Update force kernel to be integrated
Signed-off-by: Rafael Ravedutti <rafaelravedutti@gmail.com>
This commit is contained in:
Rafael Ravedutti
parent
ec556eb117
commit
9b615cf0b3
@ -10,224 +10,153 @@ computeForce:
|
|||||||
push r12
|
push r12
|
||||||
push r13
|
push r13
|
||||||
push r14
|
push r14
|
||||||
|
mov r9d, DWORD PTR [4+rsi] # r9d <- atom->Nlocal
|
||||||
# r9d <- atom->Nlocal
|
vmovsd xmm2, QWORD PTR [72+rdi] # xmm2 <- param->cutforce
|
||||||
mov r9d, DWORD PTR [4+rsi]
|
vmovsd xmm1, QWORD PTR [8+rdi] # xmm1 <- param->sigma6
|
||||||
# xmm2 <- param->cutforce
|
vmovsd xmm0, QWORD PTR [rdi] # xmm0 <- param->epsilon
|
||||||
vmovsd xmm2, QWORD PTR [72+rdi]
|
mov r13, QWORD PTR [64+rsi] # r13 <- atom->fx
|
||||||
# xmm1 <- param->sigma6
|
mov r14, QWORD PTR [72+rsi] # r14 <- atom->fy
|
||||||
vmovsd xmm1, QWORD PTR [8+rdi]
|
mov rdi, QWORD PTR [80+rsi] # rdi <- atom->fz
|
||||||
# xmm0 <- param->epsilon
|
test r9d, r9d # atom->Nlocal <= 0
|
||||||
vmovsd xmm0, QWORD PTR [rdi]
|
jle ..exit_func
|
||||||
|
|
||||||
# r13 <- atom->fx
|
|
||||||
mov r13, QWORD PTR [64+rsi]
|
|
||||||
# r14 <- atom->fy
|
|
||||||
mov r14, QWORD PTR [72+rsi]
|
|
||||||
# rdi <- atom->fz
|
|
||||||
mov rdi, QWORD PTR [80+rsi]
|
|
||||||
|
|
||||||
# atom->Nlocal <= 0
|
|
||||||
test r9d, r9d
|
|
||||||
jle ..B1.30
|
|
||||||
|
|
||||||
..B1.2:
|
..B1.2:
|
||||||
# r10d <- 0
|
xor r10d, r10d # r10d <- 0
|
||||||
xor r10d, r10d
|
mov ecx, r9d # ecx <- atom->Nlocal
|
||||||
# ecx <- atom->Nlocal
|
xor r8d, r8d # r8d <- 0
|
||||||
mov ecx, r9d
|
mov r11d, 1 # r11d <- 1
|
||||||
# r8d <- 0
|
xor eax, eax # eax <- 0
|
||||||
xor r8d, r8d
|
shr ecx, 1 # ecx <- atom->Nlocal >> 1
|
||||||
# r11d <- 1
|
je ..B1.6 # ecx == 0
|
||||||
mov r11d, 1
|
|
||||||
# eax <- 0
|
|
||||||
xor eax, eax
|
|
||||||
# ecx <- atom->Nlocal >> 1
|
|
||||||
shr ecx, 1
|
|
||||||
je ..B1.6
|
|
||||||
|
|
||||||
# Init forces to zero loop
|
# Init forces to zero loop
|
||||||
..B1.4:
|
..B1.4:
|
||||||
# fx[i] <- 0
|
mov QWORD PTR [r8+r13], rax # fx[i] <- 0
|
||||||
mov QWORD PTR [r8+r13], rax
|
mov QWORD PTR [r8+r14], rax # fy[i] <- 0
|
||||||
# i++
|
mov QWORD PTR [r8+rdi], rax # fz[i] <- 0
|
||||||
