361 lines
19 KiB
ArmAsm
361 lines
19 KiB
ArmAsm
.intel_syntax noprefix
|
|
|
|
.text
|
|
.align 16,0x90
|
|
.globl computeForce
|
|
computeForce:
|
|
# parameter 1: rdi Parameter*
|
|
# parameter 2: rsi Atom*
|
|
# parameter 3: rdx Neighbor*
|
|
push rbp
|
|
push r12
|
|
push r13
|
|
push r14
|
|
push r15
|
|
push rbx
|
|
mov r9d, DWORD PTR [4+rsi] # r9d <- atom->Nlocal
|
|
vmovsd xmm2, QWORD PTR [96+rdi] # xmm2 <- param->cutforce
|
|
vmovsd xmm1, QWORD PTR [32+rdi] # xmm1 <- param->sigma6
|
|
vmovsd xmm0, QWORD PTR [24+rdi] # xmm0 <- param->epsilon
|
|
mov r13, QWORD PTR [64+rsi] # r13 <- atom->fx
|
|
mov r14, QWORD PTR [72+rsi] # r14 <- atom->fy
|
|
mov rdi, QWORD PTR [80+rsi] # rdi <- atom->fz
|
|
test r9d, r9d # atom->Nlocal <= 0
|
|
jle ..atom_loop_exit
|
|
xor r10d, r10d # r10d <- 0
|
|
mov ecx, r9d # ecx <- atom->Nlocal
|
|
xor r8d, r8d # r8d <- 0
|
|
mov r11d, 1 # r11d <- 1
|
|
xor eax, eax # eax <- 0
|
|
shr ecx, 1 # ecx <- atom->Nlocal >> 1
|
|
je ..zero_last_element # ecx == 0
|
|
|
|
# Init forces to zero loop (unroll factor = 2)
|
|
..init_force_loop:
|
|
mov QWORD PTR [r8+r13], rax # fx[i] <- 0
|
|
mov QWORD PTR [r8+r14], rax # fy[i] <- 0
|
|
mov QWORD PTR [r8+rdi], rax # fz[i] <- 0
|
|
mov QWORD PTR [8+r8+r13], rax # fx[i] <- 0
|
|
mov QWORD PTR [8+r8+r14], rax # fy[i] <- 0
|
|
mov QWORD PTR [8+r8+rdi], rax # fz[i] <- 0
|
|
add r8, 16 # i++
|
|
inc r10 # i++
|
|
cmp r10, rcx # i < Nlocal
|
|
jb ..init_force_loop
|
|
|
|
# Trick to make r11d contain value of last element to be zeroed plus 1
|
|
# Maybe we can directly put r10+10 here and zero r11d above, then remove the -1 below
|
|
lea r11d, DWORD PTR [1+r10+r10] # r11d <- i * 2 + 1
|
|
..zero_last_element:
|
|
lea ecx, DWORD PTR [-1+r11] # ecx <- i * 2
|
|
cmp ecx, r9d # i >= Nlocal
|
|
jae ..before_atom_loop
|
|
|
|
# Set last element to zero
|
|
movsxd r11, r11d # r11 <- i * 2
|
|
mov QWORD PTR [-8+r13+r11*8], rax # fx[i] <- 0
|
|
mov QWORD PTR [-8+r14+r11*8], rax # fy[i] <- 0
|
|
mov QWORD PTR [-8+rdi+r11*8], rax # fz[i] <- 0
|
|
|
|
# Initialize registers to be used within atom loop
|
|
..before_atom_loop:
|
|
vmulsd xmm15, xmm2, xmm2 # xmm15 <- cutforcesq
|
|
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 # zmm14 <- [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
|
|
### AOS
|
|
xor eax, eax
|
|
### SOA
|
|
#mov rax, QWORD PTR [24+rsi] # rax <- atom->y
|
|
#mov rsi, QWORD PTR [32+rsi] # rsi <- atom->z
|
|
###
|
|
shl r12, 2 # r12 <- neighbor->maxneighs * 4
|
|
|
|
# Register spilling
|
|
