Diffusion lammps (efficient_neuronet) Browse

cg_cmm_long.ptx
6b1774dc2052efficient_neuronet

cg_cmm_long.ptx
View Options

			.version 2.3
			.target sm_20
			.address_size 64
			// compiled with /usr/local/cuda/open64/lib//be
			// nvopencc 4.0 built on 2011-05-12

			//-----------------------------------------------------------
			// Compiling /tmp/tmpxft_00009ef5_00000000-9_lal_cg_cmm_long.cpp3.i (/home/sjplimp/ccBI#.OMEXPd)
			//-----------------------------------------------------------

			//-----------------------------------------------------------
			// Options:
			//-----------------------------------------------------------
			// Target:ptx, ISA:sm_20, Endian:little, Pointer Size:64
			// -O3 (Optimization level)
			// -g0 (Debug level)
			// -m2 (Report advisories)
			//-----------------------------------------------------------

			.file 1 "<command-line>"
			.file 2 "/tmp/tmpxft_00009ef5_00000000-8_lal_cg_cmm_long.cudafe2.gpu"
			.file 3 "/usr/lib/gcc/x86_64-redhat-linux/4.4.5/include/stddef.h"
			.file 4 "/usr/local/cuda/include/crt/device_runtime.h"
			.file 5 "/usr/local/cuda/include/host_defines.h"
			.file 6 "/usr/local/cuda/include/builtin_types.h"
			.file 7 "/usr/local/cuda/include/device_types.h"
			.file 8 "/usr/local/cuda/include/driver_types.h"
			.file 9 "/usr/local/cuda/include/surface_types.h"
			.file 10 "/usr/local/cuda/include/texture_types.h"
			.file 11 "/usr/local/cuda/include/vector_types.h"
			.file 12 "/usr/local/cuda/include/device_launch_parameters.h"
			.file 13 "/usr/local/cuda/include/crt/storage_class.h"
			.file 14 "/usr/include/bits/types.h"
			.file 15 "/usr/include/time.h"
			.file 16 "lal_cg_cmm_long.cu"
			.file 17 "/usr/local/cuda/include/common_functions.h"
			.file 18 "/usr/local/cuda/include/math_functions.h"
			.file 19 "/usr/local/cuda/include/math_constants.h"
			.file 20 "/usr/local/cuda/include/device_functions.h"
			.file 21 "/usr/local/cuda/include/sm_11_atomic_functions.h"
			.file 22 "/usr/local/cuda/include/sm_12_atomic_functions.h"
			.file 23 "/usr/local/cuda/include/sm_13_double_functions.h"
			.file 24 "/usr/local/cuda/include/sm_20_atomic_functions.h"
			.file 25 "/usr/local/cuda/include/sm_20_intrinsics.h"
			.file 26 "/usr/local/cuda/include/surface_functions.h"
			.file 27 "/usr/local/cuda/include/texture_fetch_functions.h"
			.file 28 "/usr/local/cuda/include/math_functions_dbl_ptx3.h"

			.global .texref pos_tex;
			.global .texref q_tex;

			.entry kernel_pair (
			.param .u64 __cudaparm_kernel_pair_x_,
			.param .u64 __cudaparm_kernel_pair_lj1,
			.param .u64 __cudaparm_kernel_pair_lj3,
			.param .s32 __cudaparm_kernel_pair_lj_types,
			.param .u64 __cudaparm_kernel_pair_sp_lj_in,
			.param .u64 __cudaparm_kernel_pair_dev_nbor,
			.param .u64 __cudaparm_kernel_pair_dev_packed,
			.param .u64 __cudaparm_kernel_pair_ans,
			.param .u64 __cudaparm_kernel_pair___val_paramengv,
			.param .s32 __cudaparm_kernel_pair_eflag,
			.param .s32 __cudaparm_kernel_pair_vflag,
			.param .s32 __cudaparm_kernel_pair_inum,
			.param .s32 __cudaparm_kernel_pair_nbor_pitch,
			.param .u64 __cudaparm_kernel_pair_q_,
			.param .f32 __cudaparm_kernel_pair_cut_coulsq,
			.param .f32 __cudaparm_kernel_pair_qqrd2e,
			.param .f32 __cudaparm_kernel_pair_g_ewald,
			.param .s32 __cudaparm_kernel_pair_t_per_atom)
			{
			.reg .u32 %r<86>;
			.reg .u64 %rd<66>;
			.reg .f32 %f<175>;
			.reg .pred %p<23>;
			.shared .align 16 .b8 __cuda___cuda_local_var_32541_33_non_const_sp_lj112[32];
			.shared .align 4 .b8 __cuda___cuda_local_var_32642_55_non_const_red_acc144[3072];
			// __cuda_local_var_32553_10_non_const_f = 64
			// __cuda_local_var_32555_9_non_const_virial = 16
			// __cuda_local_var_32588_43_non_const_inv1 = 40
			// __cuda_local_var_32588_49_non_const_inv2 = 44
			// __cuda_local_var_32588_55_non_const_prefactor = 52
			// __cuda_local_var_32588_66_non_const__erfc = 48
			.loc 16 36 0
			$LDWbegin_kernel_pair:
			.loc 16 41 0
			ld.param.u64 %rd1, [__cudaparm_kernel_pair_sp_lj_in];
			ldu.global.f32 %f1, [%rd1+0];
			.loc 16 42 0
			ld.global.f32 %f2, [%rd1+4];
			.loc 16 43 0
			ld.global.f32 %f3, [%rd1+8];
			.loc 16 44 0
			ld.global.f32 %f4, [%rd1+12];
			st.shared.v4.f32 [__cuda___cuda_local_var_32541_33_non_const_sp_lj112+0], {%f1,%f2,%f3,%f4};
			.loc 16 45 0
			ld.global.f32 %f5, [%rd1+16];
			.loc 16 46 0
			ld.global.f32 %f6, [%rd1+20];
			.loc 16 47 0
			ld.global.f32 %f7, [%rd1+24];
			.loc 16 48 0
			ld.global.f32 %f8, [%rd1+28];
			st.shared.v4.f32 [__cuda___cuda_local_var_32541_33_non_const_sp_lj112+16], {%f5,%f6,%f7,%f8};
			.loc 16 56 0
			mov.f32 %f9, 0f00000000; // 0
			mov.f32 %f10, %f9;
			mov.f32 %f11, 0f00000000; // 0
			mov.f32 %f12, %f11;
			mov.f32 %f13, 0f00000000; // 0
			mov.f32 %f14, %f13;
			mov.f32 %f15, 0f00000000; // 0