inc r10
|
mov QWORD PTR [8+r8+r13], rax # fx[i] <- 0
|
||||||
# fy[i] <- 0
|
mov QWORD PTR [8+r8+r14], rax # fy[i] <- 0
|
||||||
mov QWORD PTR [r8+r14], rax
|
mov QWORD PTR [8+r8+rdi], rax # fz[i] <- 0
|
||||||
# fz[i] <- 0
|
add r8, 16 # i++
|
||||||
mov QWORD PTR [r8+rdi], rax
|
inc r10 # i++
|
||||||
# fx[i] <- 0
|
cmp r10, rcx # i < Nlocal
|
||||||
mov QWORD PTR [8+r8+r13], rax
|
|
||||||
# fy[i] <- 0
|
|
||||||
mov QWORD PTR [8+r8+r14], rax
|
|
||||||
# fz[i] <- 0
|
|
||||||
mov QWORD PTR [8+r8+rdi], rax
|
|
||||||
# i++
|
|
||||||
add r8, 16
|
|
||||||
# i < Nlocal
|
|
||||||
cmp r10, rcx
|
|
||||||
jb ..B1.4
|
jb ..B1.4
|
||||||
|
|
||||||
..B1.5:
|
..B1.5:
|
||||||
# r11d <- i * 2 + 1
|
lea r11d, DWORD PTR [1+r10+r10] # r11d <- i * 2 + 1
|
||||||
lea r11d, DWORD PTR [1+r10+r10]
|
|
||||||
..B1.6:
|
..B1.6:
|
||||||
# r11d <- i * 2
|
lea ecx, DWORD PTR [-1+r11] # r11d <- i * 2
|
||||||
lea ecx, DWORD PTR [-1+r11]
|
cmp ecx, r9d # i < Nlocal
|
||||||
# i < Nlocal
|
|
||||||
cmp ecx, r9d
|
|
||||||
jae ..B1.8
|
jae ..B1.8
|
||||||
|
|
||||||
..B1.7:
|
..B1.7:
|
||||||
|
movsxd r11, r11d # r11 <- i * 2
|
||||||
# r11 <- i * 2
|
mov QWORD PTR [-8+r13+r11*8], rax # fx[i] <- 0
|
||||||
movsxd r11, r11d
|
mov QWORD PTR [-8+r14+r11*8], rax # fy[i] <- 0
|
||||||
# fx[i] <- 0
|
mov QWORD PTR [-8+rdi+r11*8], rax # fz[i] <- 0
|
||||||
mov QWORD PTR [-8+r13+r11*8], rax
|
|
||||||
# fy[i] <- 0
|
|
||||||
mov QWORD PTR [-8+r14+r11*8], rax
|
|
||||||
# fz[i] <- 0
|
|
||||||
mov QWORD PTR [-8+rdi+r11*8], rax
|
|
||||||
|
|
||||||
..B1.8:
|
..B1.8:
|
||||||
|
vmulsd xmm15, xmm2, xmm2 # xmm15 <- cutforcesq
|
||||||
|
xor r8d, r8d # r8d <- 0
|
||||||
|
vmovdqu32 ymm18, YMMWORD PTR .L_2il0floatpacket.0[rip] # ymm18 <- [8, ...]
|
||||||
|
vmulsd xmm0, xmm0, QWORD PTR .L_2il0floatpacket.3[rip] # xmm0 <- 48 * epsilon
|
||||||
|
vmovdqu32 ymm17, YMMWORD PTR .L_2il0floatpacket.1[rip] # ymm17 <- [0..7]
|
||||||
|
vmovups zmm7, ZMMWORD PTR .L_2il0floatpacket.4[rip] # zmm7 <- [0.5, ...]
|
||||||
|
vbroadcastsd zmm16, xmm15 # zmm16 <- [cutforcesq, ...]
|
||||||
|
vbroadcastsd zmm15, xmm1 # zmm15 <- [param->sigma6, ...]
|
||||||
|
vbroadcastsd zmm14, xmm0 # zmm16 <- [48 * epsilon, ...]
