958 lines
31 KiB
C
958 lines
31 KiB
C
const char * coul_long =
|
|
" .version 2.3\n"
|
|
" .target sm_20\n"
|
|
" .address_size 64\n"
|
|
" .global .texref pos_tex;\n"
|
|
" .global .texref q_tex;\n"
|
|
" .entry kernel_pair (\n"
|
|
" .param .u64 __cudaparm_kernel_pair_x_,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_lj1,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_lj3,\n"
|
|
" .param .s32 __cudaparm_kernel_pair_lj_types,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_sp_cl_in,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_dev_nbor,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_dev_packed,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_ans,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_engv,\n"
|
|
" .param .s32 __cudaparm_kernel_pair_eflag,\n"
|
|
" .param .s32 __cudaparm_kernel_pair_vflag,\n"
|
|
" .param .s32 __cudaparm_kernel_pair_inum,\n"
|
|
" .param .s32 __cudaparm_kernel_pair_nbor_pitch,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_q_,\n"
|
|
" .param .f32 __cudaparm_kernel_pair_cut_coulsq,\n"
|
|
" .param .f32 __cudaparm_kernel_pair_qqrd2e,\n"
|
|
" .param .f32 __cudaparm_kernel_pair_g_ewald,\n"
|
|
" .param .s32 __cudaparm_kernel_pair_t_per_atom)\n"
|
|
" {\n"
|
|
" .reg .u32 %r<81>;\n"
|
|
" .reg .u64 %rd<58>;\n"
|
|
" .reg .f32 %f<132>;\n"
|
|
" .reg .pred %p<19>;\n"
|
|
" .shared .align 16 .b8 __cuda___cuda_local_var_32541_33_non_const_sp_cl112[16];\n"
|
|
" .shared .align 4 .b8 __cuda___cuda_local_var_32611_37_non_const_red_acc128[3072];\n"
|
|
" .loc 16 36 0\n"
|
|
"$LDWbegin_kernel_pair:\n"
|
|
" .loc 16 41 0\n"
|
|
" ld.param.u64 %rd1, [__cudaparm_kernel_pair_sp_cl_in];\n"
|
|
" ldu.global.f32 %f1, [%rd1+0];\n"
|
|
" .loc 16 42 0\n"
|
|
" ld.global.f32 %f2, [%rd1+4];\n"
|
|
" .loc 16 43 0\n"
|
|
" ld.global.f32 %f3, [%rd1+8];\n"
|
|
" .loc 16 44 0\n"
|
|
" ld.global.f32 %f4, [%rd1+12];\n"
|
|
" st.shared.v4.f32 [__cuda___cuda_local_var_32541_33_non_const_sp_cl112+0], {%f1,%f2,%f3,%f4};\n"
|
|
" .loc 16 51 0\n"
|
|
" mov.f32 %f5, 0f00000000; \n"
|
|
" mov.f32 %f6, %f5;\n"
|
|
" mov.f32 %f7, 0f00000000; \n"
|
|
" mov.f32 %f8, %f7;\n"
|
|
" mov.f32 %f9, 0f00000000; \n"
|
|
" mov.f32 %f10, %f9;\n"
|
|
" mov.f32 %f11, 0f00000000; \n"
|
|
" mov.f32 %f12, %f11;\n"
|
|
" mov.f32 %f13, 0f00000000; \n"
|
|
" mov.f32 %f14, %f13;\n"
|
|
" mov.f32 %f15, 0f00000000; \n"
|
|
" mov.f32 %f16, %f15;\n"
|
|
" ld.param.s32 %r1, [__cudaparm_kernel_pair_t_per_atom];\n"
|
|
" cvt.s32.u32 %r2, %tid.x;\n"
|
|
" div.s32 %r3, %r2, %r1;\n"
|
|
" cvt.s32.u32 %r4, %ntid.x;\n"
|
|
" div.s32 %r5, %r4, %r1;\n"
|
|
" cvt.s32.u32 %r6, %ctaid.x;\n"
|
|
" mul.lo.s32 %r7, %r6, %r5;\n"
|
|
" add.s32 %r8, %r3, %r7;\n"
|
|
" ld.param.s32 %r9, [__cudaparm_kernel_pair_inum];\n"
|
|
" setp.ge.s32 %p1, %r8, %r9;\n"
|
|
" @%p1 bra $Lt_0_25858;\n"
|
|
" .loc 16 56 0\n"
|
|
" cvt.s64.s32 %rd2, %r8;\n"
|
|
" mul.wide.s32 %rd3, %r8, 4;\n"
|
|
" ld.param.u64 %rd4, [__cudaparm_kernel_pair_dev_nbor];\n"
|
|
" add.u64 %rd5, %rd3, %rd4;\n"
|
|
" ld.global.s32 %r10, [%rd5+0];\n"
|
|
" ld.param.s32 %r11, [__cudaparm_kernel_pair_nbor_pitch];\n"
|
|
" cvt.s64.s32 %rd6, %r11;\n"
|
|
" mul.wide.s32 %rd7, %r11, 4;\n"
|
|
" add.u64 %rd8, %rd7, %rd5;\n"
|
|
" ld.global.s32 %r12, [%rd8+0];\n"
|
|
" sub.s32 %r13, %r1, 1;\n"
|
|
" and.b32 %r14, %r13, %r2;\n"
|
|
" cvt.s64.s32 %rd9, %r14;\n"
|
|
" mul.wide.s32 %rd10, %r14, 4;\n"
|
|
" ld.param.u64 %rd11, [__cudaparm_kernel_pair_dev_packed];\n"
|
|
" setp.ne.u64 %p2, %rd11, %rd4;\n"
|
|
" @%p2 bra $Lt_0_19458;\n"
|
|
" cvt.s32.s64 %r15, %rd6;\n"
|
|
" mul.lo.s32 %r16, %r15, %r1;\n"
|
|
" mov.s32 %r17, %r16;\n"
|
|
" mul.lo.s32 %r18, %r13, %r8;\n"
|
|
" add.s32 %r19, %r15, %r18;\n"
|
|
" cvt.s64.s32 %rd12, %r19;\n"
|
|
" mul.wide.s32 %rd13, %r19, 4;\n"
|
|
" add.u64 %rd14, %rd8, %rd13;\n"
|
|
" and.b32 %r20, %r13, %r12;\n"
|
|
" cvt.s64.s32 %rd15, %r20;\n"
|
|
" div.s32 %r21, %r12, %r1;\n"
|
|
" mul.lo.s32 %r22, %r16, %r21;\n"
|
|
" cvt.s64.s32 %rd16, %r22;\n"
|
|
" add.u64 %rd17, %rd15, %rd16;\n"
|
|
" mul.lo.u64 %rd18, %rd17, 4;\n"
|
|
" add.u64 %rd19, %rd14, %rd18;\n"
|
|
" add.u64 %rd20, %rd10, %rd14;\n"
|
|
" bra.uni $Lt_0_19202;\n"
|
|
"$Lt_0_19458:\n"
|
|
" add.u64 %rd21, %rd7, %rd8;\n"
|
|
" ld.global.s32 %r23, [%rd21+0];\n"
|
|
" cvt.s64.s32 %rd22, %r23;\n"
