body line 637, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       677     0\n"
"       cvt.u8.u32      %r11, %r9;\n"
"       cvt.u64.u32     %rd2, %r11;\n"
"       .loc    3       637     0\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_2_false_false_g_idata];\n"
"       .loc    3       677     0\n"
"       add.u64         %rd3, %rd2, %rd1;\n"
"       ld.global.u8    %rh2, [%rd3+0];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+3], %rh2;\n"
"       .loc    3       678     0\n"
"       ld.global.u8    %rh3, [%rd3+1];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+2], %rh3;\n"
"       .loc    3       679     0\n"
"       ld.global.u8    %rh4, [%rd3+2];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+1], %rh4;\n"
"       .loc    3       680     0\n"
"       ld.global.u8    %rh5, [%rd3+3];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+0], %rh5;\n"
"       add.u32         %r9, %r8, %r9;\n"
"       setp.lt.u32     %p2, %r9, %r10;\n"
"       @%p2 bra        $Lt_49_16898;\n"
"       ld.local.f32    %f1, [__cuda_local_var_22761_19_pack_0+0];\n"
"       bra.uni         $Lt_49_16386;\n"
"$Lt_49_18434:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_49_16386:\n"
"       .loc    3       692     0\n"
"       mov.u64         %rd4, __smem;\n"
"       cvt.u64.u32     %rd5, %r3;\n"
"       mul.wide.u32    %rd6, %r3, 4;\n"
"       add.u64         %rd7, %rd4, %rd6;\n"
"       st.shared.f32   [%rd7+0], %f1;\n"
"       .loc    3       693     0\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 31;\n"
"       setp.gt.u32     %p3, %r3, %r12;\n"
"       @%p3 bra        $Lt_49_17410;\n"
"       .loc    3       714     0\n"
"       ld.volatile.shared.f32  %f2, [%rd7+4];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       st.volatile.shared.f32  [%rd7+0], %f1;\n"
"$Lt_49_17410:\n"
"       mov.u32         %r13, 0;\n"
"       setp.ne.u32     %p4, %r3, %r13;\n"
"       @%p4 bra        $Lt_49_17922;\n"
"       .loc    3       719     0\n"
"       ld.shared.f32   %f3, [__smem+0];\n"
"       ld.param.u64    %rd8, [__cudaparm_packed_float_reduce_2_false_false_g_odata];\n"
"       cvt.u64.u32     %rd9, %r1;\n"
"       mul.wide.u32    %rd10, %r1, 4;\n"
"       add.u64         %rd11, %rd8, %rd10;\n"
"       st.global.f32   [%rd11+0], %f3;\n"
"$Lt_49_17922:\n"
"       .loc    3       732     0\n"
"       exit;\n"
"$LDWend_packed_float_reduce_2_false_false:\n"
"       } // packed_float_reduce_2_false_false\n"
"\n"
"       .entry packed_float_reduce_2_false_true (\n"
"               .param .u64 __cudaparm_packed_float_reduce_2_false_true_g_idata,\n"
"               .param .u64 __cudaparm_packed_float_reduce_2_false_true_g_odata,\n"
"               .param .u32 __cudaparm_packed_float_reduce_2_false_true_n)\n"
"       {\n"
"       .reg .u16 %rh<7>;\n"
"       .reg .u32 %r<15>;\n"
"       .reg .u64 %rd<13>;\n"
"       .reg .f32 %f<5>;\n"
"       .reg .pred %p<6>;\n"
"       .local .align 4 .b8 __cuda_local_var_22761_19_pack_0[4];\n"
"       .loc    3       733     0\n"
"$LDWbegin_packed_float_reduce_2_false_true:\n"
"       .loc    3       637     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 4;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_2_false_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_50_18434;\n"
"       mul.lo.u32      %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_2_false_true_n];\n"
"       mul.lo.u32      %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 16;\n"
"       add.u32         %r9, %r6, 8;\n"
"       add.u32         %r10, %r7, 8;\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_2_false_true_g_idata];\n"
"$Lt_50_16898:\n"
" //<loop> Loop body line 637, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       677     0\n"
"       cvt.u8.u32      %r11, %r9;\n"
"       cvt.u64.u32     %rd2, %r11;\n"
"       .loc    3       637     0\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_2_false_true_g_idata];\n"
"       .loc    3       677     0\n"
"       add.u64         %rd3, %rd2, %rd1;\n"
"       ld.global.u8    %rh2, [%rd3+0];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+3], %rh2;\n"