			mov.f32 %f16, %f15;
			mov.f32 %f17, 0f00000000; // 0
			mov.f32 %f18, %f17;
			mov.f32 %f19, 0f00000000; // 0
			mov.f32 %f20, %f19;
			ld.param.s32 %r1, [__cudaparm_kernel_pair_t_per_atom];
			cvt.s32.u32 %r2, %tid.x;
			div.s32 %r3, %r2, %r1;
			cvt.s32.u32 %r4, %ntid.x;
			div.s32 %r5, %r4, %r1;
			cvt.s32.u32 %r6, %ctaid.x;
			mul.lo.s32 %r7, %r6, %r5;
			add.s32 %r8, %r3, %r7;
			ld.param.s32 %r9, [__cudaparm_kernel_pair_inum];
			setp.ge.s32 %p1, %r8, %r9;
			@%p1 bra $Lt_0_33538;
			.loc 16 61 0
			cvt.s64.s32 %rd2, %r8;
			mul.wide.s32 %rd3, %r8, 4;
			ld.param.u64 %rd4, [__cudaparm_kernel_pair_dev_nbor];
			add.u64 %rd5, %rd3, %rd4;
			ld.global.s32 %r10, [%rd5+0];
			ld.param.s32 %r11, [__cudaparm_kernel_pair_nbor_pitch];
			cvt.s64.s32 %rd6, %r11;
			mul.wide.s32 %rd7, %r11, 4;
			add.u64 %rd8, %rd7, %rd5;
			ld.global.s32 %r12, [%rd8+0];
			sub.s32 %r13, %r1, 1;
			and.b32 %r14, %r13, %r2;
			cvt.s64.s32 %rd9, %r14;
			mul.wide.s32 %rd10, %r14, 4;
			ld.param.u64 %rd11, [__cudaparm_kernel_pair_dev_packed];
			setp.ne.u64 %p2, %rd11, %rd4;
			@%p2 bra $Lt_0_24066;
			cvt.s32.s64 %r15, %rd6;
			mul.lo.s32 %r16, %r15, %r1;
			mov.s32 %r17, %r16;
			mul.lo.s32 %r18, %r13, %r8;
			add.s32 %r19, %r15, %r18;
			cvt.s64.s32 %rd12, %r19;
			mul.wide.s32 %rd13, %r19, 4;
			add.u64 %rd14, %rd8, %rd13;
			and.b32 %r20, %r13, %r12;
			cvt.s64.s32 %rd15, %r20;
			div.s32 %r21, %r12, %r1;
			mul.lo.s32 %r22, %r16, %r21;
			cvt.s64.s32 %rd16, %r22;
			add.u64 %rd17, %rd15, %rd16;
			mul.lo.u64 %rd18, %rd17, 4;
			add.u64 %rd19, %rd14, %rd18;
			add.u64 %rd20, %rd10, %rd14;
			bra.uni $Lt_0_23810;
			$Lt_0_24066:
			add.u64 %rd21, %rd7, %rd8;
			ld.global.s32 %r23, [%rd21+0];
			cvt.s64.s32 %rd22, %r23;
			mul.wide.s32 %rd23, %r23, 4;
			add.u64 %rd24, %rd11, %rd23;
			cvt.s64.s32 %rd25, %r12;
			mul.wide.s32 %rd26, %r12, 4;
			add.u64 %rd19, %rd24, %rd26;
			mov.s32 %r17, %r1;
			add.u64 %rd20, %rd10, %rd24;
			$Lt_0_23810:
			.loc 16 64 0
			mov.u32 %r24, %r10;
			mov.s32 %r25, 0;
			mov.u32 %r26, %r25;
			mov.s32 %r27, 0;
			mov.u32 %r28, %r27;
			mov.s32 %r29, 0;
			mov.u32 %r30, %r29;
			tex.1d.v4.f32.s32 {%f21,%f22,%f23,%f24},[pos_tex,{%r24,%r26,%r28,%r30}];
			mov.f32 %f25, %f21;
			mov.f32 %f26, %f22;
			mov.f32 %f27, %f23;
			mov.f32 %f28, %f24;
			.loc 16 65 0
			mov.u32 %r31, %r10;
			mov.s32 %r32, 0;
			mov.u32 %r33, %r32;
			mov.s32 %r34, 0;
			mov.u32 %r35, %r34;
			mov.s32 %r36, 0;
			mov.u32 %r37, %r36;
			tex.1d.v4.f32.s32 {%f29,%f30,%f31,%f32},[q_tex,{%r31,%r33,%r35,%r37}];
			mov.f32 %f33, %f29;
			setp.ge.u64 %p3, %rd20, %rd19;
			@%p3 bra $Lt_0_35074;
			cvt.rzi.ftz.s32.f32 %r38, %f28;
			cvt.s64.s32 %rd27, %r17;
			ld.param.s32 %r39, [__cudaparm_kernel_pair_lj_types];
			mul.lo.s32 %r40, %r39, %r38;
			ld.param.u64 %rd28, [__cudaparm_kernel_pair_lj1];
			mov.f32 %f34, 0f00000000; // 0
			mov.f32 %f35, 0f00000000; // 0
			mov.f32 %f36, 0f00000000; // 0
			mov.f32 %f37, 0f00000000; // 0
			mov.f32 %f38, 0f00000000; // 0
			mov.u64 %rd29, __cuda___cuda_local_var_32541_33_non_const_sp_lj112;
			$Lt_0_24834:
			//<loop> Loop body line 65, nesting depth: 1, estimated iterations: unknown
			.loc 16 69 0
			ld.global.s32 %r41, [%rd20+0];
			.loc 16 72 0
			shr.s32 %r42, %r41, 30;
			and.b32 %r43, %r42, 3;
			cvt.s64.s32 %rd30, %r43;
			mul.wide.s32 %rd31, %r43, 4;
			add.u64 %rd32, %rd29, %rd31;
			ld.shared.f32 %f39, [%rd32+0];
			.loc 16 73 0
			mov.f32 %f40, 0f3f800000; // 1
			ld.shared.f32 %f41, [%rd32+16];
			sub.ftz.f32 %f42, %f40, %f41;
			.loc 16 76 0
			and.b32 %r44, %r41, 1073741823;
			mov.u32 %r45, %r44;
			mov.s32 %r46, 0;
			mov.u32 %r47, %r46;
			mov.s32 %r48, 0;
			mov.u32 %r49, %r48;
			mov.s32 %r50, 0;
			mov.u32 %r51, %r50;
			tex.1d.v4.f32.s32 {%f43,%f44,%f45,%f46},[pos_tex,{%r45,%r47,%r49,%r51}];
			mov.f32 %f47, %f43;
			mov.f32 %f48, %f44;
			mov.f32 %f49, %f45;
			mov.f32 %f50, %f46;
			cvt.rzi.ftz.s32.f32 %r52, %f50;
			sub.ftz.f32 %f51, %f26, %f48;
			sub.ftz.f32 %f52, %f25, %f47;
			sub.ftz.f32 %f53, %f27, %f49;
			mul.ftz.f32 %f54, %f51, %f51;
			fma.rn.ftz.f32 %f55, %f52, %f52, %f54;
			fma.rn.ftz.f32 %f56, %f53, %f53, %f55;
			add.s32 %r53, %r52, %r40;
			cvt.s64.s32 %rd33, %r53;
			mul.wide.s32 %rd34, %r53, 16;
			add.u64 %rd35, %rd34, %rd28;
			ld.global.f32 %f57, [%rd35+0];
			setp.gt.ftz.f32 %p4, %f57, %f56;
			@!%p4 bra $Lt_0_28674;
			rcp.approx.ftz.f32 %f58, %f56;
			ld.global.f32 %f59, [%rd35+4];
			setp.lt.ftz.f32 %p5, %f56, %f59;
			@!%p5 bra $Lt_0_25858;
			ld.param.u64 %rd36, [__cudaparm_kernel_pair_lj3];
			add.u64 %rd37, %rd36, %rd34;
			ld.global.f32 %f60, [%rd37+0];
			mov.f32 %f61, 0f40000000; // 2
			setp.eq.ftz.f32 %p6, %f60, %f61;
			@!%p6 bra $Lt_0_26370;
			.loc 16 92 0
			mul.ftz.f32 %f62, %f58, %f58;
			mov.f32 %f63, %f62;