|
|
" mul.wide.s32 %rd23, %r23, 4;\n"
|
|
" add.u64 %rd24, %rd11, %rd23;\n"
|
|
" cvt.s64.s32 %rd25, %r12;\n"
|
|
" mul.wide.s32 %rd26, %r12, 4;\n"
|
|
" add.u64 %rd19, %rd24, %rd26;\n"
|
|
" mov.s32 %r17, %r1;\n"
|
|
" add.u64 %rd20, %rd10, %rd24;\n"
|
|
"$Lt_0_19202:\n"
|
|
" .loc 16 59 0\n"
|
|
" mov.u32 %r24, %r10;\n"
|
|
" mov.s32 %r25, 0;\n"
|
|
" mov.u32 %r26, %r25;\n"
|
|
" mov.s32 %r27, 0;\n"
|
|
" mov.u32 %r28, %r27;\n"
|
|
" mov.s32 %r29, 0;\n"
|
|
" mov.u32 %r30, %r29;\n"
|
|
" tex.1d.v4.f32.s32 {%f17,%f18,%f19,%f20},[pos_tex,{%r24,%r26,%r28,%r30}];\n"
|
|
" mov.f32 %f21, %f17;\n"
|
|
" mov.f32 %f22, %f18;\n"
|
|
" mov.f32 %f23, %f19;\n"
|
|
" .loc 16 60 0\n"
|
|
" mov.u32 %r31, %r10;\n"
|
|
" mov.s32 %r32, 0;\n"
|
|
" mov.u32 %r33, %r32;\n"
|
|
" mov.s32 %r34, 0;\n"
|
|
" mov.u32 %r35, %r34;\n"
|
|
" mov.s32 %r36, 0;\n"
|
|
" mov.u32 %r37, %r36;\n"
|
|
" tex.1d.v4.f32.s32 {%f24,%f25,%f26,%f27},[q_tex,{%r31,%r33,%r35,%r37}];\n"
|
|
" mov.f32 %f28, %f24;\n"
|
|
" setp.ge.u64 %p3, %rd20, %rd19;\n"
|
|
" @%p3 bra $Lt_0_27394;\n"
|
|
" cvt.s64.s32 %rd27, %r17;\n"
|
|
" ld.param.f32 %f29, [__cudaparm_kernel_pair_cut_coulsq];\n"
|
|
" mov.f32 %f30, 0f00000000; \n"
|
|
" mov.f32 %f31, 0f00000000; \n"
|
|
" mov.f32 %f32, 0f00000000; \n"
|
|
" mov.f32 %f33, 0f00000000; \n"
|
|
" mov.u64 %rd28, __cuda___cuda_local_var_32541_33_non_const_sp_cl112;\n"
|
|
"$Lt_0_20226:\n"
|
|
" .loc 16 63 0\n"
|
|
" ld.global.s32 %r38, [%rd20+0];\n"
|
|
" .loc 16 66 0\n"
|
|
" mov.f32 %f34, 0f3f800000; \n"
|
|
" shr.s32 %r39, %r38, 30;\n"
|
|
" and.b32 %r40, %r39, 3;\n"
|
|
" cvt.s64.s32 %rd29, %r40;\n"
|
|
" mul.wide.s32 %rd30, %r40, 4;\n"
|
|
" add.u64 %rd31, %rd28, %rd30;\n"
|
|
" ld.shared.f32 %f35, [%rd31+0];\n"
|
|
" sub.ftz.f32 %f36, %f34, %f35;\n"
|
|
" .loc 16 69 0\n"
|
|
" and.b32 %r41, %r38, 1073741823;\n"
|
|
" mov.u32 %r42, %r41;\n"
|
|
" mov.s32 %r43, 0;\n"
|
|
" mov.u32 %r44, %r43;\n"
|
|
" mov.s32 %r45, 0;\n"
|
|
" mov.u32 %r46, %r45;\n"
|
|
" mov.s32 %r47, 0;\n"
|
|
" mov.u32 %r48, %r47;\n"
|
|
" tex.1d.v4.f32.s32 {%f37,%f38,%f39,%f40},[pos_tex,{%r42,%r44,%r46,%r48}];\n"
|
|
" mov.f32 %f41, %f37;\n"
|
|
" mov.f32 %f42, %f38;\n"
|
|
" mov.f32 %f43, %f39;\n"
|
|
" sub.ftz.f32 %f44, %f22, %f42;\n"
|
|
" sub.ftz.f32 %f45, %f21, %f41;\n"
|
|
" sub.ftz.f32 %f46, %f23, %f43;\n"
|
|
" mul.ftz.f32 %f47, %f44, %f44;\n"
|
|
" fma.rn.ftz.f32 %f48, %f45, %f45, %f47;\n"
|
|
" fma.rn.ftz.f32 %f49, %f46, %f46, %f48;\n"
|
|
" setp.lt.ftz.f32 %p4, %f49, %f29;\n"
|
|
" @!%p4 bra $Lt_0_20994;\n"
|
|
" .loc 20 518 0\n"
|
|
" rcp.approx.ftz.f32 %f50, %f49;\n"
|
|
" rsqrt.approx.ftz.f32 %f51, %f50;\n"
|
|
" ld.param.f32 %f52, [__cudaparm_kernel_pair_g_ewald];\n"
|
|
" mul.ftz.f32 %f53, %f52, %f51;\n"
|
|
" mul.ftz.f32 %f54, %f53, %f53;\n"
|
|
" neg.ftz.f32 %f55, %f54;\n"
|
|
" mov.f32 %f56, 0f3fb8aa3b; \n"
|
|
" mul.ftz.f32 %f57, %f55, %f56;\n"
|
|
" ex2.approx.ftz.f32 %f58, %f57;\n"
|
|
" .loc 16 85 0\n"
|
|
" mov.f32 %f59, 0f3f800000; \n"
|
|
" mov.f32 %f60, 0f3ea7ba05; \n"
|
|
" fma.rn.ftz.f32 %f61, %f60, %f53, %f59;\n"
|
|
" rcp.approx.ftz.f32 %f62, %f61;\n"
|
|
" mov.f32 %f63, 0f3e827906; \n"
|
|
" mov.f32 %f64, 0fbe91a98e; \n"
|
|
" mov.f32 %f65, 0f3fb5f0e3; \n"
|
|
" mov.f32 %f66, 0fbfba00e3; \n"
|
|
" mov.f32 %f67, 0f3f87dc22; \n"
|
|
" fma.rn.ftz.f32 %f68, %f67, %f62, %f66;\n"
|
|
" fma.rn.ftz.f32 %f69, %f62, %f68, %f65;\n"
|
|
" fma.rn.ftz.f32 %f70, %f62, %f69, %f64;\n"
|
|
" fma.rn.ftz.f32 %f71, %f62, %f70, %f63;\n"
|
|
" mul.ftz.f32 %f72, %f62, %f71;\n"
|
|
" mul.ftz.f32 %f73, %f58, %f72;\n"
|
|
" .loc 16 86 0\n"
|
|
" mov.u32 %r49, %r41;\n"
|
|
" mov.s32 %r50, 0;\n"
|
|
" mov.u32 %r51, %r50;\n"
|
|
" mov.s32 %r52, 0;\n"
|
|
" mov.u32 %r53, %r52;\n"
|
|
" mov.s32 %r54, 0;\n"
|
|
" mov.u32 %r55, %r54;\n"
|
|
" tex.1d.v4.f32.s32 {%f74,%f75,%f76,%f77},[q_tex,{%r49,%r51,%r53,%r55}];\n"
|
|
" mov.f32 %f78, %f74;\n"
|
|
" .loc 16 87 0\n"
|
|
" ld.param.f32 %f79, [__cudaparm_kernel_pair_qqrd2e];\n"
|
|
" mul.ftz.f32 %f80, %f79, %f28;\n"
|
|
" mul.ftz.f32 %f81, %f80, %f78;\n"
|
|
" div.approx.ftz.f32 %f82, %f81, %f51;\n"
|
|
" mov.f32 %f83, 0f3f906ebb; \n"
|
|
" mul.ftz.f32 %f84, %f53, %f83;\n"
|
|
" fma.rn.ftz.f32 %f85, %f58, %f84, %f73;\n"
|
|
" sub.ftz.f32 %f86, %f85, %f36;\n"
|
|
" mul.ftz.f32 %f87, %f82, %f86;\n"
|
|