"       .loc    3       678     0\n"
"       ld.global.u8    %rh3, [%rd3+1];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+2], %rh3;\n"
"       .loc    3       679     0\n"
"       ld.global.u8    %rh4, [%rd3+2];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+1], %rh4;\n"
"       .loc    3       680     0\n"
"       ld.global.u8    %rh5, [%rd3+3];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+0], %rh5;\n"
"       add.u32         %r9, %r8, %r9;\n"
"       setp.lt.u32     %p2, %r9, %r10;\n"
"       @%p2 bra        $Lt_50_16898;\n"
"       ld.local.f32    %f1, [__cuda_local_var_22761_19_pack_0+0];\n"
"       bra.uni         $Lt_50_16386;\n"
"$Lt_50_18434:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_50_16386:\n"
"       .loc    3       692     0\n"
"       mov.u64         %rd4, __smem;\n"
"       cvt.u64.u32     %rd5, %r3;\n"
"       mul.wide.u32    %rd6, %r3, 4;\n"
"       add.u64         %rd7, %rd4, %rd6;\n"
"       st.shared.f32   [%rd7+0], %f1;\n"
"       .loc    3       693     0\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 31;\n"
"       setp.gt.u32     %p3, %r3, %r12;\n"
"       @%p3 bra        $Lt_50_17410;\n"
"       .loc    3       714     0\n"
"       ld.volatile.shared.f32  %f2, [%rd7+4];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       st.volatile.shared.f32  [%rd7+0], %f1;\n"
"$Lt_50_17410:\n"
"       mov.u32         %r13, 0;\n"
"       setp.ne.u32     %p4, %r3, %r13;\n"
"       @%p4 bra        $Lt_50_17922;\n"
"       .loc    3       719     0\n"
"       ld.shared.f32   %f3, [__smem+0];\n"
"       ld.param.u64    %rd8, [__cudaparm_packed_float_reduce_2_false_true_g_odata];\n"
"       cvt.u64.u32     %rd9, %r1;\n"
"       mul.wide.u32    %rd10, %r1, 4;\n"
"       add.u64         %rd11, %rd8, %rd10;\n"
"       st.global.f32   [%rd11+0], %f3;\n"
"$Lt_50_17922:\n"
"       .loc    3       734     0\n"
"       exit;\n"
"$LDWend_packed_float_reduce_2_false_true:\n"
"       } // packed_float_reduce_2_false_true\n"
"\n"
"       .entry packed_float_reduce_2_true_false (\n"
"               .param .u64 __cudaparm_packed_float_reduce_2_true_false_g_idata,\n"
"               .param .u64 __cudaparm_packed_float_reduce_2_true_false_g_odata,\n"
"               .param .u32 __cudaparm_packed_float_reduce_2_true_false_n)\n"
"       {\n"
"       .reg .u16 %rh<7>;\n"
"       .reg .u32 %r<15>;\n"
"       .reg .u64 %rd<13>;\n"
"       .reg .f32 %f<5>;\n"
"       .reg .pred %p<6>;\n"
"       .local .align 4 .b8 __cuda_local_var_22761_19_pack_0[4];\n"
"       .loc    3       735     0\n"
"$LDWbegin_packed_float_reduce_2_true_false:\n"
"       .loc    3       637     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 4;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_2_true_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_51_18434;\n"
"       mul.lo.u32      %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_2_true_false_n];\n"
"       mul.lo.u32      %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 16;\n"
"       add.u32         %r9, %r6, 8;\n"
"       add.u32         %r10, %r7, 8;\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_2_true_false_g_idata];\n"
"$Lt_51_16898:\n"
" //<loop> Loop body line 637, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       677     0\n"
"       cvt.u8.u32      %r11, %r9;\n"
"       cvt.u64.u32     %rd2, %r11;\n"
"       .loc    3       637     0\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_2_true_false_g_idata];\n"
"       .loc    3       677     0\n"
"       add.u64         %rd3, %rd2, %rd1;\n"
"       ld.global.u8    %rh2, [%rd3+0];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+3], %rh2;\n"
"       .loc    3       678     0\n"
"       ld.global.u8    %rh3, [%rd3+1];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+2], %rh3;\n"
"       .loc    3       679     0\n"
"       ld.global.u8    %rh4, [%rd3+2];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+1], %rh4;\n"
"       .loc    3       680     0\n"
"       ld.global.u8    %rh5, [%rd3+3];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+0], %rh5;\n"