|
||||||
|
movsxd r9, r9d # r9 <- atom->Nlocal
|
||||||
|
xor r10d, r10d # r10d <- 0 (i)
|
||||||
|
mov rcx, QWORD PTR [24+rdx] # rcx <- neighbor->numneigh
|
||||||
|
mov r11, QWORD PTR [8+rdx] # r11 <- neighbor->neighbors
|
||||||
|
movsxd r12, DWORD PTR [16+rdx] # r12 <- neighbor->maxneighs
|
||||||
|
mov rdx, QWORD PTR [16+rsi] # rdx <- atom->x
|
||||||
|
|
||||||
# xmm15 <- cutforcesq
|
### AOS
|
||||||
vmulsd xmm15, xmm2, xmm2
|
xor eax, eax
|
||||||
# r8d <- 0
|
### SOA
|
||||||
xor r8d, r8d
|
#mov rax, QWORD PTR [24+rsi] # rax <- atom->y
|
||||||
# ymm18 <- 8
|
#mov rsi, QWORD PTR [32+rsi] # rsi <- atom->z
|
||||||
vmovdqu32 ymm18, YMMWORD PTR .L_2il0floatpacket.0[rip]
|
###
|
||||||
# xmm0 <- 48 * epsilon
|
|
||||||
vmulsd xmm0, xmm0, QWORD PTR .L_2il0floatpacket.3[rip]
|
|
||||||
# ymm17 <- [0..7]
|
|
||||||
vmovdqu32 ymm17, YMMWORD PTR .L_2il0floatpacket.1[rip]
|
|
||||||
# zmm7 <- 0.5
|
|
||||||
vmovups zmm7, ZMMWORD PTR .L_2il0floatpacket.4[rip]
|
|
||||||
# zmm16 <- cutforcesq
|
|
||||||
vbroadcastsd zmm16, xmm15
|
|
||||||
# zmm15 <- param->sigma6
|
|
||||||
vbroadcastsd zmm15, xmm1
|
|
||||||
# zmm16 <- 48 * epsilon
|
|
||||||
vbroadcastsd zmm14, xmm0
|
|
||||||
# r9 <- atom->Nlocal
|
|
||||||
movsxd r9, r9d
|
|
||||||
# r10d <- 0 (i)
|
|
||||||
xor r10d, r10d
|
|
||||||
# rcx <- neighbor->numneigh
|
|
||||||
mov rcx, QWORD PTR [24+rdx]
|
|
||||||
# r11 <- neighbor->neighbors
|
|
||||||
mov r11, QWORD PTR [8+rdx]
|
|
||||||
# r12 <- neighbor->maxneighs
|
|
||||||
movsxd r12, DWORD PTR [16+rdx]
|
|
||||||
|
|
||||||
# rdx <- atom->x
|
shl r12, 2 # r12 <- neighbor->maxneighs * 4
|
||||||
mov rdx, QWORD PTR [16+rsi]
|
mov QWORD PTR [-32+rsp], r9 # [-32+rsp] <- atom->Nlocal
|
||||||
# rax <- atom->y
|
mov QWORD PTR [-24+rsp], rcx # [-24+rsp] <- neighbor->numneigh
|
||||||
mov rax, QWORD PTR [24+rsi]
|
mov QWORD PTR [-16+rsp], r14 # [-16+rsp] <- atom->fy
|
||||||
# rsi <- atom->z
|
mov QWORD PTR [-8+rsp], r13 # [-8+rsp] <- atom->fx
|
||||||
mov rsi, QWORD PTR [32+rsi]
|
mov QWORD PTR [-40+rsp], r15 # [-40+rsp] <- r15
|
||||||
|
mov QWORD PTR [-48+rsp], rbx # [-48+rsp] <- rbx
|
||||||
# r12 <- neighbor->maxneighs * 4
|
|
||||||
shl r12, 2
|
|
||||||
# [-32+rsp] <- atom->Nlocal
|
|
||||||
mov QWORD PTR [-32+rsp], r9
|
|
||||||
# [-24+rsp] <- neighbor->numneigh
|
|
||||||