			mov.f32 %f64, %f63;
			.loc 16 93 0
			mul.ftz.f32 %f65, %f62, %f62;
			mov.f32 %f66, %f65;
			bra.uni $Lt_0_26626;
			$Lt_0_26370:
			mov.f32 %f67, 0f3f800000; // 1
			setp.eq.ftz.f32 %p7, %f60, %f67;
			@!%p7 bra $Lt_0_26882;
			.loc 16 95 0
			rsqrt.approx.ftz.f32 %f68, %f56;
			mul.ftz.f32 %f69, %f58, %f68;
			mov.f32 %f65, %f69;
			mov.f32 %f66, %f65;
			.loc 16 96 0
			mul.ftz.f32 %f63, %f69, %f69;
			mov.f32 %f64, %f63;
			bra.uni $Lt_0_26626;
			$Lt_0_26882:
			.loc 16 98 0
			mul.ftz.f32 %f70, %f58, %f58;
			mul.ftz.f32 %f71, %f58, %f70;
			mov.f32 %f63, %f71;
			mov.f32 %f64, %f63;
			.loc 16 99 0
			mov.f32 %f65, %f71;
			mov.f32 %f66, %f65;
			$Lt_0_26626:
			$Lt_0_26114:
			.loc 16 101 0
			mul.ftz.f32 %f72, %f39, %f63;
			ld.global.v2.f32 {%f73,%f74}, [%rd35+8];
			mul.ftz.f32 %f75, %f73, %f65;
			sub.ftz.f32 %f76, %f75, %f74;
			mul.ftz.f32 %f77, %f72, %f76;
			bra.uni $Lt_0_25602;
			$Lt_0_25858:
			.loc 16 103 0
			mov.f32 %f77, 0f00000000; // 0
			$Lt_0_25602:
			ld.param.f32 %f78, [__cudaparm_kernel_pair_cut_coulsq];
			setp.gt.ftz.f32 %p8, %f78, %f56;
			@!%p8 bra $Lt_0_27394;
			.loc 20 518 0
			rsqrt.approx.ftz.f32 %f79, %f58;
			ld.param.f32 %f80, [__cudaparm_kernel_pair_g_ewald];
			mul.ftz.f32 %f81, %f80, %f79;
			mul.ftz.f32 %f82, %f81, %f81;
			neg.ftz.f32 %f83, %f82;
			mov.f32 %f84, 0f3fb8aa3b; // 1.4427
			mul.ftz.f32 %f85, %f83, %f84;
			ex2.approx.ftz.f32 %f86, %f85;
			.loc 16 110 0
			mov.f32 %f87, 0f3f800000; // 1
			mov.f32 %f88, 0f3ea7ba05; // 0.327591
			fma.rn.ftz.f32 %f89, %f88, %f81, %f87;
			rcp.approx.ftz.f32 %f90, %f89;
			mov.f32 %f91, 0f3e827906; // 0.25483
			mov.f32 %f92, 0fbe91a98e; // -0.284497
			mov.f32 %f93, 0f3fb5f0e3; // 1.42141
			mov.f32 %f94, 0fbfba00e3; // -1.45315
			mov.f32 %f95, 0f3f87dc22; // 1.06141
			fma.rn.ftz.f32 %f96, %f95, %f90, %f94;
			fma.rn.ftz.f32 %f97, %f90, %f96, %f93;
			fma.rn.ftz.f32 %f98, %f90, %f97, %f92;
			fma.rn.ftz.f32 %f99, %f90, %f98, %f91;
			mul.ftz.f32 %f100, %f90, %f99;
			mul.ftz.f32 %f101, %f86, %f100;
			mov.f32 %f102, %f101;
			.loc 16 111 0
			mov.u32 %r54, %r44;
			mov.s32 %r55, 0;
			mov.u32 %r56, %r55;
			mov.s32 %r57, 0;
			mov.u32 %r58, %r57;
			mov.s32 %r59, 0;
			mov.u32 %r60, %r59;
			tex.1d.v4.f32.s32 {%f103,%f104,%f105,%f106},[q_tex,{%r54,%r56,%r58,%r60}];
			mov.f32 %f107, %f103;
			ld.param.f32 %f108, [__cudaparm_kernel_pair_qqrd2e];
			mul.ftz.f32 %f109, %f108, %f33;
			mul.ftz.f32 %f110, %f109, %f107;
			div.approx.ftz.f32 %f111, %f110, %f79;
			mov.f32 %f112, %f111;
			.loc 16 112 0
			mov.f32 %f113, 0f3f906ebb; // 1.12838
			mul.ftz.f32 %f114, %f81, %f113;
			fma.rn.ftz.f32 %f115, %f86, %f114, %f101;
			sub.ftz.f32 %f116, %f115, %f42;
			mul.ftz.f32 %f117, %f111, %f116;
			bra.uni $Lt_0_27138;
			$Lt_0_27394:
			.loc 16 114 0
			mov.f32 %f117, 0f00000000; // 0
			$Lt_0_27138:
			.loc 16 118 0
			add.ftz.f32 %f118, %f117, %f77;
			mul.ftz.f32 %f119, %f118, %f58;
			fma.rn.ftz.f32 %f36, %f52, %f119, %f36;
			.loc 16 119 0
			fma.rn.ftz.f32 %f35, %f51, %f119, %f35;
			.loc 16 120 0
			fma.rn.ftz.f32 %f34, %f53, %f119, %f34;
			ld.param.s32 %r61, [__cudaparm_kernel_pair_eflag];
			mov.u32 %r62, 0;
			setp.le.s32 %p9, %r61, %r62;
			@%p9 bra $Lt_0_28162;
			.loc 16 123 0
			mov.f32 %f120, %f112;
			mov.f32 %f121, %f102;
			sub.ftz.f32 %f122, %f121, %f42;
			fma.rn.ftz.f32 %f123, %f120, %f122, %f37;
			selp.f32 %f37, %f123, %f37, %p8;
			@!%p5 bra $Lt_0_28162;
			.loc 16 126 0
			ld.param.u64 %rd38, [__cudaparm_kernel_pair_lj3];
			add.u64 %rd39, %rd38, %rd34;
			ld.global.v4.f32 {_,%f124,%f125,%f126}, [%rd39+0];
			mov.f32 %f127, %f64;
			mul.ftz.f32 %f128, %f127, %f39;
			mov.f32 %f129, %f66;
			mul.ftz.f32 %f130, %f124, %f129;
			sub.ftz.f32 %f131, %f130, %f125;
			mul.ftz.f32 %f132, %f128, %f131;
			sub.ftz.f32 %f133, %f132, %f126;
			add.ftz.f32 %f38, %f38, %f133;
			$Lt_0_28162:
			$Lt_0_27650:
			ld.param.s32 %r63, [__cudaparm_kernel_pair_vflag];
			mov.u32 %r64, 0;
			setp.le.s32 %p10, %r63, %r64;
			@%p10 bra $Lt_0_28674;
			.loc 16 131 0
			mov.f32 %f134, %f10;
			mul.ftz.f32 %f135, %f52, %f52;
			fma.rn.ftz.f32 %f136, %f119, %f135, %f134;
			mov.f32 %f10, %f136;
			.loc 16 132 0
			mov.f32 %f137, %f12;
			fma.rn.ftz.f32 %f138, %f119, %f54, %f137;
			mov.f32 %f12, %f138;
			.loc 16 133 0
			mov.f32 %f139, %f14;
			mul.ftz.f32 %f140, %f53, %f53;
			fma.rn.ftz.f32 %f141, %f119, %f140, %f139;
			mov.f32 %f14, %f141;
			.loc 16 134 0
			mov.f32 %f142, %f16;
			mul.ftz.f32 %f143, %f51, %f52;
			fma.rn.ftz.f32 %f144, %f119, %f143, %f142;
			mov.f32 %f16, %f144;
			.loc 16 135 0
			mov.f32 %f145, %f18;
			mul.ftz.f32 %f146, %f52, %f53;
			fma.rn.ftz.f32 %f147, %f119, %f146, %f145;
			mov.f32 %f18, %f147;
			.loc 16 136 0
			mul.ftz.f32 %f148, %f51, %f53;
			fma.rn.ftz.f32 %f19, %f119, %f148, %f19;