"       add.u32         %r9, %r8, %r9;\n"
"       setp.lt.u32     %p2, %r9, %r10;\n"
"       @%p2 bra        $Lt_51_16898;\n"
"       ld.local.f32    %f1, [__cuda_local_var_22761_19_pack_0+0];\n"
"       bra.uni         $Lt_51_16386;\n"
"$Lt_51_18434:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_51_16386:\n"
"       .loc    3       692     0\n"
"       mov.u64         %rd4, __smem;\n"
"       cvt.u64.u32     %rd5, %r3;\n"
"       mul.wide.u32    %rd6, %r3, 4;\n"
"       add.u64         %rd7, %rd4, %rd6;\n"
"       st.shared.f32   [%rd7+0], %f1;\n"
"       .loc    3       693     0\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 31;\n"
"       setp.gt.u32     %p3, %r3, %r12;\n"
"       @%p3 bra        $Lt_51_17410;\n"
"       .loc    3       714     0\n"
"       ld.volatile.shared.f32  %f2, [%rd7+4];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       st.volatile.shared.f32  [%rd7+0], %f1;\n"
"$Lt_51_17410:\n"
"       mov.u32         %r13, 0;\n"
"       setp.ne.u32     %p4, %r3, %r13;\n"
"       @%p4 bra        $Lt_51_17922;\n"
"       .loc    3       719     0\n"
"       ld.shared.f32   %f3, [__smem+0];\n"
"       ld.param.u64    %rd8, [__cudaparm_packed_float_reduce_2_true_false_g_odata];\n"
"       cvt.u64.u32     %rd9, %r1;\n"
"       mul.wide.u32    %rd10, %r1, 4;\n"
"       add.u64         %rd11, %rd8, %rd10;\n"
"       st.global.f32   [%rd11+0], %f3;\n"
"$Lt_51_17922:\n"
"       .loc    3       736     0\n"
"       exit;\n"
"$LDWend_packed_float_reduce_2_true_false:\n"
"       } // packed_float_reduce_2_true_false\n"
"\n"
"       .entry packed_float_reduce_2_true_true (\n"
"               .param .u64 __cudaparm_packed_float_reduce_2_true_true_g_idata,\n"
"               .param .u64 __cudaparm_packed_float_reduce_2_true_true_g_odata,\n"
"               .param .u32 __cudaparm_packed_float_reduce_2_true_true_n)\n"
"       {\n"
"       .reg .u16 %rh<7>;\n"
"       .reg .u32 %r<15>;\n"
"       .reg .u64 %rd<13>;\n"
"       .reg .f32 %f<5>;\n"
"       .reg .pred %p<6>;\n"
"       .local .align 4 .b8 __cuda_local_var_22761_19_pack_0[4];\n"
"       .loc    3       737     0\n"
"$LDWbegin_packed_float_reduce_2_true_true:\n"
"       .loc    3       637     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 4;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_2_true_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_52_18434;\n"
"       mul.lo.u32      %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_2_true_true_n];\n"
"       mul.lo.u32      %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 16;\n"
"       add.u32         %r9, %r6, 8;\n"
"       add.u32         %r10, %r7, 8;\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_2_true_true_g_idata];\n"
"$Lt_52_16898:\n"
" //<loop> Loop body line 637, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       677     0\n"
"       cvt.u8.u32      %r11, %r9;\n"
"       cvt.u64.u32     %rd2, %r11;\n"
"       .loc    3       637     0\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_2_true_true_g_idata];\n"
"       .loc    3       677     0\n"
"       add.u64         %rd3, %rd2, %rd1;\n"
"       ld.global.u8    %rh2, [%rd3+0];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+3], %rh2;\n"
"       .loc    3       678     0\n"
"       ld.global.u8    %rh3, [%rd3+1];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+2], %rh3;\n"
"       .loc    3       679     0\n"
"       ld.global.u8    %rh4, [%rd3+2];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+1], %rh4;\n"
"       .loc    3       680     0\n"
"       ld.global.u8    %rh5, [%rd3+3];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+0], %rh5;\n"
"       add.u32         %r9, %r8, %r9;\n"
"       setp.lt.u32     %p2, %r9, %r10;\n"
"       @%p2 bra        $Lt_52_16898;\n"
"       ld.local.f32    %f1, [__cuda_local_var_22761_19_pack_0+0];\n"
"       bra.uni         $Lt_52_16386;\n"
"$Lt_52_18434:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_52_16386:\n"
"       .loc    3       692     0\n"
"       mov.u64         %rd4, __smem;\n"