mov QWORD PTR [-24+rsp], rcx
|
|
||||||
# [-16+rsp] <- atom->fy
|
|
||||||
mov QWORD PTR [-16+rsp], r14
|
|
||||||
# [-8+rsp] <- atom->fx
|
|
||||||
mov QWORD PTR [-8+rsp], r13
|
|
||||||
# [-40+rsp] <- r15
|
|
||||||
mov QWORD PTR [-40+rsp], r15
|
|
||||||
# [-48+rsp] <- rbx
|
|
||||||
mov QWORD PTR [-48+rsp], rbx
|
|
||||||
# zmm19 <- 0
|
|
||||||
vpxord zmm19, zmm19, zmm19
|
|
||||||
|
|
||||||
# Loop over all atoms
|
# Loop over all atoms
|
||||||
..B1.9:
|
..B1.9:
|
||||||
|
mov rcx, QWORD PTR [-24+rsp] # rcx <- neighbor->numneigh
|
||||||
|
vxorpd xmm25, xmm25, xmm25 # xmm25 <- 0
|
||||||
|
vmovapd xmm20, xmm25 # xmm20 <- 0
|
||||||
|
mov r13d, DWORD PTR [rcx+r10*4] # r13d <- neighbor->numneigh[i] (numneighs)
|
||||||
|
vmovapd xmm4, xmm20 # xmm4 <- 0
|
||||||
|
|
||||||
# rcx <- neighbor->numneigh
|
### AOS
|
||||||
mov rcx, QWORD PTR [-24+rsp]
|
vmovsd xmm8, QWORD PTR[rdx+rax] # xmm8 <- atom->x[i * 3]
|
||||||
# xmm25 <- 0
|
vmovsd xmm9, QWORD PTR[8+rdx+rax] # xmm9 <- atom->x[i * 3 + 1]
|
||||||
vxorpd xmm25, xmm25, xmm25
|
vmovsd xmm10, QWORD PTR[16+rdx+rax] # xmm10 <- atom->x[i * 3 + 2]
|
||||||
# xmm20 <- 0
|
### SOA
|
||||||
vmovapd xmm20, xmm25
|
#vmovsd xmm8, QWORD PTR [rdx+r10*8] # xmm8 <- atom->x[i]
|
||||||
# r13d <- neighbor->numneigh[i] (numneighs)
|
#vmovsd xmm9, QWORD PTR [rax+r10*8] # xmm9 <- atom->y[i]
|
||||||
mov r13d, DWORD PTR [rcx+r10*4]
|
#vmovsd xmm10, QWORD PTR [rsi+r10*8] # xmm10 <- atom->z[i]
|
||||||
# xmm4 <- 0
|
###
|
||||||
vmovapd xmm4, xmm20
|
|
||||||
# xmm8 <- atom->x[i]
|
test r13d, r13d # numneighs <= 0
|
||||||
vmovsd xmm8, QWORD PTR [rdx+r10*8]
|
jle ..exit_func
|
||||||
# xmm9 <- atom->y[i]
|
|
||||||
vmovsd xmm9, QWORD PTR [rax+r10*8]
|
|
||||||
# xmm9 <- atom->z[i]
|
|
||||||
vmovsd xmm10, QWORD PTR [rsi+r10*8]
|
|
||||||
# numneighs <= 0
|
|
||||||
test r13d, r13d
|
|
||||||
jle ..B1.27
|
|
||||||
|
|
||||||
..B1.10:
|
..B1.10:
|
||||||
|
vpxord zmm13, zmm13, zmm13 # zmm13 <- 0 (fix)
|
||||||
|
vmovaps zmm12, zmm13 # zmm12 <- 0 (fiy)
|
||||||
|
vmovaps zmm11, zmm12 # zmm11 <- 0 (fiz)
|
||||||
|
|
||||||
# zmm13 <- 0
|
mov r14d, r13d # r14d <- numneighs
|
||||||
vmovaps zmm13, zmm19
|
xor r11d, r11d # r11d <- 0
|
||||||
# zmm12 <- 0
|
and r14d, -8 # r14d <- numneighs & (-8)
|
||||||
vmovaps zmm12, zmm13
|
lea r9d, DWORD PTR [8+r11] # r9d <- 8 (why lea?)