" mul.ftz.f32 %f88, %f50, %f87;\n"
|
|
" .loc 16 89 0\n"
|
|
" fma.rn.ftz.f32 %f32, %f45, %f88, %f32;\n"
|
|
" .loc 16 90 0\n"
|
|
" fma.rn.ftz.f32 %f31, %f44, %f88, %f31;\n"
|
|
" .loc 16 91 0\n"
|
|
" fma.rn.ftz.f32 %f30, %f46, %f88, %f30;\n"
|
|
" .loc 16 78 0\n"
|
|
" sub.ftz.f32 %f89, %f73, %f36;\n"
|
|
" fma.rn.ftz.f32 %f90, %f82, %f89, %f33;\n"
|
|
" ld.param.s32 %r56, [__cudaparm_kernel_pair_eflag];\n"
|
|
" mov.s32 %r57, 0;\n"
|
|
" setp.gt.s32 %p5, %r56, %r57;\n"
|
|
" selp.f32 %f33, %f90, %f33, %p5;\n"
|
|
" ld.param.s32 %r58, [__cudaparm_kernel_pair_vflag];\n"
|
|
" mov.u32 %r59, 0;\n"
|
|
" setp.le.s32 %p6, %r58, %r59;\n"
|
|
" @%p6 bra $Lt_0_20994;\n"
|
|
" .loc 16 97 0\n"
|
|
" mov.f32 %f91, %f6;\n"
|
|
" mul.ftz.f32 %f92, %f45, %f45;\n"
|
|
" fma.rn.ftz.f32 %f93, %f88, %f92, %f91;\n"
|
|
" mov.f32 %f6, %f93;\n"
|
|
" .loc 16 98 0\n"
|
|
" mov.f32 %f94, %f8;\n"
|
|
" fma.rn.ftz.f32 %f95, %f88, %f47, %f94;\n"
|
|
" mov.f32 %f8, %f95;\n"
|
|
" .loc 16 99 0\n"
|
|
" mov.f32 %f96, %f10;\n"
|
|
" mul.ftz.f32 %f97, %f46, %f46;\n"
|
|
" fma.rn.ftz.f32 %f98, %f88, %f97, %f96;\n"
|
|
" mov.f32 %f10, %f98;\n"
|
|
" .loc 16 100 0\n"
|
|
" mov.f32 %f99, %f12;\n"
|
|
" mul.ftz.f32 %f100, %f44, %f45;\n"
|
|
" fma.rn.ftz.f32 %f101, %f88, %f100, %f99;\n"
|
|
" mov.f32 %f12, %f101;\n"
|
|
" .loc 16 101 0\n"
|
|
" mov.f32 %f102, %f14;\n"
|
|
" mul.ftz.f32 %f103, %f45, %f46;\n"
|
|
" fma.rn.ftz.f32 %f104, %f88, %f103, %f102;\n"
|
|
" mov.f32 %f14, %f104;\n"
|
|
" .loc 16 102 0\n"
|
|
" mul.ftz.f32 %f105, %f44, %f46;\n"
|
|
" fma.rn.ftz.f32 %f15, %f88, %f105, %f15;\n"
|
|
" mov.f32 %f16, %f15;\n"
|
|
"$Lt_0_20994:\n"
|
|
"$Lt_0_20482:\n"
|
|
" .loc 16 62 0\n"
|
|
" mul.lo.u64 %rd32, %rd27, 4;\n"
|
|
" add.u64 %rd20, %rd20, %rd32;\n"
|
|
" setp.lt.u64 %p7, %rd20, %rd19;\n"
|
|
" @%p7 bra $Lt_0_20226;\n"
|
|
" bra.uni $Lt_0_19714;\n"
|
|
"$Lt_0_27394:\n"
|
|
" mov.f32 %f30, 0f00000000; \n"
|
|
" mov.f32 %f31, 0f00000000; \n"
|
|
" mov.f32 %f32, 0f00000000; \n"
|
|
" mov.f32 %f33, 0f00000000; \n"
|
|
"$Lt_0_19714:\n"
|
|
" mov.u32 %r60, 1;\n"
|
|
" setp.le.s32 %p8, %r1, %r60;\n"
|
|
" @%p8 bra $Lt_0_23810;\n"
|
|
" .loc 16 112 0\n"
|
|
" mov.u64 %rd33, __cuda___cuda_local_var_32611_37_non_const_red_acc128;\n"
|
|
" cvt.s64.s32 %rd34, %r2;\n"
|
|
" mul.wide.s32 %rd35, %r2, 4;\n"
|
|
" add.u64 %rd36, %rd33, %rd35;\n"
|
|
" mov.f32 %f106, %f32;\n"
|
|
" st.shared.f32 [%rd36+0], %f106;\n"
|
|
" .loc 16 113 0\n"
|
|
" mov.f32 %f107, %f31;\n"
|
|
" st.shared.f32 [%rd36+512], %f107;\n"
|
|
" .loc 16 114 0\n"
|
|
" mov.f32 %f108, %f30;\n"
|
|
" st.shared.f32 [%rd36+1024], %f108;\n"
|
|
" .loc 16 115 0\n"
|
|
" mov.f32 %f109, %f33;\n"
|
|
" st.shared.f32 [%rd36+1536], %f109;\n"
|
|
" .loc 16 117 0\n"
|
|
" shr.s32 %r61, %r1, 31;\n"
|
|
" mov.s32 %r62, 1;\n"
|
|
" and.b32 %r63, %r61, %r62;\n"
|
|
" add.s32 %r64, %r63, %r1;\n"
|
|
" shr.s32 %r65, %r64, 1;\n"
|
|
" mov.s32 %r66, %r65;\n"
|
|
" mov.u32 %r67, 0;\n"
|
|
" setp.ne.u32 %p9, %r65, %r67;\n"
|
|
" @!%p9 bra $Lt_0_22274;\n"
|
|
"$Lt_0_22786:\n"
|
|
" setp.ge.u32 %p10, %r14, %r66;\n"
|
|
" @%p10 bra $Lt_0_23042;\n"
|
|
" .loc 16 120 0\n"
|
|
" add.u32 %r68, %r2, %r66;\n"
|
|
" cvt.u64.u32 %rd37, %r68;\n"
|
|
" mul.wide.u32 %rd38, %r68, 4;\n"
|
|
" add.u64 %rd39, %rd33, %rd38;\n"
|
|
" ld.shared.f32 %f110, [%rd39+0];\n"
|
|
" add.ftz.f32 %f106, %f110, %f106;\n"
|
|
" st.shared.f32 [%rd36+0], %f106;\n"
|
|
" ld.shared.f32 %f111, [%rd39+512];\n"
|
|
" add.ftz.f32 %f107, %f111, %f107;\n"
|
|
" st.shared.f32 [%rd36+512], %f107;\n"
|
|
" ld.shared.f32 %f112, [%rd39+1024];\n"
|
|
" add.ftz.f32 %f108, %f112, %f108;\n"
|
|
" st.shared.f32 [%rd36+1024], %f108;\n"
|
|
" ld.shared.f32 %f113, [%rd39+1536];\n"
|
|
" add.ftz.f32 %f109, %f113, %f109;\n"
|
|
" st.shared.f32 [%rd36+1536], %f109;\n"
|
|
"$Lt_0_23042:\n"
|
|
" .loc 16 117 0\n"
|
|
" shr.u32 %r66, %r66, 1;\n"
|
|
" mov.u32 %r69, 0;\n"
|
|
" setp.ne.u32 %p11, %r66, %r69;\n"
|
|
" @%p11 bra $Lt_0_22786;\n"
|
|
"$Lt_0_22274:\n"
|
|
" .loc 16 124 0\n"
|
|
" mov.f32 %f32, %f106;\n"
|
|
" .loc 16 125 0\n"
|
|
" mov.f32 %f31, %f107;\n"
|
|
" .loc 16 126 0\n"
|
|
" mov.f32 %f30, %f108;\n"
|
|
" .loc 16 127 0\n"
|
|
" mov.f32 %f33, %f109;\n"
|
|
" ld.param.s32 %r70, [__cudaparm_kernel_pair_vflag];\n"
|
|
" mov.u32 %r71, 0;\n"
|
|