			mov.f32 %f20, %f19;
			$Lt_0_28674:
			$Lt_0_25090:
			.loc 16 68 0
			mul.lo.u64 %rd40, %rd27, 4;
			add.u64 %rd20, %rd20, %rd40;
			setp.lt.u64 %p11, %rd20, %rd19;
			@%p11 bra $Lt_0_24834;
			bra.uni $Lt_0_24322;
			$Lt_0_35074:
			mov.f32 %f34, 0f00000000; // 0
			mov.f32 %f35, 0f00000000; // 0
			mov.f32 %f36, 0f00000000; // 0
			mov.f32 %f37, 0f00000000; // 0
			mov.f32 %f38, 0f00000000; // 0
			$Lt_0_24322:
			mov.u32 %r65, 1;
			setp.le.s32 %p12, %r1, %r65;
			@%p12 bra $Lt_0_31490;
			.loc 16 141 0
			mov.u64 %rd41, __cuda___cuda_local_var_32642_55_non_const_red_acc144;
			cvt.s64.s32 %rd42, %r2;
			mul.wide.s32 %rd43, %r2, 4;
			add.u64 %rd44, %rd41, %rd43;
			mov.f32 %f149, %f36;
			st.shared.f32 [%rd44+0], %f149;
			mov.f32 %f150, %f35;
			st.shared.f32 [%rd44+512], %f150;
			mov.f32 %f151, %f34;
			st.shared.f32 [%rd44+1024], %f151;
			mov.f32 %f152, %f38;
			st.shared.f32 [%rd44+1536], %f152;
			mov.f32 %f153, %f37;
			st.shared.f32 [%rd44+2048], %f153;
			shr.s32 %r66, %r1, 31;
			mov.s32 %r67, 1;
			and.b32 %r68, %r66, %r67;
			add.s32 %r69, %r68, %r1;
			shr.s32 %r70, %r69, 1;
			mov.s32 %r71, %r70;
			mov.u32 %r72, 0;
			setp.ne.u32 %p13, %r70, %r72;
			@!%p13 bra $Lt_0_29954;
			$Lt_0_30466:
			setp.ge.u32 %p14, %r14, %r71;
			@%p14 bra $Lt_0_30722;
			add.u32 %r73, %r2, %r71;
			cvt.u64.u32 %rd45, %r73;
			mul.wide.u32 %rd46, %r73, 4;
			add.u64 %rd47, %rd41, %rd46;
			ld.shared.f32 %f154, [%rd47+0];
			add.ftz.f32 %f149, %f154, %f149;
			st.shared.f32 [%rd44+0], %f149;
			ld.shared.f32 %f155, [%rd47+512];
			add.ftz.f32 %f150, %f155, %f150;
			st.shared.f32 [%rd44+512], %f150;
			ld.shared.f32 %f156, [%rd47+1024];
			add.ftz.f32 %f151, %f156, %f151;
			st.shared.f32 [%rd44+1024], %f151;
			ld.shared.f32 %f157, [%rd47+1536];
			add.ftz.f32 %f152, %f157, %f152;
			st.shared.f32 [%rd44+1536], %f152;
			ld.shared.f32 %f158, [%rd47+2048];
			add.ftz.f32 %f153, %f158, %f153;
			st.shared.f32 [%rd44+2048], %f153;
			$Lt_0_30722:
			shr.u32 %r71, %r71, 1;
			mov.u32 %r74, 0;
			setp.ne.u32 %p15, %r71, %r74;
			@%p15 bra $Lt_0_30466;
			$Lt_0_29954:
			mov.f32 %f36, %f149;
			mov.f32 %f35, %f150;
			mov.f32 %f34, %f151;
			mov.f32 %f38, %f152;
			mov.f32 %f37, %f153;
			ld.param.s32 %r75, [__cudaparm_kernel_pair_vflag];
			mov.u32 %r76, 0;
			setp.le.s32 %p16, %r75, %r76;
			@%p16 bra $Lt_0_31490;
			mov.f32 %f149, %f10;
			st.shared.f32 [%rd44+0], %f149;
			mov.f32 %f150, %f12;
			st.shared.f32 [%rd44+512], %f150;
			mov.f32 %f151, %f14;
			st.shared.f32 [%rd44+1024], %f151;
			mov.f32 %f152, %f16;
			st.shared.f32 [%rd44+1536], %f152;
			mov.f32 %f153, %f18;
			st.shared.f32 [%rd44+2048], %f153;
			mov.f32 %f159, %f19;
			st.shared.f32 [%rd44+2560], %f159;
			mov.s32 %r77, %r70;
			@!%p13 bra $Lt_0_32002;
			$Lt_0_32514:
			setp.ge.u32 %p17, %r14, %r77;
			@%p17 bra $Lt_0_32770;
			add.u32 %r78, %r2, %r77;
			cvt.u64.u32 %rd48, %r78;
			mul.wide.u32 %rd49, %r78, 4;
			add.u64 %rd50, %rd41, %rd49;
			ld.shared.f32 %f160, [%rd50+0];
			add.ftz.f32 %f149, %f160, %f149;
			st.shared.f32 [%rd44+0], %f149;
			ld.shared.f32 %f161, [%rd50+512];
			add.ftz.f32 %f150, %f161, %f150;
			st.shared.f32 [%rd44+512], %f150;
			ld.shared.f32 %f162, [%rd50+1024];
			add.ftz.f32 %f151, %f162, %f151;
			st.shared.f32 [%rd44+1024], %f151;
			ld.shared.f32 %f163, [%rd50+1536];
			add.ftz.f32 %f152, %f163, %f152;
			st.shared.f32 [%rd44+1536], %f152;
			ld.shared.f32 %f164, [%rd50+2048];
			add.ftz.f32 %f153, %f164, %f153;
			st.shared.f32 [%rd44+2048], %f153;
			ld.shared.f32 %f165, [%rd50+2560];
			add.ftz.f32 %f159, %f165, %f159;
			st.shared.f32 [%rd44+2560], %f159;
			$Lt_0_32770:
			shr.u32 %r77, %r77, 1;
			mov.u32 %r79, 0;
			setp.ne.u32 %p18, %r77, %r79;
			@%p18 bra $Lt_0_32514;
			$Lt_0_32002:
			mov.f32 %f10, %f149;
			mov.f32 %f12, %f150;
			mov.f32 %f14, %f151;
			mov.f32 %f16, %f152;
			mov.f32 %f18, %f153;
			mov.f32 %f20, %f159;
			$Lt_0_31490:
			$Lt_0_29442:
			mov.u32 %r80, 0;
			setp.ne.s32 %p19, %r14, %r80;
			@%p19 bra $Lt_0_33538;
			ld.param.u64 %rd51, [__cudaparm_kernel_pair___val_paramengv];
			add.u64 %rd52, %rd51, %rd3;
			ld.param.s32 %r81, [__cudaparm_kernel_pair_eflag];
			mov.u32 %r82, 0;
			setp.le.s32 %p20, %r81, %r82;
			@%p20 bra $Lt_0_34050;
			st.global.f32 [%rd52+0], %f38;
			cvt.s64.s32 %rd53, %r9;
			mul.wide.s32 %rd54, %r9, 4;
			add.u64 %rd55, %rd54, %rd52;
			st.global.f32 [%rd55+0], %f37;
			add.u64 %rd52, %rd54, %rd55;
			$Lt_0_34050:
			ld.param.s32 %r83, [__cudaparm_kernel_pair_vflag];
			mov.u32 %r84, 0;
			setp.le.s32 %p21, %r83, %r84;
			@%p21 bra $Lt_0_34562;
			mov.f32 %f166, %f10;
			st.global.f32 [%rd52+0], %f166;
			cvt.s64.s32 %rd56, %r9;
			mul.wide.s32 %rd57, %r9, 4;
			add.u64 %rd58, %rd57, %rd52;
			mov.f32 %f167, %f12;
			st.global.f32 [%rd58+0], %f167;