|
||||||
# zmm11 <- 0
|
cmp r14d, r9d # r14d < r9d
|
||||||
vmovaps zmm11, zmm12
|
jl ..B1.33
|
||||||
|
|
||||||
# numneighs < 8
|
# cmp r13d, 8 # numneighs < 8
|
||||||
cmp r13d, 8
|
# jl ..B1.32
|
||||||
jl ..B1.32
|
#..B1.11:
|
||||||
|
# cmp r13d, 1200 # numneighs < 1200
|
||||||
..B1.11:
|
# jl ..B1.31
|
||||||
|
#..B1.12:
|
||||||
# numneighs < 1200
|
# mov rcx, r12
|
||||||
cmp r13d, 1200
|
# imul rcx, r8
|
||||||
jl ..B1.31
|
# add rcx, r11 # rcx <- &neighbor->neighbors[neighbor->maxneighs * i (r8)]
|
||||||
|
# mov r9, rcx # r9 <- neighs
|
||||||
..B1.12:
|
# and r9, 63 # r9 <- neighs & 63
|
||||||
|
# test r9d, 3 # (r9d & 3) == 0 => r9d divisible by 8
|
||||||
mov rcx, r12
|
# je ..B1.14
|
||||||
imul rcx, r8
|
#..B1.13:
|
||||||
add rcx, r11
|
# xor r9d, r9d # r9d <- 0
|
||||||
mov r9, rcx
|
# jmp ..B1.16
|
||||||
and r9, 63
|
#..B1.14:
|
||||||
test r9d, 3
|
# test r9d, r9d # r9d == 0
|
||||||
je ..B1.14
|
# je ..B1.16
|
||||||
|
#..B1.15:
|
||||||
..B1.13:
|
# neg r9d
|
||||||
|
# add r9d, 64
|
||||||
xor r9d, r9d
|
# shr r9d, 2 # r9d <- (64 - r9d) / 4
|
||||||
jmp ..B1.16
|
# cmp r13d, r9d # numneighs < r9d
|
||||||
|
# cmovl r9d, r13d # r9d <- MIN(numneighs, r9d)
|
||||||
..B1.14:
|
#..B1.16:
|
||||||
|
# mov ebx, r13d
|
||||||
test r9d, r9d
|
# sub ebx, r9d
|
||||||
je ..B1.16
|
# and ebx, 7
|
||||||
|
# neg ebx
|
||||||
..B1.15:
|
# add ebx, r13d # ebx <- -((numneighs - r9d) & 7) + numneighs
|
||||||
|
# cmp r9d, 1 # r9d < 1
|
||||||
neg r9d
|
# jb ..B1.20
|
||||||
add r9d, 64
|
#..B1.20:
|
||||||
shr r9d, 2
|
# lea ecx, DWORD PTR [8+r9] # ecx <- r9d[1]
|
||||||
cmp r13d, r9d
|
# cmp ebx, ecx # -((numneighs - r9d) & 7) + numneighs < neighs
|
||||||
cmovl r9d, r13d
|
# jl ..B1.24
|
||||||
|
|
||||||
..B1.16:
|
|
||||||
|
|
||||||
mov ebx, r13d
|
|
||||||
sub ebx, r9d
|
|
||||||
and ebx, 7
|
|
||||||
neg ebx
|
|
||||||
add ebx, r13d
|
|
||||||
cmp r9d, 1
|
|
||||||
jb ..B1.20
|
|
||||||
|
|
||||||
..B1.20:
|
|
||||||
lea ecx, DWORD PTR [8+r9]
|
|
||||||
cmp ebx, ecx