"       cvt.u64.u32     %rd5, %r3;\n"
"       mul.wide.u32    %rd6, %r3, 4;\n"
"       add.u64         %rd7, %rd4, %rd6;\n"
"       st.shared.f32   [%rd7+0], %f1;\n"
"       .loc    3       693     0\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 31;\n"
"       setp.gt.u32     %p3, %r3, %r12;\n"
"       @%p3 bra        $Lt_52_17410;\n"
"       .loc    3       714     0\n"
"       ld.volatile.shared.f32  %f2, [%rd7+4];\n"
"       add.f32         %f1, %f2, %f1;\n"
"       st.volatile.shared.f32  [%rd7+0], %f1;\n"
"$Lt_52_17410:\n"
"       mov.u32         %r13, 0;\n"
"       setp.ne.u32     %p4, %r3, %r13;\n"
"       @%p4 bra        $Lt_52_17922;\n"
"       .loc    3       719     0\n"
"       ld.shared.f32   %f3, [__smem+0];\n"
"       ld.param.u64    %rd8, [__cudaparm_packed_float_reduce_2_true_true_g_odata];\n"
"       cvt.u64.u32     %rd9, %r1;\n"
"       mul.wide.u32    %rd10, %r1, 4;\n"
"       add.u64         %rd11, %rd8, %rd10;\n"
"       st.global.f32   [%rd11+0], %f3;\n"
"$Lt_52_17922:\n"
"       .loc    3       738     0\n"
"       exit;\n"
"$LDWend_packed_float_reduce_2_true_true:\n"
"       } // packed_float_reduce_2_true_true\n"
"\n"
"       .entry packed_float_reduce_4_false_false (\n"
"               .param .u64 __cudaparm_packed_float_reduce_4_false_false_g_idata,\n"
"               .param .u64 __cudaparm_packed_float_reduce_4_false_false_g_odata,\n"
"               .param .u32 __cudaparm_packed_float_reduce_4_false_false_n)\n"
"       {\n"
"       .reg .u16 %rh<7>;\n"
"       .reg .u32 %r<15>;\n"
"       .reg .u64 %rd<13>;\n"
"       .reg .f32 %f<7>;\n"
"       .reg .pred %p<6>;\n"
"       .local .align 4 .b8 __cuda_local_var_22761_19_pack_0[4];\n"
"       .loc    3       740     0\n"
"$LDWbegin_packed_float_reduce_4_false_false:\n"
"       .loc    3       637     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 8;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_4_false_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_53_18178;\n"
"       mul.lo.u32      %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_4_false_false_n];\n"
"       mul.lo.u32      %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 32;\n"
"       add.u32         %r9, %r6, 16;\n"
"       add.u32         %r10, %r7, 16;\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_4_false_false_g_idata];\n"
"$Lt_53_16642:\n"
" //<loop> Loop body line 637, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       677     0\n"
"       cvt.u8.u32      %r11, %r9;\n"
"       cvt.u64.u32     %rd2, %r11;\n"
"       .loc    3       637     0\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_4_false_false_g_idata];\n"
"       .loc    3       677     0\n"
"       add.u64         %rd3, %rd2, %rd1;\n"
"       ld.global.u8    %rh2, [%rd3+0];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+3], %rh2;\n"
"       .loc    3       678     0\n"
"       ld.global.u8    %rh3, [%rd3+1];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+2], %rh3;\n"
"       .loc    3       679     0\n"
"       ld.global.u8    %rh4, [%rd3+2];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+1], %rh4;\n"
"       .loc    3       680     0\n"
"       ld.global.u8    %rh5, [%rd3+3];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+0], %rh5;\n"
"       add.u32         %r9, %r8, %r9;\n"
"       setp.lt.u32     %p2, %r9, %r10;\n"
"       @%p2 bra        $Lt_53_16642;\n"
"       ld.local.f32    %f1, [__cuda_local_var_22761_19_pack_0+0];\n"
"       bra.uni         $Lt_53_16130;\n"
"$Lt_53_18178:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_53_16130:\n"
"       .loc    3       692     0\n"
"       mov.u64         %rd4, __smem;\n"
"       cvt.u64.u32     %rd5, %r3;\n"
"       mul.wide.u32    %rd6, %r3, 4;\n"
"       add.u64         %rd7, %rd4, %rd6;\n"
"       st.shared.f32   [%rd7+0], %f1;\n"
"       .loc    3       693     0\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 31;\n"
"       setp.gt.u32     %p3, %r3, %r12;\n"
"       @%p3 bra        $Lt_53_17154;\n"
"       .loc    3       713     0\n"