" setp.le.s32 %p12, %r70, %r71;\n"
|
|
" @%p12 bra $Lt_0_23810;\n"
|
|
" .loc 16 131 0\n"
|
|
" mov.f32 %f106, %f6;\n"
|
|
" st.shared.f32 [%rd36+0], %f106;\n"
|
|
" mov.f32 %f107, %f8;\n"
|
|
" st.shared.f32 [%rd36+512], %f107;\n"
|
|
" mov.f32 %f108, %f10;\n"
|
|
" st.shared.f32 [%rd36+1024], %f108;\n"
|
|
" mov.f32 %f109, %f12;\n"
|
|
" st.shared.f32 [%rd36+1536], %f109;\n"
|
|
" mov.f32 %f114, %f14;\n"
|
|
" st.shared.f32 [%rd36+2048], %f114;\n"
|
|
" mov.f32 %f115, %f15;\n"
|
|
" st.shared.f32 [%rd36+2560], %f115;\n"
|
|
" .loc 16 133 0\n"
|
|
" mov.s32 %r72, %r65;\n"
|
|
" @!%p9 bra $Lt_0_24322;\n"
|
|
"$Lt_0_24834:\n"
|
|
" setp.ge.u32 %p13, %r14, %r72;\n"
|
|
" @%p13 bra $Lt_0_25090;\n"
|
|
" .loc 16 136 0\n"
|
|
" add.u32 %r73, %r2, %r72;\n"
|
|
" cvt.u64.u32 %rd40, %r73;\n"
|
|
" mul.wide.u32 %rd41, %r73, 4;\n"
|
|
" add.u64 %rd42, %rd33, %rd41;\n"
|
|
" ld.shared.f32 %f116, [%rd42+0];\n"
|
|
" add.ftz.f32 %f106, %f116, %f106;\n"
|
|
" st.shared.f32 [%rd36+0], %f106;\n"
|
|
" ld.shared.f32 %f117, [%rd42+512];\n"
|
|
" add.ftz.f32 %f107, %f117, %f107;\n"
|
|
" st.shared.f32 [%rd36+512], %f107;\n"
|
|
" ld.shared.f32 %f118, [%rd42+1024];\n"
|
|
" add.ftz.f32 %f108, %f118, %f108;\n"
|
|
" st.shared.f32 [%rd36+1024], %f108;\n"
|
|
" ld.shared.f32 %f119, [%rd42+1536];\n"
|
|
" add.ftz.f32 %f109, %f119, %f109;\n"
|
|
" st.shared.f32 [%rd36+1536], %f109;\n"
|
|
" ld.shared.f32 %f120, [%rd42+2048];\n"
|
|
" add.ftz.f32 %f114, %f120, %f114;\n"
|
|
" st.shared.f32 [%rd36+2048], %f114;\n"
|
|
" ld.shared.f32 %f121, [%rd42+2560];\n"
|
|
" add.ftz.f32 %f115, %f121, %f115;\n"
|
|
" st.shared.f32 [%rd36+2560], %f115;\n"
|
|
"$Lt_0_25090:\n"
|
|
" .loc 16 133 0\n"
|
|
" shr.u32 %r72, %r72, 1;\n"
|
|
" mov.u32 %r74, 0;\n"
|
|
" setp.ne.u32 %p14, %r72, %r74;\n"
|
|
" @%p14 bra $Lt_0_24834;\n"
|
|
"$Lt_0_24322:\n"
|
|
" .loc 16 141 0\n"
|
|
" mov.f32 %f6, %f106;\n"
|
|
" mov.f32 %f8, %f107;\n"
|
|
" mov.f32 %f10, %f108;\n"
|
|
" mov.f32 %f12, %f109;\n"
|
|
" mov.f32 %f14, %f114;\n"
|
|
" mov.f32 %f16, %f115;\n"
|
|
"$Lt_0_23810:\n"
|
|
"$Lt_0_21762:\n"
|
|
" mov.u32 %r75, 0;\n"
|
|
" setp.ne.s32 %p15, %r14, %r75;\n"
|
|
" @%p15 bra $Lt_0_25858;\n"
|
|
" .loc 16 147 0\n"
|
|
" ld.param.u64 %rd43, [__cudaparm_kernel_pair_engv];\n"
|
|
" add.u64 %rd44, %rd43, %rd3;\n"
|
|
" ld.param.s32 %r76, [__cudaparm_kernel_pair_eflag];\n"
|
|
" mov.u32 %r77, 0;\n"
|
|
" setp.le.s32 %p16, %r76, %r77;\n"
|
|
" @%p16 bra $Lt_0_26370;\n"
|
|
" .loc 16 149 0\n"
|
|
" mov.f32 %f122, 0f00000000; \n"
|
|
" st.global.f32 [%rd44+0], %f122;\n"
|
|
" .loc 16 150 0\n"
|
|
" cvt.s64.s32 %rd45, %r9;\n"
|
|
" mul.wide.s32 %rd46, %r9, 4;\n"
|
|
" add.u64 %rd47, %rd46, %rd44;\n"
|
|
" .loc 16 151 0\n"
|
|
" st.global.f32 [%rd47+0], %f33;\n"
|
|
" .loc 16 152 0\n"
|
|
" add.u64 %rd44, %rd46, %rd47;\n"
|
|
"$Lt_0_26370:\n"
|
|
" ld.param.s32 %r78, [__cudaparm_kernel_pair_vflag];\n"
|
|
" mov.u32 %r79, 0;\n"
|
|
" setp.le.s32 %p17, %r78, %r79;\n"
|
|
" @%p17 bra $Lt_0_26882;\n"
|
|
" .loc 16 156 0\n"
|
|
" mov.f32 %f123, %f6;\n"
|
|
" st.global.f32 [%rd44+0], %f123;\n"
|
|
" .loc 16 157 0\n"
|
|
" cvt.s64.s32 %rd48, %r9;\n"
|
|
" mul.wide.s32 %rd49, %r9, 4;\n"
|
|
" add.u64 %rd50, %rd49, %rd44;\n"
|
|
" .loc 16 156 0\n"
|
|
" mov.f32 %f124, %f8;\n"
|
|
" st.global.f32 [%rd50+0], %f124;\n"
|
|
" .loc 16 157 0\n"
|
|
" add.u64 %rd51, %rd49, %rd50;\n"
|
|
" .loc 16 156 0\n"
|
|
" mov.f32 %f125, %f10;\n"
|
|
" st.global.f32 [%rd51+0], %f125;\n"
|
|
" .loc 16 157 0\n"
|
|
" add.u64 %rd52, %rd49, %rd51;\n"
|
|
" .loc 16 156 0\n"
|
|
" mov.f32 %f126, %f12;\n"
|
|
" st.global.f32 [%rd52+0], %f126;\n"
|
|
" .loc 16 157 0\n"
|
|
" add.u64 %rd44, %rd49, %rd52;\n"
|
|
" .loc 16 156 0\n"
|
|
" mov.f32 %f127, %f14;\n"
|
|
" st.global.f32 [%rd44+0], %f127;\n"
|
|
" mov.f32 %f128, %f16;\n"
|
|
" add.u64 %rd53, %rd49, %rd44;\n"
|
|
" st.global.f32 [%rd53+0], %f128;\n"
|
|
"$Lt_0_26882:\n"
|
|
" .loc 16 160 0\n"
|
|
" ld.param.u64 %rd54, [__cudaparm_kernel_pair_ans];\n"
|
|
" mul.lo.u64 %rd55, %rd2, 16;\n"
|
|
" add.u64 %rd56, %rd54, %rd55;\n"
|
|
" mov.f32 %f129, %f130;\n"
|
|
" st.global.v4.f32 [%rd56+0], {%f32,%f31,%f30,%f129};\n"
|
|
"$Lt_0_25858:\n"
|
|
"$Lt_0_18690:\n"
|
|
" .loc 16 163 0\n"
|
|
" exit;\n"
|
|
"$LDWend_kernel_pair:\n"
|
|
" }\n"
|
|
" .entry kernel_pair_fast (\n"
|
|