			add.u64 %rd59, %rd57, %rd58;
			mov.f32 %f168, %f14;
			st.global.f32 [%rd59+0], %f168;
			add.u64 %rd60, %rd57, %rd59;
			mov.f32 %f169, %f16;
			st.global.f32 [%rd60+0], %f169;
			add.u64 %rd52, %rd57, %rd60;
			mov.f32 %f170, %f18;
			st.global.f32 [%rd52+0], %f170;
			mov.f32 %f171, %f20;
			add.u64 %rd61, %rd57, %rd52;
			st.global.f32 [%rd61+0], %f171;
			$Lt_0_34562:
			ld.param.u64 %rd62, [__cudaparm_kernel_pair_ans];
			mul.lo.u64 %rd63, %rd2, 16;
			add.u64 %rd64, %rd62, %rd63;
			mov.f32 %f172, %f173;
			st.global.v4.f32 [%rd64+0], {%f36,%f35,%f34,%f172};
			$Lt_0_33538:
			$Lt_0_23298:
			.loc 16 144 0
			exit;
			$LDWend_kernel_pair:
			} // kernel_pair

			.entry kernel_pair_fast (
			.param .u64 __cudaparm_kernel_pair_fast_x_,
			.param .u64 __cudaparm_kernel_pair_fast_lj1_in,
			.param .u64 __cudaparm_kernel_pair_fast_lj3_in,
			.param .u64 __cudaparm_kernel_pair_fast_sp_lj_in,
			.param .u64 __cudaparm_kernel_pair_fast_dev_nbor,
			.param .u64 __cudaparm_kernel_pair_fast_dev_packed,
			.param .u64 __cudaparm_kernel_pair_fast_ans,
			.param .u64 __cudaparm_kernel_pair_fast___val_paramengv,
			.param .s32 __cudaparm_kernel_pair_fast_eflag,
			.param .s32 __cudaparm_kernel_pair_fast_vflag,
			.param .s32 __cudaparm_kernel_pair_fast_inum,
			.param .s32 __cudaparm_kernel_pair_fast_nbor_pitch,
			.param .u64 __cudaparm_kernel_pair_fast_q_,
			.param .f32 __cudaparm_kernel_pair_fast_cut_coulsq,
			.param .f32 __cudaparm_kernel_pair_fast_qqrd2e,
			.param .f32 __cudaparm_kernel_pair_fast_g_ewald,
			.param .s32 __cudaparm_kernel_pair_fast_t_per_atom)
			{
			.reg .u32 %r<86>;
			.reg .u64 %rd<77>;
			.reg .f32 %f<178>;
			.reg .pred %p<25>;
			.shared .align 4 .b8 __cuda___cuda_local_var_32661_33_non_const_sp_lj3320[32];
			.shared .align 16 .b8 __cuda___cuda_local_var_32660_34_non_const_lj33360[1936];
			.shared .align 16 .b8 __cuda___cuda_local_var_32659_34_non_const_lj15296[1936];
			.shared .align 4 .b8 __cuda___cuda_local_var_32762_55_non_const_red_acc7232[3072];
			// __cuda_local_var_32671_10_non_const_f = 64
			// __cuda_local_var_32673_9_non_const_virial = 16
			// __cuda_local_var_32708_43_non_const_inv1 = 40
			// __cuda_local_var_32708_49_non_const_inv2 = 44
			// __cuda_local_var_32708_55_non_const_prefactor = 52
			// __cuda_local_var_32708_66_non_const__erfc = 48
			.loc 16 154 0
			$LDWbegin_kernel_pair_fast:
			cvt.s32.u32 %r1, %tid.x;
			mov.u32 %r2, 7;
			setp.gt.s32 %p1, %r1, %r2;
			@%p1 bra $Lt_1_24834;
			.loc 16 162 0
			mov.u64 %rd1, __cuda___cuda_local_var_32661_33_non_const_sp_lj3320;
			cvt.s64.s32 %rd2, %r1;
			mul.wide.s32 %rd3, %r1, 4;
			ld.param.u64 %rd4, [__cudaparm_kernel_pair_fast_sp_lj_in];
			add.u64 %rd5, %rd4, %rd3;
			ld.global.f32 %f1, [%rd5+0];
			add.u64 %rd6, %rd3, %rd1;
			st.shared.f32 [%rd6+0], %f1;
			$Lt_1_24834:
			mov.u64 %rd1, __cuda___cuda_local_var_32661_33_non_const_sp_lj3320;
			mov.u32 %r3, 120;
			setp.gt.s32 %p2, %r1, %r3;
			@%p2 bra $Lt_1_25346;
			.loc 16 164 0
			mov.u64 %rd7, __cuda___cuda_local_var_32660_34_non_const_lj33360;
			mov.u64 %rd8, __cuda___cuda_local_var_32659_34_non_const_lj15296;
			cvt.s64.s32 %rd9, %r1;
			mul.wide.s32 %rd10, %r1, 16;
			ld.param.u64 %rd11, [__cudaparm_kernel_pair_fast_lj1_in];
			add.u64 %rd12, %rd11, %rd10;
			add.u64 %rd13, %rd10, %rd8;
			ld.global.v4.f32 {%f2,%f3,%f4,%f5}, [%rd12+0];
			st.shared.v4.f32 [%rd13+0], {%f2,%f3,%f4,%f5};
			.loc 16 165 0
			ld.param.u64 %rd14, [__cudaparm_kernel_pair_fast_lj3_in];
			add.u64 %rd15, %rd14, %rd10;
			add.u64 %rd16, %rd10, %rd7;
			ld.global.v4.f32 {%f6,%f7,%f8,%f9}, [%rd15+0];
			st.shared.v4.f32 [%rd16+0], {%f6,%f7,%f8,%f9};
			$Lt_1_25346:
			mov.u64 %rd7, __cuda___cuda_local_var_32660_34_non_const_lj33360;
			mov.u64 %rd8, __cuda___cuda_local_var_32659_34_non_const_lj15296;
			.loc 16 174 0
			mov.f32 %f10, 0f00000000; // 0
			mov.f32 %f11, %f10;
			mov.f32 %f12, 0f00000000; // 0
			mov.f32 %f13, %f12;
			mov.f32 %f14, 0f00000000; // 0
			mov.f32 %f15, %f14;
			mov.f32 %f16, 0f00000000; // 0
			mov.f32 %f17, %f16;
			mov.f32 %f18, 0f00000000; // 0
			mov.f32 %f19, %f18;
			mov.f32 %f20, 0f00000000; // 0
			mov.f32 %f21, %f20;
			.loc 16 176 0
			bar.sync 0;
			ld.param.s32 %r4, [__cudaparm_kernel_pair_fast_t_per_atom];
			div.s32 %r5, %r1, %r4;
			cvt.s32.u32 %r6, %ntid.x;
			div.s32 %r7, %r6, %r4;
			cvt.s32.u32 %r8, %ctaid.x;
			mul.lo.s32 %r9, %r8, %r7;
			add.s32 %r10, %r5, %r9;
			ld.param.s32 %r11, [__cudaparm_kernel_pair_fast_inum];
			setp.ge.s32 %p3, %r10, %r11;
			@%p3 bra $Lt_1_36098;
			.loc 16 181 0
			cvt.s64.s32 %rd17, %r10;
			mul.wide.s32 %rd18, %r10, 4;
			ld.param.u64 %rd19, [__cudaparm_kernel_pair_fast_dev_nbor];
			add.u64 %rd20, %rd18, %rd19;
			ld.global.s32 %r12, [%rd20+0];
			ld.param.s32 %r13, [__cudaparm_kernel_pair_fast_nbor_pitch];
