nb_kernel010_x86_64_sse2.s
来自「最著名最快的分子模拟软件」· S 代码 · 共 990 行 · 第 1/2 页
S
990 行
#### $Id: nb_kernel010_x86_64_sse2.s,v 1.4.2.4 2006/09/22 08:40:31 lindahl Exp $#### Gromacs 4.0 Copyright (c) 1991-2003 ## David van der Spoel, Erik Lindahl#### This program is free software; you can redistribute it and/or## modify it under the terms of the GNU General Public License## as published by the Free Software Foundation; either version 2## of the License, or (at your option) any later version.#### To help us fund GROMACS development, we humbly ask that you cite## the research papers on the package. Check out http://www.gromacs.org## ## And Hey:## Gnomes, ROck Monsters And Chili Sauce##.globl nb_kernel010_x86_64_sse2.globl _nb_kernel010_x86_64_sse2nb_kernel010_x86_64_sse2: _nb_kernel010_x86_64_sse2: .set nb010_fshift, 16.set nb010_gid, 24.set nb010_pos, 32.set nb010_faction, 40.set nb010_charge, 48.set nb010_p_facel, 56.set nb010_argkrf, 64.set nb010_argcrf, 72.set nb010_Vc, 80.set nb010_type, 88.set nb010_p_ntype, 96.set nb010_vdwparam, 104.set nb010_Vvdw, 112.set nb010_p_tabscale, 120.set nb010_VFtab, 128.set nb010_invsqrta, 136.set nb010_dvda, 144.set nb010_p_gbtabscale, 152.set nb010_GBtab, 160.set nb010_p_nthreads, 168.set nb010_count, 176.set nb010_mtx, 184.set nb010_outeriter, 192.set nb010_inneriter, 200.set nb010_work, 208 ## stack offsets for local variables ## bottom of stack is cache-aligned for sse use .set nb010_ix, 0.set nb010_iy, 16.set nb010_iz, 32.set nb010_dx, 48.set nb010_dy, 64.set nb010_dz, 80.set nb010_two, 96.set nb010_c6, 112.set nb010_c12, 128.set nb010_six, 144.set nb010_twelve, 160.set nb010_Vvdwtot, 176.set nb010_fix, 192.set nb010_fiy, 208.set nb010_fiz, 224.set nb010_half, 240.set nb010_three, 256.set nb010_nri, 272.set nb010_iinr, 280.set nb010_jindex, 288.set nb010_jjnr, 296.set nb010_shift, 304.set nb010_shiftvec, 312.set nb010_facel, 320.set nb010_innerjjnr, 328.set nb010_is3, 336.set nb010_ii3, 340.set nb010_ntia, 344.set nb010_innerk, 348.set nb010_n, 352.set nb010_nn1, 356.set nb010_ntype, 360.set nb010_nouter, 364.set nb010_ninner, 368 push %rbp movq %rsp,%rbp push %rbx emms push %r12 push %r13 push %r14 push %r15 subq $392,%rsp ## local variable stack space (n*16+8) ## zero 32-bit iteration counters movl $0,%eax movl %eax,nb010_nouter(%rsp) movl %eax,nb010_ninner(%rsp) movl (%rdi),%edi movl %edi,nb010_nri(%rsp) movq %rsi,nb010_iinr(%rsp) movq %rdx,nb010_jindex(%rsp) movq %rcx,nb010_jjnr(%rsp) movq %r8,nb010_shift(%rsp) movq %r9,nb010_shiftvec(%rsp) movq nb010_p_ntype(%rbp),%rdi movl (%rdi),%edi movl %edi,nb010_ntype(%rsp) ## create constant floating-point factors on stack movl $0x00000000,%eax ## lower half of double two IEEE (hex) movl $0x40000000,%ebx movl %eax,nb010_two(%rsp) movl %ebx,nb010_two+4(%rsp) movsd nb010_two(%rsp),%xmm1 shufpd $0,%xmm1,%xmm1 ## splat to all elements movapd %xmm1,%xmm2 addpd %xmm1,%xmm2 ## 4.0 addpd %xmm1,%xmm2 ## six movapd %xmm2,%xmm3 addpd %xmm3,%xmm3 ## twelve movapd %xmm1,nb010_two(%rsp) movapd %xmm2,nb010_six(%rsp) movapd %xmm3,nb010_twelve(%rsp)_nb_kernel010_x86_64_sse2.nb010_threadloop: movq nb010_count(%rbp),%rsi ## pointer to sync counter movl (%rsi),%eax_nb_kernel010_x86_64_sse2.nb010_spinlock: movl %eax,%ebx ## ebx=*count=nn0 addl $1,%ebx ## ebx=nn1=nn0+10 lock cmpxchgl %ebx,(%rsi) ## write nn1 to *counter, ## if it hasnt changed. ## or reread *counter to eax. pause ## -> better p4 performance jnz _nb_kernel010_x86_64_sse2.nb010_spinlock ## if(nn1>nri) nn1=nri movl nb010_nri(%rsp),%ecx movl %ecx,%edx subl %ebx,%ecx cmovlel %edx,%ebx ## if(nn1>nri) nn1=nri ## Cleared the spinlock if we got here. ## eax contains nn0, ebx contains nn1. movl %eax,nb010_n(%rsp) movl %ebx,nb010_nn1(%rsp) subl %eax,%ebx ## calc number of outer lists movl %eax,%esi ## copy n to esi jg _nb_kernel010_x86_64_sse2.nb010_outerstart jmp _nb_kernel010_x86_64_sse2.nb010_end_nb_kernel010_x86_64_sse2.nb010_outerstart: ## ebx contains number of outer iterations addl nb010_nouter(%rsp),%ebx movl %ebx,nb010_nouter(%rsp)_nb_kernel010_x86_64_sse2.nb010_outer: movq nb010_shift(%rsp),%rax ## rax = pointer into shift[] movl (%rax,%rsi,4),%ebx ## rbx=shift[n] lea (%rbx,%rbx,2),%rbx ## rbx=3*is movl %ebx,nb010_is3(%rsp) ## store is3 movq nb010_shiftvec(%rsp),%rax ## rax = base of shiftvec[] movsd (%rax,%rbx,8),%xmm0 movsd 8(%rax,%rbx,8),%xmm1 movsd 16(%rax,%rbx,8),%xmm2 movq nb010_iinr(%rsp),%rcx ## rcx = pointer into iinr[] movl (%rcx,%rsi,4),%ebx ## ebx =ii movq nb010_type(%rbp),%rdx movl (%rdx,%rbx,4),%edx imull nb010_ntype(%rsp),%edx shll %edx movl %edx,nb010_ntia(%rsp) lea (%rbx,%rbx,2),%rbx ## rbx = 