"       ld.volatile.shared.f32  %f2, [%rd7+8];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       st.volatile.shared.f32  [%rd7+0], %f3;\n"
"       .loc    3       714     0\n"
"       ld.volatile.shared.f32  %f4, [%rd7+4];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       st.volatile.shared.f32  [%rd7+0], %f1;\n"
"$Lt_53_17154:\n"
"       mov.u32         %r13, 0;\n"
"       setp.ne.u32     %p4, %r3, %r13;\n"
"       @%p4 bra        $Lt_53_17666;\n"
"       .loc    3       719     0\n"
"       ld.shared.f32   %f5, [__smem+0];\n"
"       ld.param.u64    %rd8, [__cudaparm_packed_float_reduce_4_false_false_g_odata];\n"
"       cvt.u64.u32     %rd9, %r1;\n"
"       mul.wide.u32    %rd10, %r1, 4;\n"
"       add.u64         %rd11, %rd8, %rd10;\n"
"       st.global.f32   [%rd11+0], %f5;\n"
"$Lt_53_17666:\n"
"       .loc    3       741     0\n"
"       exit;\n"
"$LDWend_packed_float_reduce_4_false_false:\n"
"       } // packed_float_reduce_4_false_false\n"
"\n"
"       .entry packed_float_reduce_4_false_true (\n"
"               .param .u64 __cudaparm_packed_float_reduce_4_false_true_g_idata,\n"
"               .param .u64 __cudaparm_packed_float_reduce_4_false_true_g_odata,\n"
"               .param .u32 __cudaparm_packed_float_reduce_4_false_true_n)\n"
"       {\n"
"       .reg .u16 %rh<7>;\n"
"       .reg .u32 %r<15>;\n"
"       .reg .u64 %rd<13>;\n"
"       .reg .f32 %f<7>;\n"
"       .reg .pred %p<6>;\n"
"       .local .align 4 .b8 __cuda_local_var_22761_19_pack_0[4];\n"
"       .loc    3       742     0\n"
"$LDWbegin_packed_float_reduce_4_false_true:\n"
"       .loc    3       637     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 8;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_4_false_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_54_18178;\n"
"       mul.lo.u32      %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_4_false_true_n];\n"
"       mul.lo.u32      %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 32;\n"
"       add.u32         %r9, %r6, 16;\n"
"       add.u32         %r10, %r7, 16;\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_4_false_true_g_idata];\n"
"$Lt_54_16642:\n"
" //<loop> Loop body line 637, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       677     0\n"
"       cvt.u8.u32      %r11, %r9;\n"
"       cvt.u64.u32     %rd2, %r11;\n"
"       .loc    3       637     0\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_4_false_true_g_idata];\n"
"       .loc    3       677     0\n"
"       add.u64         %rd3, %rd2, %rd1;\n"
"       ld.global.u8    %rh2, [%rd3+0];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+3], %rh2;\n"
"       .loc    3       678     0\n"
"       ld.global.u8    %rh3, [%rd3+1];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+2], %rh3;\n"
"       .loc    3       679     0\n"
"       ld.global.u8    %rh4, [%rd3+2];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+1], %rh4;\n"
"       .loc    3       680     0\n"
"       ld.global.u8    %rh5, [%rd3+3];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+0], %rh5;\n"
"       add.u32         %r9, %r8, %r9;\n"
"       setp.lt.u32     %p2, %r9, %r10;\n"
"       @%p2 bra        $Lt_54_16642;\n"
"       ld.local.f32    %f1, [__cuda_local_var_22761_19_pack_0+0];\n"
"       bra.uni         $Lt_54_16130;\n"
"$Lt_54_18178:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_54_16130:\n"
"       .loc    3       692     0\n"
"       mov.u64         %rd4, __smem;\n"
"       cvt.u64.u32     %rd5, %r3;\n"
"       mul.wide.u32    %rd6, %r3, 4;\n"
"       add.u64         %rd7, %rd4, %rd6;\n"
"       st.shared.f32   [%rd7+0], %f1;\n"
"       .loc    3       693     0\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 31;\n"
"       setp.gt.u32     %p3, %r3, %r12;\n"
"       @%p3 bra        $Lt_54_17154;\n"
"       .loc    3       713     0\n"
"       ld.volatile.shared.f32  %f2, [%rd7+8];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       st.volatile.shared.f32  [%rd7+0], %f3;\n"
"       .loc    3       714     0\n"
"       ld.volatile.shared.f32  %f4, [%rd7+4];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       st.volatile.shared.f32  [%rd7+0], %f1;\n"