			cvt.s64.s32 %rd21, %r13;
			mul.wide.s32 %rd22, %r13, 4;
			add.u64 %rd23, %rd22, %rd20;
			ld.global.s32 %r14, [%rd23+0];
			sub.s32 %r15, %r4, 1;
			and.b32 %r16, %r15, %r1;
			cvt.s64.s32 %rd24, %r16;
			mul.wide.s32 %rd25, %r16, 4;
			ld.param.u64 %rd26, [__cudaparm_kernel_pair_fast_dev_packed];
			setp.ne.u64 %p4, %rd26, %rd19;
			@%p4 bra $Lt_1_26626;
			cvt.s32.s64 %r17, %rd21;
			mul.lo.s32 %r18, %r17, %r4;
			mov.s32 %r19, %r18;
			mul.lo.s32 %r20, %r15, %r10;
			add.s32 %r21, %r17, %r20;
			cvt.s64.s32 %rd27, %r21;
			mul.wide.s32 %rd28, %r21, 4;
			add.u64 %rd29, %rd23, %rd28;
			and.b32 %r22, %r15, %r14;
			cvt.s64.s32 %rd30, %r22;
			div.s32 %r23, %r14, %r4;
			mul.lo.s32 %r24, %r18, %r23;
			cvt.s64.s32 %rd31, %r24;
			add.u64 %rd32, %rd30, %rd31;
			mul.lo.u64 %rd33, %rd32, 4;
			add.u64 %rd34, %rd29, %rd33;
			add.u64 %rd35, %rd25, %rd29;
			bra.uni $Lt_1_26370;
			$Lt_1_26626:
			add.u64 %rd36, %rd22, %rd23;
			ld.global.s32 %r25, [%rd36+0];
			cvt.s64.s32 %rd37, %r25;
			mul.wide.s32 %rd38, %r25, 4;
			add.u64 %rd39, %rd26, %rd38;
			cvt.s64.s32 %rd40, %r14;
			mul.wide.s32 %rd41, %r14, 4;
			add.u64 %rd34, %rd39, %rd41;
			mov.s32 %r19, %r4;
			add.u64 %rd35, %rd25, %rd39;
			$Lt_1_26370:
			.loc 16 184 0
			mov.u32 %r26, %r12;
			mov.s32 %r27, 0;
			mov.u32 %r28, %r27;
			mov.s32 %r29, 0;
			mov.u32 %r30, %r29;
			mov.s32 %r31, 0;
			mov.u32 %r32, %r31;
			tex.1d.v4.f32.s32 {%f22,%f23,%f24,%f25},[pos_tex,{%r26,%r28,%r30,%r32}];
			mov.f32 %f26, %f22;
			mov.f32 %f27, %f23;
			mov.f32 %f28, %f24;
			mov.f32 %f29, %f25;
			.loc 16 185 0
			mov.u32 %r33, %r12;
			mov.s32 %r34, 0;
			mov.u32 %r35, %r34;
			mov.s32 %r36, 0;
			mov.u32 %r37, %r36;
			mov.s32 %r38, 0;
			mov.u32 %r39, %r38;
			tex.1d.v4.f32.s32 {%f30,%f31,%f32,%f33},[q_tex,{%r33,%r35,%r37,%r39}];
			mov.f32 %f34, %f30;
			setp.ge.u64 %p5, %rd35, %rd34;
			@%p5 bra $Lt_1_37634;
			cvt.rzi.ftz.s32.f32 %r40, %f29;
			cvt.s64.s32 %rd42, %r19;
			mul.lo.s32 %r41, %r40, 11;
			cvt.rn.f32.s32 %f35, %r41;
			mov.f32 %f36, 0f00000000; // 0
			mov.f32 %f37, 0f00000000; // 0
			mov.f32 %f38, 0f00000000; // 0
			mov.f32 %f39, 0f00000000; // 0
			mov.f32 %f40, 0f00000000; // 0
			$Lt_1_27394:
			//<loop> Loop body line 185, nesting depth: 1, estimated iterations: unknown
			.loc 16 190 0
			ld.global.s32 %r42, [%rd35+0];
			.loc 16 193 0
			shr.s32 %r43, %r42, 30;
			and.b32 %r44, %r43, 3;
			cvt.s64.s32 %rd43, %r44;
			mul.wide.s32 %rd44, %r44, 4;
			add.u64 %rd45, %rd1, %rd44;
			ld.shared.f32 %f41, [%rd45+0];
			.loc 16 194 0
			mov.f32 %f42, 0f3f800000; // 1
			ld.shared.f32 %f43, [%rd45+16];
			sub.ftz.f32 %f44, %f42, %f43;
			.loc 16 197 0
			and.b32 %r45, %r42, 1073741823;
			mov.u32 %r46, %r45;
			mov.s32 %r47, 0;
			mov.u32 %r48, %r47;
			mov.s32 %r49, 0;
			mov.u32 %r50, %r49;
			mov.s32 %r51, 0;
			mov.u32 %r52, %r51;
			tex.1d.v4.f32.s32 {%f45,%f46,%f47,%f48},[pos_tex,{%r46,%r48,%r50,%r52}];
			mov.f32 %f49, %f45;
			mov.f32 %f50, %f46;
			mov.f32 %f51, %f47;
			mov.f32 %f52, %f48;
			sub.ftz.f32 %f53, %f27, %f50;
			sub.ftz.f32 %f54, %f26, %f49;
			sub.ftz.f32 %f55, %f28, %f51;
			mul.ftz.f32 %f56, %f53, %f53;
			fma.rn.ftz.f32 %f57, %f54, %f54, %f56;
			fma.rn.ftz.f32 %f58, %f55, %f55, %f57;
			add.ftz.f32 %f59, %f35, %f52;
			cvt.rzi.ftz.s32.f32 %r53, %f59;
			cvt.s64.s32 %rd46, %r53;
			mul.wide.s32 %rd47, %r53, 16;
			add.u64 %rd48, %rd47, %rd8;
			ld.shared.f32 %f60, [%rd48+0];
			setp.gt.ftz.f32 %p6, %f60, %f58;
			@!%p6 bra $Lt_1_31234;
			rcp.approx.ftz.f32 %f61, %f58;
			ld.shared.f32 %f62, [%rd48+4];
			setp.lt.ftz.f32 %p7, %f58, %f62;
			@!%p7 bra $Lt_1_28418;
			add.u64 %rd49, %rd47, %rd7;
			ld.shared.f32 %f63, [%rd49+0];
			mov.f32 %f64, 0f40000000; // 2
			setp.eq.ftz.f32 %p8, %f63, %f64;
			@!%p8 bra $Lt_1_28930;
			.loc 16 212 0
			mul.ftz.f32 %f65, %f61, %f61;
			mov.f32 %f66, %f65;
			mov.f32 %f67, %f66;
			.loc 16 213 0
			mul.ftz.f32 %f68, %f65, %f65;
			mov.f32 %f69, %f68;
			bra.uni $Lt_1_29186;
			$Lt_1_28930:
			mov.f32 %f70, 0f3f800000; // 1
			setp.eq.ftz.f32 %p9, %f63, %f70;
			@!%p9 bra $Lt_1_29442;
			.loc 16 215 0
			rsqrt.approx.ftz.f32 %f71, %f58;
			mul.ftz.f32 %f72, %f61, %f71;
			mov.f32 %f68, %f72;
			mov.f32 %f69, %f68;
			.loc 16 216 0
			mul.ftz.f32 %f66, %f72, %f72;
			mov.f32 %f67, %f66;
			bra.uni $Lt_1_29186;
			$Lt_1_29442:
			.loc 16 218 0
			mul.ftz.f32 %f73, %f61, %f61;
			mul.ftz.f32 %f74, %f61, %f73;
			mov.f32 %f66, %f74;
			mov.f32 %f67, %f66;
			.loc 16 219 0
			mov.f32 %f68, %f74;
			mov.f32 %f69, %f68;
			$Lt_1_29186:
			$Lt_1_28674:
			.loc 16 221 0
			mul.ftz.f32 %f75, %f41, %f66;
			ld.shared.v2.f32 {%f76,%f77}, [%rd48+8];
			mul.ftz.f32 %f78, %f76, %f68;
			sub.ftz.f32 %f79, %f78, %f77;
			mul.ftz.f32 %f80, %f75, %f79;
			bra.uni $Lt_1_28162;