mov QWORD PTR [-32+rsp], r9 # [-32+rsp] <- atom->Nlocal
|
|
mov QWORD PTR [-24+rsp], rcx # [-24+rsp] <- neighbor->numneigh
|
|
mov QWORD PTR [-16+rsp], r14 # [-16+rsp] <- atom->fy
|
|
mov QWORD PTR [-8+rsp], r13 # [-8+rsp] <- atom->fx
|
|
mov QWORD PTR [-40+rsp], r15 # [-40+rsp] <- r15
|
|
mov QWORD PTR [-48+rsp], rbx # [-48+rsp] <- rbx
|
|
|
|
..atom_loop_begin:
|
|
mov rcx, QWORD PTR [-24+rsp] # rcx <- neighbor->numneigh
|
|
vxorpd xmm25, xmm25, xmm25 # xmm25 <- 0 (fix)
|
|
vmovapd xmm20, xmm25 # xmm20 <- 0 (fiy)
|
|
mov r13d, DWORD PTR [rcx+r10*4] # r13d <- neighbor->numneigh[i] (numneighs)
|
|
vmovapd xmm4, xmm20 # xmm4 <- 0 (fiz)
|
|
|
|
### AOS
|
|
vmovsd xmm8, QWORD PTR[rdx+rax] # xmm8 <- atom->x[i * 3]
|
|
vmovsd xmm9, QWORD PTR[8+rdx+rax] # xmm9 <- atom->x[i * 3 + 1]
|
|
vmovsd xmm10, QWORD PTR[16+rdx+rax] # xmm10 <- atom->x[i * 3 + 2]
|
|
### SOA
|
|
#vmovsd xmm8, QWORD PTR [rdx+r10*8] # xmm8 <- atom->x[i]
|
|
#vmovsd xmm9, QWORD PTR [rax+r10*8] # xmm9 <- atom->y[i]
|
|
#vmovsd xmm10, QWORD PTR [rsi+r10*8] # xmm10 <- atom->z[i]
|
|
###
|
|
vbroadcastsd zmm0, xmm8 # zmm0 <- atom_x(i)
|
|
vbroadcastsd zmm1, xmm9 # zmm1 <- atom_y(i)
|
|
vbroadcastsd zmm2, xmm10 # zmm2 <- atom_z(i)
|
|
test r13d, r13d # numneighs <= 0
|
|
jle ..atom_loop_exit
|
|
|
|
vpxord zmm13, zmm13, zmm13 # zmm13 <- 0 (fix)
|
|
vmovaps zmm12, zmm13 # zmm12 <- 0 (fiy)
|
|
vmovaps zmm11, zmm12 # zmm11 <- 0 (fiz)
|
|
mov rcx, r12 # rcx <- neighbor->maxneighs * 4
|
|
imul rcx, r10 # rcx <- neighbor->maxneighs * 4 * i
|
|
add rcx, r11 # rcx <- &neighbor->neighbors[neighbor->maxneighs * i]
|
|
xor r9d, r9d # r9d <- 0 (k)
|
|
mov r14d, r13d # r14d <- numneighs
|
|
cmp r14d, 8
|
|
jl ..compute_forces_remainder
|
|
|
|
# and r14d, -8 # r14d <- numneighs & (-8)
|
|
# lea r9d, DWORD PTR [8+r11] # r9d <- 8 (why lea?)
|
|
# cmp r14d, r9d # r14d < r9d
|
|
# jl ..B1.33
|
|
|
|
# cmp r13d, 8 # numneighs < 8
|
|
# jl ..B1.32
|
|
#..B1.11:
|
|
# cmp r13d, 1200 # numneighs < 1200
|
|
# jl ..B1.31
|
|
#..B1.12:
|
|
# mov r9, rcx # r9 <- neighs
|
|
# and r9, 63 # r9 <- neighs & 63
|
|
# test r9d, 3 # (r9d & 3) == 0 => r9d divisible by 8
|
|
# je ..B1.14
|
|
#..B1.13:
|
|
# xor r9d, r9d # r9d <- 0
|
|
# jmp ..B1.16
|
|
#..B1.14:
|
|
# test r9d, r9d # r9d == 0
|
|
# je ..B1.16
|
|
#..B1.15:
|
|
# neg r9d
|
|
# add r9d, 64
|
|
# shr r9d, 2 # r9d <- (64 - r9d) / 4
|
|
# cmp r13d, r9d # numneighs < r9d
|
|
# cmovl r9d, r13d # r9d <- MIN(numneighs, r9d)
|
|
#..B1.16:
|
|
# mov ebx, r13d
|
|
# sub ebx, r9d
|
|
# and ebx, 7
|
|
# neg ebx
|