|
|
||||||
jl ..B1.24
|
|
||||||
|
|
||||||
..B1.21:
|
..B1.21:
|
||||||
mov rcx, r12
|
mov rcx, r12
|
||||||
@ -246,17 +175,25 @@ push r14
|
|||||||
vmovdqu ymm3, YMMWORD PTR [rcx+r14*4]
|
vmovdqu ymm3, YMMWORD PTR [rcx+r14*4]
|
||||||
add r14, 8
|
add r14, 8
|
||||||
vpxord zmm5, zmm5, zmm5
|
vpxord zmm5, zmm5, zmm5
|
||||||
vpxord zmm4, zmm4, zmm4
|
|
||||||
vpxord zmm6, zmm6, zmm6
|
vpxord zmm6, zmm6, zmm6
|
||||||
|
|
||||||
vgatherdpd zmm5{k2}, QWORD PTR [rax+ymm3*8]
|
### AOS
|
||||||
|
vpaddd ymm4, ymm3, ymm3
|
||||||
|
vpaddd ymm3, ymm3, ymm4
|
||||||
|
vpxord zmm4, zmm4, zmm4
|
||||||
vgatherdpd zmm4{k1}, QWORD PTR [rdx+ymm3*8]
|
vgatherdpd zmm4{k1}, QWORD PTR [rdx+ymm3*8]
|
||||||
vgatherdpd zmm6{k3}, QWORD PTR [rsi+ymm3*8]
|
vgatherdpd zmm5{k2}, QWORD PTR [8+rdx+ymm3*8]
|
||||||
|
vgatherdpd zmm6{k3}, QWORD PTR [16+rdx+ymm3*8]
|
||||||
|
### SOA
|
||||||
|
#vpxord zmm4, zmm4, zmm4
|
||||||
|
#vgatherdpd zmm5{k2}, QWORD PTR [rax+ymm3*8]
|
||||||
|
#vgatherdpd zmm4{k1}, QWORD PTR [rdx+ymm3*8]
|
||||||
|
#vgatherdpd zmm6{k3}, QWORD PTR [rsi+ymm3*8]
|
||||||
|
###
|
||||||
|
|
||||||
vsubpd zmm29, zmm1, zmm5
|
vsubpd zmm29, zmm1, zmm5
|
||||||
vsubpd zmm28, zmm0, zmm4
|
vsubpd zmm28, zmm0, zmm4
|
||||||
vsubpd zmm31, zmm2, zmm6
|
vsubpd zmm31, zmm2, zmm6
|
||||||
|
|
||||||
vmulpd zmm20, zmm29, zmm29
|
vmulpd zmm20, zmm29, zmm29
|
||||||
vfmadd231pd zmm20, zmm28, zmm28
|
vfmadd231pd zmm20, zmm28, zmm28
|
||||||
vfmadd231pd zmm20, zmm31, zmm31
|
vfmadd231pd zmm20, zmm31, zmm31
|
||||||
@ -264,17 +201,13 @@ push r14
|
|||||||
# if condition cutoff radius
|
# if condition cutoff radius
|
||||||
vrcp14pd zmm27, zmm20 #-> sr2
|
vrcp14pd zmm27, zmm20 #-> sr2
|
||||||
vcmppd k5, zmm20, zmm16, 1
|
vcmppd k5, zmm20, zmm16, 1
|
||||||
|
vmulpd zmm22, zmm27, zmm15 # zmm22 <- sr2 * sigma6
|
||||||
vmulpd zmm22, zmm27, zmm15 #-> sr2 * sigma6
|
vmulpd zmm24, zmm27, zmm14 # zmm24 <- 48.0 * epsilon * sr2
|
||||||
vmulpd zmm24, zmm27, zmm14 #-> 48.0 * epsilon * sr2
|