			$Lt_1_28418:
			.loc 16 223 0
			mov.f32 %f80, 0f00000000; // 0
			$Lt_1_28162:
			ld.param.f32 %f81, [__cudaparm_kernel_pair_fast_cut_coulsq];
			setp.gt.ftz.f32 %p10, %f81, %f58;
			@!%p10 bra $Lt_1_29954;
			.loc 20 518 0
			rsqrt.approx.ftz.f32 %f82, %f61;
			ld.param.f32 %f83, [__cudaparm_kernel_pair_fast_g_ewald];
			mul.ftz.f32 %f84, %f83, %f82;
			mul.ftz.f32 %f85, %f84, %f84;
			neg.ftz.f32 %f86, %f85;
			mov.f32 %f87, 0f3fb8aa3b; // 1.4427
			mul.ftz.f32 %f88, %f86, %f87;
			ex2.approx.ftz.f32 %f89, %f88;
			.loc 16 230 0
			mov.f32 %f90, 0f3f800000; // 1
			mov.f32 %f91, 0f3ea7ba05; // 0.327591
			fma.rn.ftz.f32 %f92, %f91, %f84, %f90;
			rcp.approx.ftz.f32 %f93, %f92;
			mov.f32 %f94, 0f3e827906; // 0.25483
			mov.f32 %f95, 0fbe91a98e; // -0.284497
			mov.f32 %f96, 0f3fb5f0e3; // 1.42141
			mov.f32 %f97, 0fbfba00e3; // -1.45315
			mov.f32 %f98, 0f3f87dc22; // 1.06141
			fma.rn.ftz.f32 %f99, %f98, %f93, %f97;
			fma.rn.ftz.f32 %f100, %f93, %f99, %f96;
			fma.rn.ftz.f32 %f101, %f93, %f100, %f95;
			fma.rn.ftz.f32 %f102, %f93, %f101, %f94;
			mul.ftz.f32 %f103, %f93, %f102;
			mul.ftz.f32 %f104, %f89, %f103;
			mov.f32 %f105, %f104;
			.loc 16 231 0
			mov.u32 %r54, %r45;
			mov.s32 %r55, 0;
			mov.u32 %r56, %r55;
			mov.s32 %r57, 0;
			mov.u32 %r58, %r57;
			mov.s32 %r59, 0;
			mov.u32 %r60, %r59;
			tex.1d.v4.f32.s32 {%f106,%f107,%f108,%f109},[q_tex,{%r54,%r56,%r58,%r60}];
			mov.f32 %f110, %f106;
			ld.param.f32 %f111, [__cudaparm_kernel_pair_fast_qqrd2e];
			mul.ftz.f32 %f112, %f111, %f34;
			mul.ftz.f32 %f113, %f112, %f110;
			div.approx.ftz.f32 %f114, %f113, %f82;
			mov.f32 %f115, %f114;
			.loc 16 232 0
			mov.f32 %f116, 0f3f906ebb; // 1.12838
			mul.ftz.f32 %f117, %f84, %f116;
			fma.rn.ftz.f32 %f118, %f89, %f117, %f104;
			sub.ftz.f32 %f119, %f118, %f44;
			mul.ftz.f32 %f120, %f114, %f119;
			bra.uni $Lt_1_29698;
			$Lt_1_29954:
			.loc 16 234 0
			mov.f32 %f120, 0f00000000; // 0
			$Lt_1_29698:
			.loc 16 238 0
			add.ftz.f32 %f121, %f120, %f80;
			mul.ftz.f32 %f122, %f121, %f61;
			fma.rn.ftz.f32 %f38, %f54, %f122, %f38;
			.loc 16 239 0
			fma.rn.ftz.f32 %f37, %f53, %f122, %f37;
			.loc 16 240 0
			fma.rn.ftz.f32 %f36, %f55, %f122, %f36;
			ld.param.s32 %r61, [__cudaparm_kernel_pair_fast_eflag];
			mov.u32 %r62, 0;
			setp.le.s32 %p11, %r61, %r62;
			@%p11 bra $Lt_1_30722;
			.loc 16 243 0
			mov.f32 %f123, %f115;
			mov.f32 %f124, %f105;
			sub.ftz.f32 %f125, %f124, %f44;
			fma.rn.ftz.f32 %f126, %f123, %f125, %f39;
			selp.f32 %f39, %f126, %f39, %p10;
			@!%p7 bra $Lt_1_30722;
			.loc 16 246 0
			add.u64 %rd50, %rd47, %rd7;
			ld.shared.v4.f32 {_,%f127,%f128,%f129}, [%rd50+0];
			mov.f32 %f130, %f67;
			mul.ftz.f32 %f131, %f130, %f41;
			mov.f32 %f132, %f69;
			mul.ftz.f32 %f133, %f127, %f132;
			sub.ftz.f32 %f134, %f133, %f128;
			mul.ftz.f32 %f135, %f131, %f134;
			sub.ftz.f32 %f136, %f135, %f129;
			add.ftz.f32 %f40, %f40, %f136;
			$Lt_1_30722:
			$Lt_1_30210:
			ld.param.s32 %r63, [__cudaparm_kernel_pair_fast_vflag];
			mov.u32 %r64, 0;
			setp.le.s32 %p12, %r63, %r64;
			@%p12 bra $Lt_1_31234;
			.loc 16 251 0
			mov.f32 %f137, %f11;
			mul.ftz.f32 %f138, %f54, %f54;
			fma.rn.ftz.f32 %f139, %f122, %f138, %f137;
			mov.f32 %f11, %f139;
			.loc 16 252 0
			mov.f32 %f140, %f13;
			fma.rn.ftz.f32 %f141, %f122, %f56, %f140;
			mov.f32 %f13, %f141;
			.loc 16 253 0
			mov.f32 %f142, %f15;
			mul.ftz.f32 %f143, %f55, %f55;
			fma.rn.ftz.f32 %f144, %f122, %f143, %f142;
			mov.f32 %f15, %f144;
			.loc 16 254 0
			mov.f32 %f145, %f17;
			mul.ftz.f32 %f146, %f53, %f54;
			fma.rn.ftz.f32 %f147, %f122, %f146, %f145;
			mov.f32 %f17, %f147;
			.loc 16 255 0
			mov.f32 %f148, %f19;
			mul.ftz.f32 %f149, %f54, %f55;
			fma.rn.ftz.f32 %f150, %f122, %f149, %f148;
			mov.f32 %f19, %f150;
			.loc 16 256 0
			mul.ftz.f32 %f151, %f53, %f55;
			fma.rn.ftz.f32 %f20, %f122, %f151, %f20;
			mov.f32 %f21, %f20;
			$Lt_1_31234:
			$Lt_1_27650:
			.loc 16 189 0
			mul.lo.u64 %rd51, %rd42, 4;
			add.u64 %rd35, %rd35, %rd51;
			setp.lt.u64 %p13, %rd35, %rd34;
			@%p13 bra $Lt_1_27394;
			bra.uni $Lt_1_26882;
			$Lt_1_37634:
			mov.f32 %f36, 0f00000000; // 0
			mov.f32 %f37, 0f00000000; // 0
			mov.f32 %f38, 0f00000000; // 0
			mov.f32 %f39, 0f00000000; // 0
			mov.f32 %f40, 0f00000000; // 0
			$Lt_1_26882:
			mov.u32 %r65, 1;
			setp.le.s32 %p14, %r4, %r65;
			@%p14 bra $Lt_1_34050;
			.loc 16 261 0
			mov.u64 %rd52, __cuda___cuda_local_var_32762_55_non_const_red_acc7232;
			cvt.s64.s32 %rd53, %r1;
			mul.wide.s32 %rd54, %r1, 4;
			add.u64 %rd55, %rd52, %rd54;
			mov.f32 %f152, %f38;
			st.shared.f32 [%rd55+0], %f152;
			mov.f32 %f153, %f37;
			st.shared.f32 [%rd55+512], %f153;
			mov.f32 %f154, %f36;
			st.shared.f32 [%rd55+1024], %f154;
			mov.f32 %f155, %f40;
			st.shared.f32 [%rd55+1536], %f155;