|
# add ebx, r13d # ebx <- -((numneighs - r9d) & 7) + numneighs
|
|
# cmp r9d, 1 # r9d < 1
|
|
# jb ..B1.20
|
|
#..B1.20:
|
|
# lea ecx, DWORD PTR [8+r9] # ecx <- r9d[1]
|
|
# cmp ebx, ecx # -((numneighs - r9d) & 7) + numneighs < neighs
|
|
# jl ..B1.24
|
|
|
|
..compute_forces:
|
|
vpcmpeqb k1, xmm0, xmm0
|
|
vpcmpeqb k2, xmm0, xmm0
|
|
vpcmpeqb k3, xmm0, xmm0
|
|
vmovdqu ymm3, YMMWORD PTR [rcx+r9*4]
|
|
vpxord zmm5, zmm5, zmm5
|
|
vpxord zmm6, zmm6, zmm6
|
|
|
|
### AOS
|
|
vpaddd ymm4, ymm3, ymm3
|
|
vpaddd ymm3, ymm3, ymm4
|
|
vpxord zmm4, zmm4, zmm4
|
|
vgatherdpd zmm4{k1}, [rdx+ymm3*8]
|
|
vgatherdpd zmm5{k2}, [8+rdx+ymm3*8]
|
|
vgatherdpd zmm6{k3}, [16+rdx+ymm3*8]
|
|
### SOA
|
|
#vpxord zmm4, zmm4, zmm4
|
|
#vgatherdpd zmm5{k2}, [rax+ymm3*8]
|
|
#vgatherdpd zmm4{k1}, [rdx+ymm3*8]
|
|
#vgatherdpd zmm6{k3}, [rsi+ymm3*8]
|
|
###
|
|
|
|
vsubpd zmm29, zmm1, zmm5 # zmm29 <- atom_y(i) - atom_y(j) -- dely
|
|
vsubpd zmm28, zmm0, zmm4 # zmm28 <- atom_x(i) - atom_x(j) -- delx
|
|
vsubpd zmm31, zmm2, zmm6 # zmm31 <- atom_z(i) - atom_z(j) -- delz
|
|
vmulpd zmm20, zmm29, zmm29 # zmm20 <- dely * dely
|
|
vfmadd231pd zmm20, zmm28, zmm28 # zmm20 <- dely * dely + delx * delx
|
|
vfmadd231pd zmm20, zmm31, zmm31 # zmm20 <- zmm20 + delz * delz -- rsq
|
|
|
|
# Cutoff radius condition
|
|
vrcp14pd zmm27, zmm20 # zmm27 <- 1.0 / rsq (sr2)
|
|
vcmppd k5, zmm20, zmm16, 1 # k5 <- rsq < cutforcesq
|
|
vmulpd zmm22, zmm27, zmm15 # zmm22 <- sr2 * sigma6
|
|
vmulpd zmm24, zmm27, zmm14 # zmm24 <- 48.0 * epsilon * sr2
|
|
vmulpd zmm25, zmm27, zmm22 # zmm25 <- sr2 * sigma6 * sr2
|
|
vmulpd zmm23, zmm27, zmm25 # zmm23 <- sr2 * sigma6 * sr2 * sr2
|
|
vfmsub213pd zmm27, zmm25, zmm7 # zmm27 <- sr2 * sigma * sr2 * sr2 - 0.5
|
|
vmulpd zmm26, zmm23, zmm24 # zmm26 <- 48.0 * epsilon * sr2 * sr2 * sigma6 * sr2
|
|
vmulpd zmm30, zmm26, zmm27 # zmm30 <- force
|
|
vfmadd231pd zmm13{k5}, zmm30, zmm28 # fix += force * delx
|
|
vfmadd231pd zmm12{k5}, zmm30, zmm29 # fiy += force * dely
|
|
vfmadd231pd zmm11{k5}, zmm30, zmm31 # fiz += force * delz
|
|
sub r14, 8
|
|
add r9, 8
|
|
cmp r14, 8
|
|
jge ..compute_forces
|
|
|
|
# Check if there are remaining neighbors to be computed
|
|
..compute_forces_remainder:
|
|
cmp r14, 1
|
|
jl ..sum_up_forces
|
|
|
|
vpbroadcastd ymm0, r14d
|
|
vpcmpgtd k1, ymm0, ymm17
|
|
kmovw r15d, k1
|
|
vmovdqu32 ymm3{k3}{z}, YMMWORD PTR [rcx+r9*4]
|
|
kmovw k2, k1
|
|
kmovw k3, k1
|
|
vpxord zmm5, zmm5, zmm5
|
|
vpxord zmm6, zmm6, zmm6
|
|
|
|
### AOS
|
|
vpaddd ymm4, ymm3, ymm3
|
|