" .param .u64 __cudaparm_kernel_pair_fast_x_,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_fast_lj1_in,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_fast_lj3_in,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_fast_sp_cl_in,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_fast_dev_nbor,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_fast_dev_packed,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_fast_ans,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_fast_engv,\n"
|
|
" .param .s32 __cudaparm_kernel_pair_fast_eflag,\n"
|
|
" .param .s32 __cudaparm_kernel_pair_fast_vflag,\n"
|
|
" .param .s32 __cudaparm_kernel_pair_fast_inum,\n"
|
|
" .param .s32 __cudaparm_kernel_pair_fast_nbor_pitch,\n"
|
|
" .param .u64 __cudaparm_kernel_pair_fast_q_,\n"
|
|
" .param .f32 __cudaparm_kernel_pair_fast_cut_coulsq,\n"
|
|
" .param .f32 __cudaparm_kernel_pair_fast_qqrd2e,\n"
|
|
" .param .f32 __cudaparm_kernel_pair_fast_g_ewald,\n"
|
|
" .param .s32 __cudaparm_kernel_pair_fast_t_per_atom)\n"
|
|
" {\n"
|
|
" .reg .u32 %r<82>;\n"
|
|
" .reg .u64 %rd<62>;\n"
|
|
" .reg .f32 %f<129>;\n"
|
|
" .reg .pred %p<20>;\n"
|
|
" .shared .align 4 .b8 __cuda___cuda_local_var_32678_33_non_const_sp_cl3304[16];\n"
|
|
" .shared .align 4 .b8 __cuda___cuda_local_var_32748_37_non_const_red_acc3320[3072];\n"
|
|
" .loc 16 173 0\n"
|
|
"$LDWbegin_kernel_pair_fast:\n"
|
|
" cvt.s32.u32 %r1, %tid.x;\n"
|
|
" mov.u32 %r2, 3;\n"
|
|
" setp.gt.s32 %p1, %r1, %r2;\n"
|
|
" @%p1 bra $Lt_1_19458;\n"
|
|
" .loc 16 179 0\n"
|
|
" mov.u64 %rd1, __cuda___cuda_local_var_32678_33_non_const_sp_cl3304;\n"
|
|
" cvt.s64.s32 %rd2, %r1;\n"
|
|
" mul.wide.s32 %rd3, %r1, 4;\n"
|
|
" ld.param.u64 %rd4, [__cudaparm_kernel_pair_fast_sp_cl_in];\n"
|
|
" add.u64 %rd5, %rd4, %rd3;\n"
|
|
" ld.global.f32 %f1, [%rd5+0];\n"
|
|
" add.u64 %rd6, %rd3, %rd1;\n"
|
|
" st.shared.f32 [%rd6+0], %f1;\n"
|
|
"$Lt_1_19458:\n"
|
|
" mov.u64 %rd1, __cuda___cuda_local_var_32678_33_non_const_sp_cl3304;\n"
|
|
" .loc 16 186 0\n"
|
|
" mov.f32 %f2, 0f00000000; \n"
|
|
" mov.f32 %f3, %f2;\n"
|
|
" mov.f32 %f4, 0f00000000; \n"
|
|
" mov.f32 %f5, %f4;\n"
|
|
" mov.f32 %f6, 0f00000000; \n"
|
|
" mov.f32 %f7, %f6;\n"
|
|
" mov.f32 %f8, 0f00000000; \n"
|
|
" mov.f32 %f9, %f8;\n"
|
|
" mov.f32 %f10, 0f00000000; \n"
|
|
" mov.f32 %f11, %f10;\n"
|
|
" mov.f32 %f12, 0f00000000; \n"
|
|
" mov.f32 %f13, %f12;\n"
|
|
" .loc 16 188 0\n"
|
|
" bar.sync 0;\n"
|
|
" ld.param.s32 %r3, [__cudaparm_kernel_pair_fast_t_per_atom];\n"
|
|
" div.s32 %r4, %r1, %r3;\n"
|
|
" cvt.s32.u32 %r5, %ntid.x;\n"
|
|
" div.s32 %r6, %r5, %r3;\n"
|
|
" cvt.s32.u32 %r7, %ctaid.x;\n"
|
|
" mul.lo.s32 %r8, %r7, %r6;\n"
|
|
" add.s32 %r9, %r4, %r8;\n"
|
|
" ld.param.s32 %r10, [__cudaparm_kernel_pair_fast_inum];\n"
|
|
" setp.ge.s32 %p2, %r9, %r10;\n"
|
|
" @%p2 bra $Lt_1_27138;\n"
|
|
" .loc 16 193 0\n"
|
|
" cvt.s64.s32 %rd7, %r9;\n"
|
|
" mul.wide.s32 %rd8, %r9, 4;\n"
|
|
" ld.param.u64 %rd9, [__cudaparm_kernel_pair_fast_dev_nbor];\n"
|
|
" add.u64 %rd10, %rd8, %rd9;\n"
|
|
" ld.global.s32 %r11, [%rd10+0];\n"
|
|
" ld.param.s32 %r12, [__cudaparm_kernel_pair_fast_nbor_pitch];\n"
|
|
" cvt.s64.s32 %rd11, %r12;\n"
|
|
" mul.wide.s32 %rd12, %r12, 4;\n"
|
|
" add.u64 %rd13, %rd12, %rd10;\n"
|
|
" ld.global.s32 %r13, [%rd13+0];\n"
|
|
" sub.s32 %r14, %r3, 1;\n"
|
|
" and.b32 %r15, %r14, %r1;\n"
|
|
" cvt.s64.s32 %rd14, %r15;\n"
|
|
" mul.wide.s32 %rd15, %r15, 4;\n"
|
|
" ld.param.u64 %rd16, [__cudaparm_kernel_pair_fast_dev_packed];\n"
|
|
" setp.ne.u64 %p3, %rd16, %rd9;\n"
|
|
" @%p3 bra $Lt_1_20738;\n"
|
|
" cvt.s32.s64 %r16, %rd11;\n"
|
|
" mul.lo.s32 %r17, %r16, %r3;\n"
|
|
" mov.s32 %r18, %r17;\n"
|
|
" mul.lo.s32 %r19, %r14, %r9;\n"
|
|
" add.s32 %r20, %r16, %r19;\n"
|
|
" cvt.s64.s32 %rd17, %r20;\n"
|
|
" mul.wide.s32 %rd18, %r20, 4;\n"
|
|
" add.u64 %rd19, %rd13, %rd18;\n"
|
|
" and.b32 %r21, %r14, %r13;\n"
|
|
" cvt.s64.s32 %rd20, %r21;\n"
|
|
" div.s32 %r22, %r13, %r3;\n"
|
|
" mul.lo.s32 %r23, %r17, %r22;\n"
|
|
" cvt.s64.s32 %rd21, %r23;\n"
|
|
" add.u64 %rd22, %rd20, %rd21;\n"
|
|
" mul.lo.u64 %rd23, %rd22, 4;\n"
|
|
" add.u64 %rd24, %rd19, %rd23;\n"
|
|
" add.u64 %rd25, %rd15, %rd19;\n"
|
|
" bra.uni $Lt_1_20482;\n"
|
|
"$Lt_1_20738:\n"
|
|
" add.u64 %rd26, %rd12, %rd13;\n"
|
|