			mov.f32 %f156, %f39;
			st.shared.f32 [%rd55+2048], %f156;
			shr.s32 %r66, %r4, 31;
			mov.s32 %r67, 1;
			and.b32 %r68, %r66, %r67;
			add.s32 %r69, %r68, %r4;
			shr.s32 %r70, %r69, 1;
			mov.s32 %r71, %r70;
			mov.u32 %r72, 0;
			setp.ne.u32 %p15, %r70, %r72;
			@!%p15 bra $Lt_1_32514;
			$Lt_1_33026:
			setp.ge.u32 %p16, %r16, %r71;
			@%p16 bra $Lt_1_33282;
			add.u32 %r73, %r1, %r71;
			cvt.u64.u32 %rd56, %r73;
			mul.wide.u32 %rd57, %r73, 4;
			add.u64 %rd58, %rd52, %rd57;
			ld.shared.f32 %f157, [%rd58+0];
			add.ftz.f32 %f152, %f157, %f152;
			st.shared.f32 [%rd55+0], %f152;
			ld.shared.f32 %f158, [%rd58+512];
			add.ftz.f32 %f153, %f158, %f153;
			st.shared.f32 [%rd55+512], %f153;
			ld.shared.f32 %f159, [%rd58+1024];
			add.ftz.f32 %f154, %f159, %f154;
			st.shared.f32 [%rd55+1024], %f154;
			ld.shared.f32 %f160, [%rd58+1536];
			add.ftz.f32 %f155, %f160, %f155;
			st.shared.f32 [%rd55+1536], %f155;
			ld.shared.f32 %f161, [%rd58+2048];
			add.ftz.f32 %f156, %f161, %f156;
			st.shared.f32 [%rd55+2048], %f156;
			$Lt_1_33282:
			shr.u32 %r71, %r71, 1;
			mov.u32 %r74, 0;
			setp.ne.u32 %p17, %r71, %r74;
			@%p17 bra $Lt_1_33026;
			$Lt_1_32514:
			mov.f32 %f38, %f152;
			mov.f32 %f37, %f153;
			mov.f32 %f36, %f154;
			mov.f32 %f40, %f155;
			mov.f32 %f39, %f156;
			ld.param.s32 %r75, [__cudaparm_kernel_pair_fast_vflag];
			mov.u32 %r76, 0;
			setp.le.s32 %p18, %r75, %r76;
			@%p18 bra $Lt_1_34050;
			mov.f32 %f152, %f11;
			st.shared.f32 [%rd55+0], %f152;
			mov.f32 %f153, %f13;
			st.shared.f32 [%rd55+512], %f153;
			mov.f32 %f154, %f15;
			st.shared.f32 [%rd55+1024], %f154;
			mov.f32 %f155, %f17;
			st.shared.f32 [%rd55+1536], %f155;
			mov.f32 %f156, %f19;
			st.shared.f32 [%rd55+2048], %f156;
			mov.f32 %f162, %f20;
			st.shared.f32 [%rd55+2560], %f162;
			mov.s32 %r77, %r70;
			@!%p15 bra $Lt_1_34562;
			$Lt_1_35074:
			setp.ge.u32 %p19, %r16, %r77;
			@%p19 bra $Lt_1_35330;
			add.u32 %r78, %r1, %r77;
			cvt.u64.u32 %rd59, %r78;
			mul.wide.u32 %rd60, %r78, 4;
			add.u64 %rd61, %rd52, %rd60;
			ld.shared.f32 %f163, [%rd61+0];
			add.ftz.f32 %f152, %f163, %f152;
			st.shared.f32 [%rd55+0], %f152;
			ld.shared.f32 %f164, [%rd61+512];
			add.ftz.f32 %f153, %f164, %f153;
			st.shared.f32 [%rd55+512], %f153;
			ld.shared.f32 %f165, [%rd61+1024];
			add.ftz.f32 %f154, %f165, %f154;
			st.shared.f32 [%rd55+1024], %f154;
			ld.shared.f32 %f166, [%rd61+1536];
			add.ftz.f32 %f155, %f166, %f155;
			st.shared.f32 [%rd55+1536], %f155;
			ld.shared.f32 %f167, [%rd61+2048];
			add.ftz.f32 %f156, %f167, %f156;
			st.shared.f32 [%rd55+2048], %f156;
			ld.shared.f32 %f168, [%rd61+2560];
			add.ftz.f32 %f162, %f168, %f162;
			st.shared.f32 [%rd55+2560], %f162;
			$Lt_1_35330:
			shr.u32 %r77, %r77, 1;
			mov.u32 %r79, 0;
			setp.ne.u32 %p20, %r77, %r79;
			@%p20 bra $Lt_1_35074;
			$Lt_1_34562:
			mov.f32 %f11, %f152;
			mov.f32 %f13, %f153;
			mov.f32 %f15, %f154;
			mov.f32 %f17, %f155;
			mov.f32 %f19, %f156;
			mov.f32 %f21, %f162;
			$Lt_1_34050:
			$Lt_1_32002:
			mov.u32 %r80, 0;
			setp.ne.s32 %p21, %r16, %r80;
			@%p21 bra $Lt_1_36098;
			ld.param.u64 %rd62, [__cudaparm_kernel_pair_fast___val_paramengv];
			add.u64 %rd63, %rd62, %rd18;
			ld.param.s32 %r81, [__cudaparm_kernel_pair_fast_eflag];
			mov.u32 %r82, 0;
			setp.le.s32 %p22, %r81, %r82;
			@%p22 bra $Lt_1_36610;
			st.global.f32 [%rd63+0], %f40;
			cvt.s64.s32 %rd64, %r11;
			mul.wide.s32 %rd65, %r11, 4;
			add.u64 %rd66, %rd65, %rd63;
			st.global.f32 [%rd66+0], %f39;
			add.u64 %rd63, %rd65, %rd66;
			$Lt_1_36610:
			ld.param.s32 %r83, [__cudaparm_kernel_pair_fast_vflag];
			mov.u32 %r84, 0;
			setp.le.s32 %p23, %r83, %r84;
			@%p23 bra $Lt_1_37122;
			mov.f32 %f169, %f11;
			st.global.f32 [%rd63+0], %f169;
			cvt.s64.s32 %rd67, %r11;
			mul.wide.s32 %rd68, %r11, 4;
			add.u64 %rd69, %rd68, %rd63;
			mov.f32 %f170, %f13;
			st.global.f32 [%rd69+0], %f170;
			add.u64 %rd70, %rd68, %rd69;
			mov.f32 %f171, %f15;
			st.global.f32 [%rd70+0], %f171;
			add.u64 %rd71, %rd68, %rd70;
			mov.f32 %f172, %f17;
			st.global.f32 [%rd71+0], %f172;
			add.u64 %rd63, %rd68, %rd71;
			mov.f32 %f173, %f19;
			st.global.f32 [%rd63+0], %f173;
			mov.f32 %f174, %f21;
			add.u64 %rd72, %rd68, %rd63;
			st.global.f32 [%rd72+0], %f174;
			$Lt_1_37122:
			ld.param.u64 %rd73, [__cudaparm_kernel_pair_fast_ans];
			mul.lo.u64 %rd74, %rd17, 16;
			add.u64 %rd75, %rd73, %rd74;
			mov.f32 %f175, %f176;
			st.global.v4.f32 [%rd75+0], {%f38,%f37,%f36,%f175};
			$Lt_1_36098:
			$Lt_1_25858:
			.loc 16 264 0
			exit;
			$LDWend_kernel_pair_fast:
			} // kernel_pair_fast

lammps/lib/gpu/cg_cmm_long.ptx6b1774dc2052efficient_neuronet

cg_cmm_long.ptxView Options

lammps/lib/gpu/cg_cmm_long.ptx
6b1774dc2052efficient_neuronet

cg_cmm_long.ptx
View Options