vpaddd ymm3, ymm3, ymm4
|
|
vpxord zmm4, zmm4, zmm4
|
|
vgatherdpd zmm4{k1}, [rdx+ymm3*8]
|
|
vgatherdpd zmm5{k2}, [8+rdx+ymm3*8]
|
|
vgatherdpd zmm6{k3}, [16+rdx+ymm3*8]
|
|
### SOA
|
|
#vpxord zmm4, zmm4, zmm4
|
|
#vgatherdpd zmm5{k2}, [rax+ymm3*8]
|
|
#vgatherdpd zmm4{k1}, [rdx+ymm3*8]
|
|
#vgatherdpd zmm6{k3}, [rsi+ymm3*8]
|
|
###
|
|
|
|
vsubpd zmm29, zmm1, zmm5 # zmm29 <- atom_y(i) - atom_y(j) -- dely
|
|
vsubpd zmm28, zmm0, zmm4 # zmm28 <- atom_x(i) - atom_x(j) -- delx
|
|
vsubpd zmm31, zmm2, zmm6 # zmm31 <- atom_z(i) - atom_z(j) -- delz
|
|
vmulpd zmm20, zmm29, zmm29 # zmm20 <- dely * dely
|
|
vfmadd231pd zmm20, zmm28, zmm28 # zmm20 <- dely * dely + delx * delx
|
|
vfmadd231pd zmm20, zmm31, zmm31 # zmm20 <- zmm20 + delz * delz -- rsq
|
|
|
|
# Cutoff radius condition
|
|
vrcp14pd zmm27, zmm20 # zmm27 <- 1.0 / rsq (sr2)
|
|
vcmppd k5, zmm20, zmm16, 1 # k5 <- rsq < cutforcesq
|
|
kmovw r9d, k5 # r9d <- rsq < cutforcesq
|
|
and r15d, r9d # r15d <- rsq < cutforcesq && k < numneighs
|
|
kmovw k3, r15d # k3 <- rsq < cutforcesq && k < numneighs
|
|
vmulpd zmm22, zmm27, zmm15 # zmm22 <- sr2 * sigma6
|
|
vmulpd zmm24, zmm27, zmm14 # zmm24 <- 48.0 * epsilon * sr2
|
|
vmulpd zmm25, zmm27, zmm22 # zmm25 <- sr2 * sigma6 * sr2
|
|
vmulpd zmm23, zmm27, zmm25 # zmm23 <- sr2 * sigma6 * sr2 * sr2
|
|
vfmsub213pd zmm27, zmm25, zmm7 # zmm27 <- sr2 * sigma * sr2 * sr2 - 0.5
|
|
vmulpd zmm26, zmm23, zmm24 # zmm26 <- 48.0 * epsilon * sr2 * sr2 * sigma6 * sr2
|
|
vmulpd zmm30, zmm26, zmm27 # zmm30 <- force
|
|
vfmadd231pd zmm13{k5}, zmm30, zmm28 # fix += force * delx
|
|
vfmadd231pd zmm12{k5}, zmm30, zmm29 # fiy += force * dely
|
|
vfmadd231pd zmm11{k5}, zmm30, zmm31 # fiz += force * delz
|
|
|
|
# Forces are currently separated in different lanes of zmm registers, hence it is necessary to permutate
|
|
# and add them (reduction) to obtain the final contribution for the current atom
|
|
..sum_up_forces:
|
|
vmovups zmm10, ZMMWORD PTR .L_2il0floatpacket.6[rip]
|
|
vpermd zmm0, zmm10, zmm11
|
|
vpermd zmm5, zmm10, zmm12
|
|
vpermd zmm21, zmm10, zmm13
|
|
vaddpd zmm11, zmm0, zmm11
|
|
vaddpd zmm12, zmm5, zmm12
|
|
vaddpd zmm13, zmm21, zmm13
|
|
vpermpd zmm1, zmm11, 78
|
|
vpermpd zmm6, zmm12, 78
|
|
vpermpd zmm22, zmm13, 78
|
|
vaddpd zmm2, zmm11, zmm1
|
|
vaddpd zmm8, zmm12, zmm6
|
|
vaddpd zmm23, zmm13, zmm22
|
|
vpermpd zmm3, zmm2, 177
|
|
vpermpd zmm9, zmm8, 177
|
|
vpermpd zmm24, zmm23, 177
|
|
vaddpd zmm4, zmm2, zmm3
|
|
vaddpd zmm20, zmm8, zmm9
|
|
vaddpd zmm25, zmm23, zmm24
|
|
|
|
..atom_loop_exit:
|
