Source code

Revision control

Copy as Markdown

Other Tools

; This Source Code Form is subject to the terms of the Mozilla Public
; License, v. 2.0. If a copy of the MPL was not distributed with this
; file, You can obtain one at http://mozilla.org/MPL/2.0/.
;
; This code is converted from mpi_amd64_gas.asm for MASM for x64.
;
; ------------------------------------------------------------------------
;
; Implementation of s_mpv_mul_set_vec which exploits
; the 64X64->128 bit unsigned multiply instruction.
;
; ------------------------------------------------------------------------
; r = a * digit, r and a are vectors of length len
; returns the carry digit
; r and a are 64 bit aligned.
;
; uint64_t
; s_mpv_mul_set_vec64(uint64_t *r, uint64_t *a, int len, uint64_t digit)
;
.CODE
s_mpv_mul_set_vec64 PROC
; compatibilities for paramenter registers
;
; About GAS and MASM, the usage of parameter registers are different.
push rdi
push rsi
mov rdi, rcx
mov rsi, rdx
mov edx, r8d
mov rcx, r9
xor rax, rax
test rdx, rdx
jz L17
mov r8, rdx
xor r9, r9
L15:
cmp r8, 8
jb L16
mov rax, [rsi]
mov r11, [8+rsi]
mul rcx
add rax, r9
adc rdx, 0
mov [0+rdi], rax
mov r9, rdx
mov rax,r11
mov r11, [16+rsi]
mul rcx
add rax,r9
adc rdx,0
mov [8+rdi],rax
mov r9,rdx
mov rax,r11
mov r11, [24+rsi]
mul rcx
add rax,r9
adc rdx,0
mov [16+rdi],rax
mov r9,rdx
mov rax,r11
mov r11, [32+rsi]
mul rcx
add rax,r9
adc rdx,0
mov [24+rdi],rax
mov r9,rdx
mov rax,r11
mov r11, [40+rsi]
mul rcx
add rax,r9
adc rdx,0
mov [32+rdi],rax
mov r9,rdx
mov rax,r11
mov r11, [48+rsi]
mul rcx
add rax,r9
adc rdx,0
mov [40+rdi],rax
mov r9,rdx
mov rax,r11
mov r11, [56+rsi]
mul rcx
add rax,r9
adc rdx,0
mov [48+rdi],rax
mov r9,rdx
mov rax,r11
mul rcx
add rax,r9
adc rdx,0
mov [56+rdi],rax
mov r9,rdx
add rsi, 64
add rdi, 64
sub r8, 8
jz L17
jmp L15
L16:
mov rax, [0+rsi]
mul rcx
add rax, r9
adc rdx,0
mov [0+rdi],rax
mov r9,rdx
dec r8
jz L17
mov rax, [8+rsi]
mul rcx
add rax,r9
adc rdx,0
mov [8+rdi], rax
mov r9, rdx
dec r8
jz L17
mov rax, [16+rsi]
mul rcx
add rax, r9
adc rdx, 0
mov [16+rdi],rax
mov r9,rdx
dec r8
jz L17
mov rax, [24+rsi]
mul rcx
add rax, r9
adc rdx, 0
mov [24+rdi], rax
mov r9, rdx
dec r8
jz L17
mov rax, [32+rsi]
mul rcx
add rax, r9
adc rdx, 0
mov [32+rdi],rax
mov r9, rdx
dec r8
jz L17
mov rax, [40+rsi]
mul rcx
add rax, r9
adc rdx, 0
mov [40+rdi], rax
mov r9, rdx
dec r8
jz L17
mov rax, [48+rsi]
mul rcx
add rax, r9
adc rdx, 0
mov [48+rdi], rax
mov r9, rdx
dec r8
jz L17
L17:
mov rax, r9
pop rsi
pop rdi
ret
s_mpv_mul_set_vec64 ENDP
;------------------------------------------------------------------------
;
; Implementation of s_mpv_mul_add_vec which exploits