"$Lt_54_17154:\n"
"       mov.u32         %r13, 0;\n"
"       setp.ne.u32     %p4, %r3, %r13;\n"
"       @%p4 bra        $Lt_54_17666;\n"
"       .loc    3       719     0\n"
"       ld.shared.f32   %f5, [__smem+0];\n"
"       ld.param.u64    %rd8, [__cudaparm_packed_float_reduce_4_false_true_g_odata];\n"
"       cvt.u64.u32     %rd9, %r1;\n"
"       mul.wide.u32    %rd10, %r1, 4;\n"
"       add.u64         %rd11, %rd8, %rd10;\n"
"       st.global.f32   [%rd11+0], %f5;\n"
"$Lt_54_17666:\n"
"       .loc    3       743     0\n"
"       exit;\n"
"$LDWend_packed_float_reduce_4_false_true:\n"
"       } // packed_float_reduce_4_false_true\n"
"\n"
"       .entry packed_float_reduce_4_true_false (\n"
"               .param .u64 __cudaparm_packed_float_reduce_4_true_false_g_idata,\n"
"               .param .u64 __cudaparm_packed_float_reduce_4_true_false_g_odata,\n"
"               .param .u32 __cudaparm_packed_float_reduce_4_true_false_n)\n"
"       {\n"
"       .reg .u16 %rh<7>;\n"
"       .reg .u32 %r<15>;\n"
"       .reg .u64 %rd<13>;\n"
"       .reg .f32 %f<7>;\n"
"       .reg .pred %p<6>;\n"
"       .local .align 4 .b8 __cuda_local_var_22761_19_pack_0[4];\n"
"       .loc    3       744     0\n"
"$LDWbegin_packed_float_reduce_4_true_false:\n"
"       .loc    3       637     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 8;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_4_true_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_55_18178;\n"
"       mul.lo.u32      %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_4_true_false_n];\n"
"       mul.lo.u32      %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 32;\n"
"       add.u32         %r9, %r6, 16;\n"
"       add.u32         %r10, %r7, 16;\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_4_true_false_g_idata];\n"
"$Lt_55_16642:\n"
" //<loop> Loop body line 637, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       677     0\n"
"       cvt.u8.u32      %r11, %r9;\n"
"       cvt.u64.u32     %rd2, %r11;\n"
"       .loc    3       637     0\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_4_true_false_g_idata];\n"
"       .loc    3       677     0\n"
"       add.u64         %rd3, %rd2, %rd1;\n"
"       ld.global.u8    %rh2, [%rd3+0];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+3], %rh2;\n"
"       .loc    3       678     0\n"
"       ld.global.u8    %rh3, [%rd3+1];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+2], %rh3;\n"
"       .loc    3       679     0\n"
"       ld.global.u8    %rh4, [%rd3+2];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+1], %rh4;\n"
"       .loc    3       680     0\n"
"       ld.global.u8    %rh5, [%rd3+3];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+0], %rh5;\n"
"       add.u32         %r9, %r8, %r9;\n"
"       setp.lt.u32     %p2, %r9, %r10;\n"
"       @%p2 bra        $Lt_55_16642;\n"
"       ld.local.f32    %f1, [__cuda_local_var_22761_19_pack_0+0];\n"
"       bra.uni         $Lt_55_16130;\n"
"$Lt_55_18178:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_55_16130:\n"
"       .loc    3       692     0\n"
"       mov.u64         %rd4, __smem;\n"
"       cvt.u64.u32     %rd5, %r3;\n"
"       mul.wide.u32    %rd6, %r3, 4;\n"
"       add.u64         %rd7, %rd4, %rd6;\n"
"       st.shared.f32   [%rd7+0], %f1;\n"
"       .loc    3       693     0\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 31;\n"
"       setp.gt.u32     %p3, %r3, %r12;\n"
"       @%p3 bra        $Lt_55_17154;\n"
"       .loc    3       713     0\n"
"       ld.volatile.shared.f32  %f2, [%rd7+8];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       st.volatile.shared.f32  [%rd7+0], %f3;\n"
"       .loc    3       714     0\n"
"       ld.volatile.shared.f32  %f4, [%rd7+4];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       st.volatile.shared.f32  [%rd7+0], %f1;\n"
"$Lt_55_17154:\n"
"       mov.u32         %r13, 0;\n"
"       setp.ne.u32     %p4, %r3, %r13;\n"
"       @%p4 bra        $Lt_55_17666;\n"
"       .loc    3       719     0\n"
"       ld.shared.f32   %f5, [__smem+0];\n"