|
mov rcx, QWORD PTR [-8+rsp] #84.9[spill]
|
|
mov rbx, QWORD PTR [-16+rsp] #85.9[spill]
|
|
|
|
### AOS
|
|
add rax, 24
|
|
###
|
|
|
|
vaddsd xmm0, xmm25, QWORD PTR [rcx+r10*8] #84.9
|
|
vmovsd QWORD PTR [rcx+r10*8], xmm0 #84.9
|
|
vaddsd xmm1, xmm20, QWORD PTR [rbx+r10*8] #85.9
|
|
vmovsd QWORD PTR [rbx+r10*8], xmm1 #85.9
|
|
vaddsd xmm2, xmm4, QWORD PTR [rdi+r10*8] #86.9
|
|
vmovsd QWORD PTR [rdi+r10*8], xmm2 #86.9
|
|
inc r10 #55.5
|
|
cmp r10, QWORD PTR [-32+rsp] #55.5[spill]
|
|
jb ..atom_loop_begin
|
|
vzeroupper #93.12
|
|
vxorpd xmm0, xmm0, xmm0 #93.12
|
|
pop rbx
|
|
pop r15
|
|
pop r14 #93.12
|
|
pop r13 #93.12
|
|
pop r12 #93.12
|
|
pop rbp #93.12
|
|
ret #93.12
|
|
|
|
.type computeForce,@function
|
|
.size computeForce,.-computeForce
|
|
|
|
|
|
..LNcomputeForce.0:
|
|
.data
|
|
# -- End computeForce
|
|
.section .rodata, "a"
|
|
.align 64
|
|
.align 64
|
|
.L_2il0floatpacket.2:
|
|
.long 0x00000000,0x3ff00000,0x00000000,0x3ff00000,0x00000000,0x3ff00000,0x00000000,0x3ff00000,0x00000000,0x3ff00000,0x00000000,0x3ff00000,0x00000000,0x3ff00000,0x00000000,0x3ff00000
|
|
.type .L_2il0floatpacket.2,@object
|
|
.size .L_2il0floatpacket.2,64
|
|
.align 64
|
|
.L_2il0floatpacket.4:
|
|
.long 0x00000000,0x3fe00000,0x00000000,0x3fe00000,0x00000000,0x3fe00000,0x00000000,0x3fe00000,0x00000000,0x3fe00000,0x00000000,0x3fe00000,0x00000000,0x3fe00000,0x00000000,0x3fe00000
|
|
.type .L_2il0floatpacket.4,@object
|
|
.size .L_2il0floatpacket.4,64
|
|
.align 64
|
|
.L_2il0floatpacket.6:
|
|
.long 0x00000008,0x00000009,0x0000000a,0x0000000b,0x0000000c,0x0000000d,0x0000000e,0x0000000f,0x00000008,0x00000009,0x0000000a,0x0000000b,0x0000000c,0x0000000d,0x0000000e,0x0000000f
|
|
.type .L_2il0floatpacket.6,@object
|
|
.size .L_2il0floatpacket.6,64
|
|
.align 32
|
|
.L_2il0floatpacket.0:
|
|
.long 0x00000008,0x00000008,0x00000008,0x00000008,0x00000008,0x00000008,0x00000008,0x00000008
|
|
.type .L_2il0floatpacket.0,@object
|
|
.size .L_2il0floatpacket.0,32
|
|
.align 32
|
|
.L_2il0floatpacket.1:
|
|
.long 0x00000000,0x00000001,0x00000002,0x00000003,0x00000004,0x00000005,0x00000006,0x00000007
|
|
.type .L_2il0floatpacket.1,@object
|
|
.size .L_2il0floatpacket.1,32
|
|
.align 8
|
|
.L_2il0floatpacket.3:
|
|
.long 0x00000000,0x40480000
|
|
.type .L_2il0floatpacket.3,@object
|
|
.size .L_2il0floatpacket.3,8
|
|
.align 8
|
|
.L_2il0floatpacket.5:
|
|
.long 0x00000000,0x3ff00000
|
|
.type .L_2il0floatpacket.5,@object
|
|
.size .L_2il0floatpacket.5,8
|
|
.data
|
|
.section .note.GNU-stack, ""
|
|
# End
|