; the 64X64->128 bit unsigned multiply instruction.
;
;------------------------------------------------------------------------
; r += a * digit, r and a are vectors of length len
; returns the carry digit
; r and a are 64 bit aligned.
;
; uint64_t
; s_mpv_mul_add_vec64(uint64_t *r, uint64_t *a, int len, uint64_t digit)
;
s_mpv_mul_add_vec64 PROC
; compatibilities for paramenter registers
;
; About GAS and MASM, the usage of parameter registers are different.
push rdi
push rsi
mov rdi, rcx
mov rsi, rdx
mov edx, r8d
mov rcx, r9
xor rax, rax
test rdx, rdx
jz L27
mov r8, rdx
xor r9, r9
L25:
cmp r8, 8
jb L26
mov rax, [0+rsi]
mov r10, [0+rdi]
mov r11, [8+rsi]
mul rcx
add rax,r10
adc rdx,0
mov r10, [8+rdi]
add rax,r9
adc rdx,0
mov [0+rdi],rax
mov r9,rdx
mov rax,r11
mov r11, [16+rsi]
mul rcx
add rax,r10
adc rdx,0
mov r10, [16+rdi]
add rax,r9
adc rdx,0
mov [8+rdi],rax
mov r9,rdx
mov rax,r11
mov r11, [24+rsi]
mul rcx
add rax,r10
adc rdx,0
mov r10, [24+rdi]
add rax,r9
adc rdx,0
mov [16+rdi],rax
mov r9,rdx
mov rax,r11
mov r11, [32+rsi]
mul rcx
add rax,r10
adc rdx,0
mov r10, [32+rdi]
add rax,r9
adc rdx,0
mov [24+rdi],rax
mov r9,rdx
mov rax,r11
mov r11, [40+rsi]
mul rcx
add rax,r10
adc rdx,0
mov r10, [40+rdi]
add rax,r9
adc rdx,0
mov [32+rdi],rax
mov r9,rdx
mov rax,r11
mov r11, [48+rsi]
mul rcx
add rax,r10
adc rdx,0
mov r10, [48+rdi]
add rax,r9
adc rdx,0
mov [40+rdi],rax
mov r9,rdx
mov rax,r11
mov r11, [56+rsi]
mul rcx
add rax,r10
adc rdx,0
mov r10, [56+rdi]
add rax,r9
adc rdx,0
mov [48+rdi],rax
mov r9,rdx
mov rax,r11
mul rcx
add rax,r10
adc rdx,0
add rax,r9
adc rdx,0
mov [56+rdi],rax
mov r9,rdx
add rsi,64
add rdi,64
sub r8, 8
jz L27
jmp L25
L26:
mov rax, [0+rsi]
mov r10, [0+rdi]
mul rcx
add rax,r10
adc rdx,0
add rax,r9
adc rdx,0
mov [0+rdi],rax
mov r9,rdx
dec r8
jz L27
mov rax, [8+rsi]
mov r10, [8+rdi]
mul rcx
add rax,r10
adc rdx,0
add rax,r9
adc rdx,0
mov [8+rdi],rax
mov r9,rdx
dec r8
jz L27
mov rax, [16+rsi]
mov r10, [16+rdi]
mul rcx
add rax,r10
adc rdx,0
add rax,r9
adc rdx,0
mov [16+rdi],rax
mov r9,rdx
dec r8
jz L27
mov rax, [24+rsi]
mov r10, [24+rdi]
mul rcx
add rax,r10
adc rdx,0
add rax,r9
adc rdx,0
mov [24+rdi],rax
mov r9,rdx
dec r8
jz L27
mov rax, [32+rsi]
mov r10, [32+rdi]
mul rcx
add rax,r10
adc rdx,0
add rax,r9
adc rdx,0
mov [32+rdi],rax
mov r9,rdx
dec r8
jz L27
mov rax, [40+rsi]
mov r10, [40+rdi]
mul rcx
add rax,r10
adc rdx,0
add rax,r9
adc rdx,0
mov [40+rdi],rax
mov r9,rdx
dec r8
jz L27
mov rax, [48+rsi]
mov r10, [48+rdi]
mul rcx
add rax,r10
adc rdx,0
add rax, r9
adc rdx, 0
mov [48+rdi], rax
mov r9, rdx
dec r8
jz L27
L27:
mov rax, r9
pop rsi
pop rdi
ret
s_mpv_mul_add_vec64 ENDP
END