vmulpd zmm25, zmm27, zmm22 # zmm25 <- sr2 * sigma6 * sr2
|
||||||
|
vmulpd zmm23, zmm27, zmm25 # zmm23 <- sr2 * sigma6 * sr2 * sr2
|
||||||
vmulpd zmm25, zmm27, zmm22 #-> sr2 * sigma6 * sr2
|
vfmsub213pd zmm27, zmm25, zmm7 # zmm27 <- sr2 * sigma * sr2 * sr2 - 0.5
|
||||||
vmulpd zmm23, zmm27, zmm25 #-> sr2 * sigma6 * sr2 * sr2
|
vmulpd zmm26, zmm23, zmm24 # zmm26 <- 48.0 * epsilon * sr2 * sr2 * sigma6 * sr2
|
||||||
|
vmulpd zmm30, zmm26, zmm27 # zmm30 <- force
|
||||||
vfmsub213pd zmm27, zmm25, zmm7 #-> sr2 * sigma * sr2 * sr2 - 0.5
|
|
||||||
vmulpd zmm26, zmm23, zmm24 #-> 48.0 * epsilon * sr2 * sr2 * sigma6 * sr2
|
|
||||||
vmulpd zmm30, zmm26, zmm27 #->
|
|
||||||
|
|
||||||
vfmadd231pd zmm13{k5}, zmm30, zmm28
|
vfmadd231pd zmm13{k5}, zmm30, zmm28
|
||||||
vfmadd231pd zmm12{k5}, zmm30, zmm29
|
vfmadd231pd zmm12{k5}, zmm30, zmm29
|
||||||
vfmadd231pd zmm11{k5}, zmm30, zmm31
|
vfmadd231pd zmm11{k5}, zmm30, zmm31
|
||||||
@ -304,9 +237,14 @@ push r14
|
|||||||
vaddpd zmm25, zmm23, zmm24
|
vaddpd zmm25, zmm23, zmm24
|
||||||
|
|
||||||
#exit function
|
#exit function
|
||||||
..B1.27:
|
..exit_func:
|
||||||
mov rcx, QWORD PTR [-8+rsp] #84.9[spill]
|
mov rcx, QWORD PTR [-8+rsp] #84.9[spill]
|
||||||
mov rbx, QWORD PTR [-16+rsp] #85.9[spill]
|
mov rbx, QWORD PTR [-16+rsp] #85.9[spill]
|
||||||
|
|
||||||
|
### AOS
|
||||||
|
add rax, 24
|
||||||
|
###
|
||||||
|
|
||||||
movsxd r8, r10d #55.32
|
movsxd r8, r10d #55.32
|
||||||
inc r8 #55.32
|
inc r8 #55.32
|
||||||
vaddsd xmm0, xmm25, QWORD PTR [rcx+r10*8] #84.9
|
vaddsd xmm0, xmm25, QWORD PTR [rcx+r10*8] #84.9
|
||||||
@ -318,8 +256,6 @@ push r14
|
|||||||
inc r10 #55.5
|
inc r10 #55.5
|
||||||
cmp r10, QWORD PTR [-32+rsp] #55.5[spill]
|
cmp r10, QWORD PTR [-32+rsp] #55.5[spill]
|
||||||
jb ..B1.9
|
jb ..B1.9
|
||||||
|
|
||||||
|
|
||||||
vzeroupper #93.12
|
vzeroupper #93.12
|
||||||
vxorpd xmm0, xmm0, xmm0 #93.12
|
vxorpd xmm0, xmm0, xmm0 #93.12
|
||||||
pop r14 #93.12
|
pop r14 #93.12
|
||||||
|
|||||||
Loading…
Reference in New Issue
Block a user