" ld.global.s32 %r24, [%rd26+0];\n"
|
|
" cvt.s64.s32 %rd27, %r24;\n"
|
|
" mul.wide.s32 %rd28, %r24, 4;\n"
|
|
" add.u64 %rd29, %rd16, %rd28;\n"
|
|
" cvt.s64.s32 %rd30, %r13;\n"
|
|
" mul.wide.s32 %rd31, %r13, 4;\n"
|
|
" add.u64 %rd24, %rd29, %rd31;\n"
|
|
" mov.s32 %r18, %r3;\n"
|
|
" add.u64 %rd25, %rd15, %rd29;\n"
|
|
"$Lt_1_20482:\n"
|
|
" .loc 16 196 0\n"
|
|
" mov.u32 %r25, %r11;\n"
|
|
" mov.s32 %r26, 0;\n"
|
|
" mov.u32 %r27, %r26;\n"
|
|
" mov.s32 %r28, 0;\n"
|
|
" mov.u32 %r29, %r28;\n"
|
|
" mov.s32 %r30, 0;\n"
|
|
" mov.u32 %r31, %r30;\n"
|
|
" tex.1d.v4.f32.s32 {%f14,%f15,%f16,%f17},[pos_tex,{%r25,%r27,%r29,%r31}];\n"
|
|
" mov.f32 %f18, %f14;\n"
|
|
" mov.f32 %f19, %f15;\n"
|
|
" mov.f32 %f20, %f16;\n"
|
|
" .loc 16 197 0\n"
|
|
" mov.u32 %r32, %r11;\n"
|
|
" mov.s32 %r33, 0;\n"
|
|
" mov.u32 %r34, %r33;\n"
|
|
" mov.s32 %r35, 0;\n"
|
|
" mov.u32 %r36, %r35;\n"
|
|
" mov.s32 %r37, 0;\n"
|
|
" mov.u32 %r38, %r37;\n"
|
|
" tex.1d.v4.f32.s32 {%f21,%f22,%f23,%f24},[q_tex,{%r32,%r34,%r36,%r38}];\n"
|
|
" mov.f32 %f25, %f21;\n"
|
|
" setp.ge.u64 %p4, %rd25, %rd24;\n"
|
|
" @%p4 bra $Lt_1_28674;\n"
|
|
" cvt.s64.s32 %rd32, %r18;\n"
|
|
" ld.param.f32 %f26, [__cudaparm_kernel_pair_fast_cut_coulsq];\n"
|
|
" mov.f32 %f27, 0f00000000; \n"
|
|
" mov.f32 %f28, 0f00000000; \n"
|
|
" mov.f32 %f29, 0f00000000; \n"
|
|
" mov.f32 %f30, 0f00000000; \n"
|
|
"$Lt_1_21506:\n"
|
|
" .loc 16 200 0\n"
|
|
" ld.global.s32 %r39, [%rd25+0];\n"
|
|
" .loc 16 203 0\n"
|
|
" mov.f32 %f31, 0f3f800000; \n"
|
|
" shr.s32 %r40, %r39, 30;\n"
|
|
" and.b32 %r41, %r40, 3;\n"
|
|
" cvt.s64.s32 %rd33, %r41;\n"
|
|
" mul.wide.s32 %rd34, %r41, 4;\n"
|
|
" add.u64 %rd35, %rd1, %rd34;\n"
|
|
" ld.shared.f32 %f32, [%rd35+0];\n"
|
|
" sub.ftz.f32 %f33, %f31, %f32;\n"
|
|
" .loc 16 206 0\n"
|
|
" and.b32 %r42, %r39, 1073741823;\n"
|
|
" mov.u32 %r43, %r42;\n"
|
|
" mov.s32 %r44, 0;\n"
|
|
" mov.u32 %r45, %r44;\n"
|
|
" mov.s32 %r46, 0;\n"
|
|
" mov.u32 %r47, %r46;\n"
|
|
" mov.s32 %r48, 0;\n"
|
|
" mov.u32 %r49, %r48;\n"
|
|
" tex.1d.v4.f32.s32 {%f34,%f35,%f36,%f37},[pos_tex,{%r43,%r45,%r47,%r49}];\n"
|
|
" mov.f32 %f38, %f34;\n"
|
|
" mov.f32 %f39, %f35;\n"
|
|
" mov.f32 %f40, %f36;\n"
|
|
" sub.ftz.f32 %f41, %f19, %f39;\n"
|
|
" sub.ftz.f32 %f42, %f18, %f38;\n"
|
|
" sub.ftz.f32 %f43, %f20, %f40;\n"
|
|
" mul.ftz.f32 %f44, %f41, %f41;\n"
|
|
" fma.rn.ftz.f32 %f45, %f42, %f42, %f44;\n"
|
|
" fma.rn.ftz.f32 %f46, %f43, %f43, %f45;\n"
|
|
" setp.lt.ftz.f32 %p5, %f46, %f26;\n"
|
|
" @!%p5 bra $Lt_1_22274;\n"
|
|
" .loc 20 518 0\n"
|
|
" rcp.approx.ftz.f32 %f47, %f46;\n"
|
|
" rsqrt.approx.ftz.f32 %f48, %f47;\n"
|
|
" ld.param.f32 %f49, [__cudaparm_kernel_pair_fast_g_ewald];\n"
|
|
" mul.ftz.f32 %f50, %f49, %f48;\n"
|
|
" mul.ftz.f32 %f51, %f50, %f50;\n"
|
|
" neg.ftz.f32 %f52, %f51;\n"
|
|
" mov.f32 %f53, 0f3fb8aa3b; \n"
|
|
" mul.ftz.f32 %f54, %f52, %f53;\n"
|
|
" ex2.approx.ftz.f32 %f55, %f54;\n"
|
|
" .loc 16 222 0\n"
|
|
" mov.f32 %f56, 0f3f800000; \n"
|
|
" mov.f32 %f57, 0f3ea7ba05; \n"
|
|
" fma.rn.ftz.f32 %f58, %f57, %f50, %f56;\n"
|
|
" rcp.approx.ftz.f32 %f59, %f58;\n"
|
|
" mov.f32 %f60, 0f3e827906; \n"
|
|
" mov.f32 %f61, 0fbe91a98e; \n"
|
|
" mov.f32 %f62, 0f3fb5f0e3; \n"
|
|
" mov.f32 %f63, 0fbfba00e3; \n"
|
|
" mov.f32 %f64, 0f3f87dc22; \n"
|
|
" fma.rn.ftz.f32 %f65, %f64, %f59, %f63;\n"
|
|
" fma.rn.ftz.f32 %f66, %f59, %f65, %f62;\n"
|
|
" fma.rn.ftz.f32 %f67, %f59, %f66, %f61;\n"
|
|
" fma.rn.ftz.f32 %f68, %f59, %f67, %f60;\n"
|
|
" mul.ftz.f32 %f69, %f59, %f68;\n"
|
|
" mul.ftz.f32 %f70, %f55, %f69;\n"
|
|
" .loc 16 223 0\n"
|
|
" mov.u32 %r50, %r42;\n"
|
|
" mov.s32 %r51, 0;\n"
|
|
" mov.u32 %r52, %r51;\n"
|
|
" mov.s32 %r53, 0;\n"
|
|
" mov.u32 %r54, %r53;\n"
|
|
" mov.s32 %r55, 0;\n"
|
|
" mov.u32 %r56, %r55;\n"
|
|
" tex.1d.v4.f32.s32 {%f71,%f72,%f73,%f74},[q_tex,{%r50,%r52,%r54,%r56}];\n"
|
|
" mov.f32 %f75, %f71;\n"
|
|
" .loc 16 224 0\n"
|
|
" ld.param.f32 %f76, [__cudaparm_kernel_pair_fast_qqrd2e];\n"
|
|
" mul.ftz.f32 %f77, %f76, %f25;\n"
|
|
" mul.ftz.f32 %f78, %f77, %f75;\n"
|
|
" div.approx.ftz.f32 %f79, %f78, %f48;\n"
|
|
" mov.f32 %f80, 0f3f906ebb; \n"
|
|
" mul.ftz.f32 %f81, %f50, %f80;\n"
|
|
" fma.rn.ftz.f32 %f82, %f55, %f81, %f70;\n"
|
|
" sub.ftz.f32 %f83, %f82, %f33;\n"
|
|
" mul.ftz.f32 %f84, %f79, %f83;\n"
|
|