3*ii=ii3 movq nb010_pos(%rbp),%rax ## rax = base of pos[] addsd (%rax,%rbx,8),%xmm0 addsd 8(%rax,%rbx,8),%xmm1 addsd 16(%rax,%rbx,8),%xmm2 shufpd $0,%xmm0,%xmm0 shufpd $0,%xmm1,%xmm1 shufpd $0,%xmm2,%xmm2 movapd %xmm0,nb010_ix(%rsp) movapd %xmm1,nb010_iy(%rsp) movapd %xmm2,nb010_iz(%rsp) movl %ebx,nb010_ii3(%rsp) ## clear vvdwtot (xmm12) and i forces (xmm13-xmm15) xorpd %xmm12,%xmm12 movapd %xmm12,%xmm13 movapd %xmm12,%xmm14 movapd %xmm12,%xmm15 movq nb010_jindex(%rsp),%rax movl (%rax,%rsi,4),%ecx ## jindex[n] movl 4(%rax,%rsi,4),%edx ## jindex[n+1] subl %ecx,%edx ## number of innerloop atoms movq nb010_pos(%rbp),%rsi movq nb010_faction(%rbp),%rdi movq nb010_jjnr(%rsp),%rax shll $2,%ecx addq %rcx,%rax movq %rax,nb010_innerjjnr(%rsp) ## pointer to jjnr[nj0] movl %edx,%ecx subl $2,%edx addl nb010_ninner(%rsp),%ecx movl %ecx,nb010_ninner(%rsp) addl $0,%edx movl %edx,nb010_innerk(%rsp) ## number of innerloop atoms jge _nb_kernel010_x86_64_sse2.nb010_unroll_loop jmp _nb_kernel010_x86_64_sse2.nb010_checksingle_nb_kernel010_x86_64_sse2.nb010_unroll_loop: ## twice unrolled innerloop here movq nb010_innerjjnr(%rsp),%rdx ## pointer to jjnr[k] movl (%rdx),%r8d movl 4(%rdx),%r9d addq $8,nb010_innerjjnr(%rsp) ## advance pointer (unrolled 2) movq nb010_pos(%rbp),%rsi ## base of pos[] lea (%r8,%r8,2),%rax ## replace jnr with j3 lea (%r9,%r9,2),%rbx ## move two coordinates to xmm0-xmm2 movlpd (%rsi,%rax,8),%xmm4 movlpd 8(%rsi,%rax,8),%xmm5 movlpd 16(%rsi,%rax,8),%xmm6 movhpd (%rsi,%rbx,8),%xmm4 movhpd 8(%rsi,%rbx,8),%xmm5 movhpd 16(%rsi,%rbx,8),%xmm6 ## calc dr subpd nb010_ix(%rsp),%xmm4 subpd nb010_iy(%rsp),%xmm5 subpd nb010_iz(%rsp),%xmm6 movq nb010_type(%rbp),%rsi ## store dr movapd %xmm4,%xmm9 movapd %xmm5,%xmm10 movapd %xmm6,%xmm11 movl (%rsi,%r8,4),%r8d movl (%rsi,%r9,4),%r9d ## square it mulpd %xmm4,%xmm4 mulpd %xmm5,%xmm5 mulpd %xmm6,%xmm6 addpd %xmm5,%xmm4 addpd %xmm6,%xmm4 ## rsq in xmm4 movq nb010_vdwparam(%rbp),%rsi shll %r8d shll %r9d cvtpd2ps %xmm4,%xmm6 rcpps %xmm6,%xmm6 cvtps2pd %xmm6,%xmm6 ## lu in low xmm6 movl nb010_ntia(%rsp),%edi addl %edi,%r8d addl %edi,%r9d ## 1/x lookup seed in xmm6 movapd nb010_two(%rsp),%xmm0 movapd %xmm4,%xmm5 mulpd %xmm6,%xmm4 ## lu*rsq subpd %xmm4,%xmm0 ## 2-lu*rsq mulpd %xmm0,%xmm6 ## (new lu) movlpd (%rsi,%r8,8),%xmm8 ## c6a movlpd 