"       ld.param.u64    %rd8, [__cudaparm_packed_float_reduce_4_true_false_g_odata];\n"
"       cvt.u64.u32     %rd9, %r1;\n"
"       mul.wide.u32    %rd10, %r1, 4;\n"
"       add.u64         %rd11, %rd8, %rd10;\n"
"       st.global.f32   [%rd11+0], %f5;\n"
"$Lt_55_17666:\n"
"       .loc    3       745     0\n"
"       exit;\n"
"$LDWend_packed_float_reduce_4_true_false:\n"
"       } // packed_float_reduce_4_true_false\n"
"\n"
"       .entry packed_float_reduce_4_true_true (\n"
"               .param .u64 __cudaparm_packed_float_reduce_4_true_true_g_idata,\n"
"               .param .u64 __cudaparm_packed_float_reduce_4_true_true_g_odata,\n"
"               .param .u32 __cudaparm_packed_float_reduce_4_true_true_n)\n"
"       {\n"
"       .reg .u16 %rh<7>;\n"
"       .reg .u32 %r<15>;\n"
"       .reg .u64 %rd<13>;\n"
"       .reg .f32 %f<7>;\n"
"       .reg .pred %p<6>;\n"
"       .local .align 4 .b8 __cuda_local_var_22761_19_pack_0[4];\n"
"       .loc    3       746     0\n"
"$LDWbegin_packed_float_reduce_4_true_true:\n"
"       .loc    3       637     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 8;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_4_true_true_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_56_18178;\n"
"       mul.lo.u32      %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_4_true_true_n];\n"
"       mul.lo.u32      %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 32;\n"
"       add.u32         %r9, %r6, 16;\n"
"       add.u32         %r10, %r7, 16;\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_4_true_true_g_idata];\n"
"$Lt_56_16642:\n"
" //<loop> Loop body line 637, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       677     0\n"
"       cvt.u8.u32      %r11, %r9;\n"
"       cvt.u64.u32     %rd2, %r11;\n"
"       .loc    3       637     0\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_4_true_true_g_idata];\n"
"       .loc    3       677     0\n"
"       add.u64         %rd3, %rd2, %rd1;\n"
"       ld.global.u8    %rh2, [%rd3+0];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+3], %rh2;\n"
"       .loc    3       678     0\n"
"       ld.global.u8    %rh3, [%rd3+1];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+2], %rh3;\n"
"       .loc    3       679     0\n"
"       ld.global.u8    %rh4, [%rd3+2];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+1], %rh4;\n"
"       .loc    3       680     0\n"
"       ld.global.u8    %rh5, [%rd3+3];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+0], %rh5;\n"
"       add.u32         %r9, %r8, %r9;\n"
"       setp.lt.u32     %p2, %r9, %r10;\n"
"       @%p2 bra        $Lt_56_16642;\n"
"       ld.local.f32    %f1, [__cuda_local_var_22761_19_pack_0+0];\n"
"       bra.uni         $Lt_56_16130;\n"
"$Lt_56_18178:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_56_16130:\n"
"       .loc    3       692     0\n"
"       mov.u64         %rd4, __smem;\n"
"       cvt.u64.u32     %rd5, %r3;\n"
"       mul.wide.u32    %rd6, %r3, 4;\n"
"       add.u64         %rd7, %rd4, %rd6;\n"
"       st.shared.f32   [%rd7+0], %f1;\n"
"       .loc    3       693     0\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 31;\n"
"       setp.gt.u32     %p3, %r3, %r12;\n"
"       @%p3 bra        $Lt_56_17154;\n"
"       .loc    3       713     0\n"
"       ld.volatile.shared.f32  %f2, [%rd7+8];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       st.volatile.shared.f32  [%rd7+0], %f3;\n"
"       .loc    3       714     0\n"
"       ld.volatile.shared.f32  %f4, [%rd7+4];\n"
"       add.f32         %f1, %f4, %f3;\n"
"       st.volatile.shared.f32  [%rd7+0], %f1;\n"
"$Lt_56_17154:\n"
"       mov.u32         %r13, 0;\n"
"       setp.ne.u32     %p4, %r3, %r13;\n"
"       @%p4 bra        $Lt_56_17666;\n"
"       .loc    3       719     0\n"
"       ld.shared.f32   %f5, [__smem+0];\n"
"       ld.param.u64    %rd8, [__cudaparm_packed_float_reduce_4_true_true_g_odata];\n"
"       cvt.u64.u32     %rd9, %r1;\n"
"       mul.wide.u32    %rd10, %r1, 4;\n"
"       add.u64         %rd11, %rd8, %rd10;\n"
"       st.global.f32   [%rd11+0], %f5;\n"
"$Lt_56_17666:\n"