" mul.ftz.f32 %f85, %f47, %f84;\n"
|
|
" .loc 16 226 0\n"
|
|
" fma.rn.ftz.f32 %f29, %f42, %f85, %f29;\n"
|
|
" .loc 16 227 0\n"
|
|
" fma.rn.ftz.f32 %f28, %f41, %f85, %f28;\n"
|
|
" .loc 16 228 0\n"
|
|
" fma.rn.ftz.f32 %f27, %f43, %f85, %f27;\n"
|
|
" .loc 16 215 0\n"
|
|
" sub.ftz.f32 %f86, %f70, %f33;\n"
|
|
" fma.rn.ftz.f32 %f87, %f79, %f86, %f30;\n"
|
|
" ld.param.s32 %r57, [__cudaparm_kernel_pair_fast_eflag];\n"
|
|
" mov.s32 %r58, 0;\n"
|
|
" setp.gt.s32 %p6, %r57, %r58;\n"
|
|
" selp.f32 %f30, %f87, %f30, %p6;\n"
|
|
" ld.param.s32 %r59, [__cudaparm_kernel_pair_fast_vflag];\n"
|
|
" mov.u32 %r60, 0;\n"
|
|
" setp.le.s32 %p7, %r59, %r60;\n"
|
|
" @%p7 bra $Lt_1_22274;\n"
|
|
" .loc 16 234 0\n"
|
|
" mov.f32 %f88, %f3;\n"
|
|
" mul.ftz.f32 %f89, %f42, %f42;\n"
|
|
" fma.rn.ftz.f32 %f90, %f85, %f89, %f88;\n"
|
|
" mov.f32 %f3, %f90;\n"
|
|
" .loc 16 235 0\n"
|
|
" mov.f32 %f91, %f5;\n"
|
|
" fma.rn.ftz.f32 %f92, %f85, %f44, %f91;\n"
|
|
" mov.f32 %f5, %f92;\n"
|
|
" .loc 16 236 0\n"
|
|
" mov.f32 %f93, %f7;\n"
|
|
" mul.ftz.f32 %f94, %f43, %f43;\n"
|
|
" fma.rn.ftz.f32 %f95, %f85, %f94, %f93;\n"
|
|
" mov.f32 %f7, %f95;\n"
|
|
" .loc 16 237 0\n"
|
|
" mov.f32 %f96, %f9;\n"
|
|
" mul.ftz.f32 %f97, %f41, %f42;\n"
|
|
" fma.rn.ftz.f32 %f98, %f85, %f97, %f96;\n"
|
|
" mov.f32 %f9, %f98;\n"
|
|
" .loc 16 238 0\n"
|
|
" mov.f32 %f99, %f11;\n"
|
|
" mul.ftz.f32 %f100, %f42, %f43;\n"
|
|
" fma.rn.ftz.f32 %f101, %f85, %f100, %f99;\n"
|
|
" mov.f32 %f11, %f101;\n"
|
|
" .loc 16 239 0\n"
|
|
" mul.ftz.f32 %f102, %f41, %f43;\n"
|
|
" fma.rn.ftz.f32 %f12, %f85, %f102, %f12;\n"
|
|
" mov.f32 %f13, %f12;\n"
|
|
"$Lt_1_22274:\n"
|
|
"$Lt_1_21762:\n"
|
|
" .loc 16 199 0\n"
|
|
" mul.lo.u64 %rd36, %rd32, 4;\n"
|
|
" add.u64 %rd25, %rd25, %rd36;\n"
|
|
" setp.lt.u64 %p8, %rd25, %rd24;\n"
|
|
" @%p8 bra $Lt_1_21506;\n"
|
|
" bra.uni $Lt_1_20994;\n"
|
|
"$Lt_1_28674:\n"
|
|
" mov.f32 %f27, 0f00000000; \n"
|
|
" mov.f32 %f28, 0f00000000; \n"
|
|
" mov.f32 %f29, 0f00000000; \n"
|
|
" mov.f32 %f30, 0f00000000; \n"
|
|
"$Lt_1_20994:\n"
|
|
" mov.u32 %r61, 1;\n"
|
|
" setp.le.s32 %p9, %r3, %r61;\n"
|
|
" @%p9 bra $Lt_1_25090;\n"
|
|
" .loc 16 249 0\n"
|
|
" mov.u64 %rd37, __cuda___cuda_local_var_32748_37_non_const_red_acc3320;\n"
|
|
" cvt.s64.s32 %rd38, %r1;\n"
|
|
" mul.wide.s32 %rd39, %r1, 4;\n"
|
|
" add.u64 %rd40, %rd37, %rd39;\n"
|
|
" mov.f32 %f103, %f29;\n"
|
|
" st.shared.f32 [%rd40+0], %f103;\n"
|
|
" .loc 16 250 0\n"
|
|
" mov.f32 %f104, %f28;\n"
|
|
" st.shared.f32 [%rd40+512], %f104;\n"
|
|
" .loc 16 251 0\n"
|
|
" mov.f32 %f105, %f27;\n"
|
|
" st.shared.f32 [%rd40+1024], %f105;\n"
|
|
" .loc 16 252 0\n"
|
|
" mov.f32 %f106, %f30;\n"
|
|
" st.shared.f32 [%rd40+1536], %f106;\n"
|
|
" .loc 16 254 0\n"
|
|
" shr.s32 %r62, %r3, 31;\n"
|
|
" mov.s32 %r63, 1;\n"
|
|
" and.b32 %r64, %r62, %r63;\n"
|
|
" add.s32 %r65, %r64, %r3;\n"
|
|
" shr.s32 %r66, %r65, 1;\n"
|
|
" mov.s32 %r67, %r66;\n"
|
|
" mov.u32 %r68, 0;\n"
|
|
" setp.ne.u32 %p10, %r66, %r68;\n"
|
|
" @!%p10 bra $Lt_1_23554;\n"
|
|
"$Lt_1_24066:\n"
|
|
" setp.ge.u32 %p11, %r15, %r67;\n"
|
|
" @%p11 bra $Lt_1_24322;\n"
|
|
" .loc 16 257 0\n"
|
|
" add.u32 %r69, %r1, %r67;\n"
|
|
" cvt.u64.u32 %rd41, %r69;\n"
|
|
" mul.wide.u32 %rd42, %r69, 4;\n"
|
|
" add.u64 %rd43, %rd37, %rd42;\n"
|
|
" ld.shared.f32 %f107, [%rd43+0];\n"
|
|
" add.ftz.f32 %f103, %f107, %f103;\n"
|
|
" st.shared.f32 [%rd40+0], %f103;\n"
|
|
" ld.shared.f32 %f108, [%rd43+512];\n"
|
|
" add.ftz.f32 %f104, %f108, %f104;\n"
|
|
" st.shared.f32 [%rd40+512], %f104;\n"
|
|
" ld.shared.f32 %f109, [%rd43+1024];\n"
|
|
" add.ftz.f32 %f105, %f109, %f105;\n"
|
|
" st.shared.f32 [%rd40+1024], %f105;\n"
|
|
" ld.shared.f32 %f110, [%rd43+1536];\n"
|
|
" add.ftz.f32 %f106, %f110, %f106;\n"
|
|
" st.shared.f32 [%rd40+1536], %f106;\n"
|
|
"$Lt_1_24322:\n"
|
|
" .loc 16 254 0\n"
|
|
" shr.u32 %r67, %r67, 1;\n"
|
|
" mov.u32 %r70, 0;\n"
|
|
" setp.ne.u32 %p12, %r67, %r70;\n"
|
|
" @%p12 bra $Lt_1_24066;\n"
|
|
"$Lt_1_23554:\n"
|
|
" .loc 16 261 0\n"
|
|
" mov.f32 %f29, %f103;\n"
|
|
" .loc 16 262 0\n"
|
|
" mov.f32 %f28, %f104;\n"
|
|
" .loc 16 263 0\n"
|
|
" mov.f32 %f27, %f105;\n"
|
|
" .loc 16 264 0\n"
|
|
" mov.f32 %f30, %f106;\n"
|
|
" ld.param.s32 %r71, [__cudaparm_kernel_pair_fast_vflag];\n"
|
|
" mov.u32 %r72, 0;\n"
|
|