8(%rsi,%r8,8),%xmm4 ## c612b movapd nb010_two(%rsp),%xmm0 mulpd %xmm6,%xmm5 ## lu*rsq subpd %xmm5,%xmm0 ## 2-lu*rsq mulpd %xmm6,%xmm0 ## xmm0=rinvsq movhpd (%rsi,%r9,8),%xmm8 ## c6a c12a movhpd 8(%rsi,%r9,8),%xmm4 ## c6b c12b movapd %xmm0,%xmm1 mulpd %xmm0,%xmm1 mulpd %xmm0,%xmm1 ## xmm1=rinvsix movapd %xmm1,%xmm2 mulpd %xmm2,%xmm2 ## xmm2=rinvtwelve ## c6 in xmm7, c12 in xmm8 mulpd %xmm8,%xmm1 ## mult by c6 mulpd %xmm4,%xmm2 ## mult by c12 movapd %xmm2,%xmm5 subpd %xmm1,%xmm5 ## Vvdw=Vvdw12-Vvdw6 mulpd nb010_six(%rsp),%xmm1 mulpd nb010_twelve(%rsp),%xmm2 subpd %xmm1,%xmm2 mulpd %xmm2,%xmm0 ## xmm4=total fscal movapd %xmm0,%xmm4 ## the fj's - start by accumulating forces from memory movq nb010_faction(%rbp),%rdi movlpd (%rdi,%rax,8),%xmm6 movlpd 8(%rdi,%rax,8),%xmm7 movlpd 16(%rdi,%rax,8),%xmm8 movhpd (%rdi,%rbx,8),%xmm6 movhpd 8(%rdi,%rbx,8),%xmm7 movhpd 16(%rdi,%rbx,8),%xmm8 ## increment potential addpd %xmm5,%xmm12 mulpd %xmm4,%xmm9 mulpd %xmm4,%xmm10 mulpd %xmm4,%xmm11 ## now update f_i addpd %xmm9,%xmm13 addpd %xmm10,%xmm14 addpd %xmm11,%xmm15 addpd %xmm9,%xmm6 addpd %xmm10,%xmm7 addpd %xmm11,%xmm8 movlpd %xmm6,(%rdi,%rax,8) movlpd %xmm7,8(%rdi,%rax,8) movlpd %xmm8,16(%rdi,%rax,8) movhpd %xmm6,(%rdi,%rbx,8) movhpd %xmm7,8(%rdi,%rbx,8) movhpd %xmm8,16(%rdi,%rbx,8) ## should we do one more iteration? subl $2,nb010_innerk(%rsp) jl _nb_kernel010_x86_64_sse2.nb010_checksingle jmp _nb_kernel010_x86_64_sse2.nb010_unroll_loop_nb_kernel010_x86_64_sse2.nb010_checksingle: movl nb010_innerk(%rsp),%edx andl $1,%edx jnz _nb_kernel010_x86_64_sse2.nb010_dosingle jmp _nb_kernel010_x86_64_sse2.nb010_updateouterdata_nb_kernel010_x86_64_sse2.nb010_dosingle: movq nb010_pos(%rbp),%rdi movq nb010_innerjjnr(%rsp),%rcx movl (%rcx),%eax movq nb010_type(%rbp),%rsi movl (%rsi,%rax,4),%r8d movq nb010_vdwparam(%rbp),%rsi shll %r8d movl nb010_ntia(%rsp),%edi addl %edi,%r8d movsd (%rsi,%r8,8),%xmm7 ## c6 movsd 8(%rsi,%r8,8),%xmm8 ## c12 ## c6 in xmm7, c12 in xmm8 movq nb010_pos(%rbp),%rsi ## base of pos[] lea (%rax,%rax,2),%rax ## replace jnr with j3 ## move two coordinates to xmm0-xmm2 movsd (%rsi,%rax,8),%xmm4 movsd 8(%rsi,%rax,8),%xmm5 movsd 16(%rsi,%rax,8),%xmm6 ## calc dr subsd nb010_ix(%rsp),%xmm4 subsd nb010_iy(%rsp),%xmm5 subsd nb010_iz(%rsp),%xmm6 ## store dr movapd %xmm4,%xmm9 movapd %xmm5,%xmm10 movapd %xmm6,%xmm11 ## square it mulsd %xmm4,%xmm4 mulsd %xmm5,%xmm5 mulsd %xmm6,%xmm6 addsd %xmm5,%xmm4 addsd %xmm6,%xmm4 ## rsq in xmm4 cvtsd2ss %xmm4,%xmm6 rcpss %xmm6,%xmm6 cvtss2sd %xmm6,%xmm6 ## lu in low xmm6 ## 1/x lookup seed in xmm6 movapd nb010_two(%rsp),%xmm0 movapd %xmm4,%xmm5 mulsd %xmm6,%xmm4 ## lu*rsq subsd %xmm4,%xmm0 ## 2-lu*rsq mulsd %xmm0,%xmm6 ## (new lu) movapd nb010_two(%rsp),%xmm0 mulsd %xmm6,%xmm5 ## lu*rsq subsd %xmm5,%xmm0 ## 2-lu*rsq mulsd %xmm6,%xmm0 ## xmm0=rinvsq movapd %xmm0,%xmm1 mulsd %xmm0,%xmm1 mulsd %xmm0,%xmm1 ## xmm1=rinvsix movapd %xmm1,%xmm2 mulsd %xmm2,%xmm2 ## xmm2=rinvtwelve mulsd %xmm7,%xmm1 ## mult by c6 mulsd %xmm8,%xmm2 ## mult by c12 movapd %xmm2,%xmm5 subsd %xmm1,%xmm5 ## Vvdw=Vvdw12-Vvdw6 mulsd nb010_six(%rsp),%xmm1 mulsd nb010_twelve(%rsp),%xmm2 subsd %xmm1,%xmm2 mulsd %xmm2,%xmm0 ## xmm4=total fscal movapd %xmm0,%xmm4 ## increment potential addsd %xmm5,%xmm12 movq nb010_faction(%rbp),%rdi mulsd %xmm4,%xmm9 mulsd %xmm4,%xmm10 mulsd %xmm4,%xmm11 ## now update f_i addsd %xmm9,%xmm13 addsd %xmm10,%xmm14 addsd %xmm11,%xmm15 ## the fj's - start by accumulating forces from memory movsd (%rdi,%rax,8),%xmm3 movsd 8(%rdi,%rax,8),%xmm4 movsd 16(%rdi,%rax,8),%xmm5 addsd %xmm9,%xmm3 addsd %xmm10,%xmm4 addsd %xmm11,%xmm5 movsd %xmm3,(%rdi,%rax,8) movsd %xmm4,8(%rdi,%rax,8) movsd %xmm5,16(%rdi,%rax,8)_nb_kernel010_x86_64_sse2.nb010_updateouterdata: movl nb010_ii3(%rsp),%ecx movq nb010_faction(%rbp),%rdi movq nb010_fshift(%rbp),%rsi movl nb010_is3(%rsp),%edx ## accumulate i forces in xmm13, xmm14, xmm15 movhlps %xmm13,%xmm3 movhlps %xmm14,%xmm4 movhlps %xmm15,%xmm5 addsd %xmm3,%xmm13 addsd %xmm4,%xmm14 addsd %xmm5,%xmm15 ## sum is in low xmm13-xmm15 ## increment i force movsd (%rdi,%rcx,8),%xmm3 movsd 8(%rdi,%rcx,8),%xmm4 movsd 16(%rdi,%rcx,8),%xmm5 subsd %xmm13,%xmm3 subsd %xmm14,%xmm4 subsd %xmm15,%xmm5 movsd %xmm3,(%rdi,%rcx,8) movsd %xmm4,8(%rdi,%rcx,8) movsd %xmm5,16(%rdi,%rcx,8) ## increment fshift force movsd (%rsi,%rdx,8),%xmm3 movsd 8(%rsi,%rdx,8),%xmm4 movsd 16(%rsi,%rdx,8),%xmm5 subsd %xmm13,%xmm3 subsd %xmm14,%xmm4 subsd %xmm15,%xmm5 movsd %xmm3,(%rsi,%rdx,8) movsd %xmm4,8(%rsi,%rdx,8) movsd %xmm5,16(%rsi,%rdx,8) ## get n from stack⌨️ 快捷键说明
复制代码Ctrl + C
搜索代码Ctrl + F
全屏模式F11
增大字号Ctrl + =
减小字号Ctrl + -
显示快捷键?