"       .loc    3       747     0\n"
"       exit;\n"
"$LDWend_packed_float_reduce_4_true_true:\n"
"       } // packed_float_reduce_4_true_true\n"
"\n"
"       .entry packed_float_reduce_8_false_false (\n"
"               .param .u64 __cudaparm_packed_float_reduce_8_false_false_g_idata,\n"
"               .param .u64 __cudaparm_packed_float_reduce_8_false_false_g_odata,\n"
"               .param .u32 __cudaparm_packed_float_reduce_8_false_false_n)\n"
"       {\n"
"       .reg .u16 %rh<7>;\n"
"       .reg .u32 %r<15>;\n"
"       .reg .u64 %rd<13>;\n"
"       .reg .f32 %f<9>;\n"
"       .reg .pred %p<6>;\n"
"       .local .align 4 .b8 __cuda_local_var_22761_19_pack_0[4];\n"
"       .loc    3       749     0\n"
"$LDWbegin_packed_float_reduce_8_false_false:\n"
"       .loc    3       637     0\n"
"       cvt.u32.u16     %r1, %ctaid.x;\n"
"       mul24.lo.u32    %r2, %r1, 16;\n"
"       cvt.u32.u16     %r3, %tid.x;\n"
"       add.u32         %r4, %r2, %r3;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_8_false_false_n];\n"
"       setp.ge.u32     %p1, %r4, %r5;\n"
"       @%p1 bra        $Lt_57_17922;\n"
"       mul.lo.u32      %r6, %r4, 4;\n"
"       ld.param.u32    %r5, [__cudaparm_packed_float_reduce_8_false_false_n];\n"
"       mul.lo.u32      %r7, %r5, 4;\n"
"       mov.u16         %rh1, %nctaid.x;\n"
"       mul.wide.u16    %r8, %rh1, 64;\n"
"       add.u32         %r9, %r6, 32;\n"
"       add.u32         %r10, %r7, 32;\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_8_false_false_g_idata];\n"
"$Lt_57_16386:\n"
" //<loop> Loop body line 637, nesting depth: 1, estimated iterations: unknown\n"
"       .loc    3       677     0\n"
"       cvt.u8.u32      %r11, %r9;\n"
"       cvt.u64.u32     %rd2, %r11;\n"
"       .loc    3       637     0\n"
"       ld.param.u64    %rd1, [__cudaparm_packed_float_reduce_8_false_false_g_idata];\n"
"       .loc    3       677     0\n"
"       add.u64         %rd3, %rd2, %rd1;\n"
"       ld.global.u8    %rh2, [%rd3+0];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+3], %rh2;\n"
"       .loc    3       678     0\n"
"       ld.global.u8    %rh3, [%rd3+1];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+2], %rh3;\n"
"       .loc    3       679     0\n"
"       ld.global.u8    %rh4, [%rd3+2];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+1], %rh4;\n"
"       .loc    3       680     0\n"
"       ld.global.u8    %rh5, [%rd3+3];\n"
"       st.local.u8     [__cuda_local_var_22761_19_pack_0+0], %rh5;\n"
"       add.u32         %r9, %r8, %r9;\n"
"       setp.lt.u32     %p2, %r9, %r10;\n"
"       @%p2 bra        $Lt_57_16386;\n"
"       ld.local.f32    %f1, [__cuda_local_var_22761_19_pack_0+0];\n"
"       bra.uni         $Lt_57_15874;\n"
"$Lt_57_17922:\n"
"       mov.f32         %f1, 0f00000000;        // 0\n"
"$Lt_57_15874:\n"
"       .loc    3       692     0\n"
"       mov.u64         %rd4, __smem;\n"
"       cvt.u64.u32     %rd5, %r3;\n"
"       mul.wide.u32    %rd6, %r3, 4;\n"
"       add.u64         %rd7, %rd4, %rd6;\n"
"       st.shared.f32   [%rd7+0], %f1;\n"
"       .loc    3       693     0\n"
"       bar.sync        0;\n"
"       mov.u32         %r12, 31;\n"
"       setp.gt.u32     %p3, %r3, %r12;\n"
"       @%p3 bra        $Lt_57_16898;\n"
"       .loc    3       712     0\n"
"       ld.volatile.shared.f32  %f2, [%rd7+16];\n"
"       add.f32         %f3, %f2, %f1;\n"
"       st.volatile.shared.f32  [%rd7+0], %f3;\n"
"       .loc    3       713     0\n"
"       ld.volatile.shared.f32  %f4, [%rd7+8];\n"
"       add.f32         %f5, %f4, %f3;\n"
"       st.volatile.shared.f32  [%rd7+0], %f5;\n"
"       .loc    3       714     0\n"
"       ld.volatile.shared.f32  %f6, [%rd7+4];\n"
"       add.f32         %f1, %f6, %f5;\n"
"       st.volatile.shared.f32  [%rd7+0], %f1;\n"
"$Lt_57_16898:\n"
"       mov.u32         %r13, 0;\n"
"       setp.ne.u32     %p4, %r3, %r13;\n"
"       @%p4 bra        $Lt_57_17410;\n"
"       .loc    3       719     0\n"
"       ld.shared.f32   %f7, [__smem+0];\n"
"       ld.param.u64    %rd8, [__cudaparm_packed_float_reduce_8_false_false_g_odata];\n"