" setp.le.s32 %p13, %r71, %r72;\n"
|
|
" @%p13 bra $Lt_1_25090;\n"
|
|
" .loc 16 268 0\n"
|
|
" mov.f32 %f103, %f3;\n"
|
|
" st.shared.f32 [%rd40+0], %f103;\n"
|
|
" mov.f32 %f104, %f5;\n"
|
|
" st.shared.f32 [%rd40+512], %f104;\n"
|
|
" mov.f32 %f105, %f7;\n"
|
|
" st.shared.f32 [%rd40+1024], %f105;\n"
|
|
" mov.f32 %f106, %f9;\n"
|
|
" st.shared.f32 [%rd40+1536], %f106;\n"
|
|
" mov.f32 %f111, %f11;\n"
|
|
" st.shared.f32 [%rd40+2048], %f111;\n"
|
|
" mov.f32 %f112, %f12;\n"
|
|
" st.shared.f32 [%rd40+2560], %f112;\n"
|
|
" .loc 16 270 0\n"
|
|
" mov.s32 %r73, %r66;\n"
|
|
" @!%p10 bra $Lt_1_25602;\n"
|
|
"$Lt_1_26114:\n"
|
|
" setp.ge.u32 %p14, %r15, %r73;\n"
|
|
" @%p14 bra $Lt_1_26370;\n"
|
|
" .loc 16 273 0\n"
|
|
" add.u32 %r74, %r1, %r73;\n"
|
|
" cvt.u64.u32 %rd44, %r74;\n"
|
|
" mul.wide.u32 %rd45, %r74, 4;\n"
|
|
" add.u64 %rd46, %rd37, %rd45;\n"
|
|
" ld.shared.f32 %f113, [%rd46+0];\n"
|
|
" add.ftz.f32 %f103, %f113, %f103;\n"
|
|
" st.shared.f32 [%rd40+0], %f103;\n"
|
|
" ld.shared.f32 %f114, [%rd46+512];\n"
|
|
" add.ftz.f32 %f104, %f114, %f104;\n"
|
|
" st.shared.f32 [%rd40+512], %f104;\n"
|
|
" ld.shared.f32 %f115, [%rd46+1024];\n"
|
|
" add.ftz.f32 %f105, %f115, %f105;\n"
|
|
" st.shared.f32 [%rd40+1024], %f105;\n"
|
|
" ld.shared.f32 %f116, [%rd46+1536];\n"
|
|
" add.ftz.f32 %f106, %f116, %f106;\n"
|
|
" st.shared.f32 [%rd40+1536], %f106;\n"
|
|
" ld.shared.f32 %f117, [%rd46+2048];\n"
|
|
" add.ftz.f32 %f111, %f117, %f111;\n"
|
|
" st.shared.f32 [%rd40+2048], %f111;\n"
|
|
" ld.shared.f32 %f118, [%rd46+2560];\n"
|
|
" add.ftz.f32 %f112, %f118, %f112;\n"
|
|
" st.shared.f32 [%rd40+2560], %f112;\n"
|
|
"$Lt_1_26370:\n"
|
|
" .loc 16 270 0\n"
|
|
" shr.u32 %r73, %r73, 1;\n"
|
|
" mov.u32 %r75, 0;\n"
|
|
" setp.ne.u32 %p15, %r73, %r75;\n"
|
|
" @%p15 bra $Lt_1_26114;\n"
|
|
"$Lt_1_25602:\n"
|
|
" .loc 16 278 0\n"
|
|
" mov.f32 %f3, %f103;\n"
|
|
" mov.f32 %f5, %f104;\n"
|
|
" mov.f32 %f7, %f105;\n"
|
|
" mov.f32 %f9, %f106;\n"
|
|
" mov.f32 %f11, %f111;\n"
|
|
" mov.f32 %f13, %f112;\n"
|
|
"$Lt_1_25090:\n"
|
|
"$Lt_1_23042:\n"
|
|
" mov.u32 %r76, 0;\n"
|
|
" setp.ne.s32 %p16, %r15, %r76;\n"
|
|
" @%p16 bra $Lt_1_27138;\n"
|
|
" .loc 16 284 0\n"
|
|
" ld.param.u64 %rd47, [__cudaparm_kernel_pair_fast_engv];\n"
|
|
" add.u64 %rd48, %rd47, %rd8;\n"
|
|
" ld.param.s32 %r77, [__cudaparm_kernel_pair_fast_eflag];\n"
|
|
" mov.u32 %r78, 0;\n"
|
|
" setp.le.s32 %p17, %r77, %r78;\n"
|
|
" @%p17 bra $Lt_1_27650;\n"
|
|
" .loc 16 286 0\n"
|
|
" mov.f32 %f119, 0f00000000; \n"
|
|
" st.global.f32 [%rd48+0], %f119;\n"
|
|
" .loc 16 287 0\n"
|
|
" cvt.s64.s32 %rd49, %r10;\n"
|
|
" mul.wide.s32 %rd50, %r10, 4;\n"
|
|
" add.u64 %rd51, %rd50, %rd48;\n"
|
|
" .loc 16 288 0\n"
|
|
" st.global.f32 [%rd51+0], %f30;\n"
|
|
" .loc 16 289 0\n"
|
|
" add.u64 %rd48, %rd50, %rd51;\n"
|
|
"$Lt_1_27650:\n"
|
|
" ld.param.s32 %r79, [__cudaparm_kernel_pair_fast_vflag];\n"
|
|
" mov.u32 %r80, 0;\n"
|
|
" setp.le.s32 %p18, %r79, %r80;\n"
|
|
" @%p18 bra $Lt_1_28162;\n"
|
|
" .loc 16 293 0\n"
|
|
" mov.f32 %f120, %f3;\n"
|
|
" st.global.f32 [%rd48+0], %f120;\n"
|
|
" .loc 16 294 0\n"
|
|
" cvt.s64.s32 %rd52, %r10;\n"
|
|
" mul.wide.s32 %rd53, %r10, 4;\n"
|
|
" add.u64 %rd54, %rd53, %rd48;\n"
|
|
" .loc 16 293 0\n"
|
|
" mov.f32 %f121, %f5;\n"
|
|
" st.global.f32 [%rd54+0], %f121;\n"
|
|
" .loc 16 294 0\n"
|
|
" add.u64 %rd55, %rd53, %rd54;\n"
|
|
" .loc 16 293 0\n"
|
|
" mov.f32 %f122, %f7;\n"
|
|
" st.global.f32 [%rd55+0], %f122;\n"
|
|
" .loc 16 294 0\n"
|
|
" add.u64 %rd56, %rd53, %rd55;\n"
|
|
" .loc 16 293 0\n"
|
|
" mov.f32 %f123, %f9;\n"
|
|
" st.global.f32 [%rd56+0], %f123;\n"
|
|
" .loc 16 294 0\n"
|
|
" add.u64 %rd48, %rd53, %rd56;\n"
|
|
" .loc 16 293 0\n"
|
|
" mov.f32 %f124, %f11;\n"
|
|
" st.global.f32 [%rd48+0], %f124;\n"
|
|
" mov.f32 %f125, %f13;\n"
|
|
" add.u64 %rd57, %rd53, %rd48;\n"
|
|
" st.global.f32 [%rd57+0], %f125;\n"
|
|
"$Lt_1_28162:\n"
|
|
" .loc 16 297 0\n"
|
|
" ld.param.u64 %rd58, [__cudaparm_kernel_pair_fast_ans];\n"
|
|
" mul.lo.u64 %rd59, %rd7, 16;\n"
|
|
" add.u64 %rd60, %rd58, %rd59;\n"
|
|
" mov.f32 %f126, %f127;\n"
|
|
" st.global.v4.f32 [%rd60+0], {%f29,%f28,%f27,%f126};\n"
|
|
"$Lt_1_27138:\n"
|
|
"$Lt_1_19970:\n"
|
|
" .loc 16 300 0\n"
|
|
" exit;\n"
|
|
"$LDWend_kernel_pair_fast:\n"
|
|
" }\n"
|
|
;
|