mpir/mpn/x86_64/k8/sumdiff_n.as

100 lines
2.5 KiB
ActionScript

; AMD64 mpn_sumdiff_n
; Copyright 2009 Jason Moxham
; This file is part of the MPIR Library.
; The MPIR Library is free software; you can redistribute it and/or modify
; it under the terms of the GNU Lesser General Public License as published
; by the Free Software Foundation; either version 2.1 of the License, or (at
; your option) any later version.
; The MPIR Library is distributed in the hope that it will be useful, but
; WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
; or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
; License for more details.
; You should have received a copy of the GNU Lesser General Public License
; along with the MPIR Library; see the file COPYING.LIB. If not, write
; to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
; Boston, MA 02110-1301, USA.
; (rdi, r8) = (rdx, r8) + (rcx, r8) (rsi, r8) = (rdx, r8) - (rcx, r8)
; return 2*add_carry + sub_borrow
%include 'yasm_mac.inc'
BITS 64
GLOBAL_FUNC mpn_sumdiff_n
lea rsi, [rsi+r8*8]
lea rdx, [rdx+r8*8]
lea rdi, [rdi+r8*8]
lea rcx, [rcx+r8*8]
neg r8
xor r9, r9
xor r10, r10
test r8, 3
jz next
lp1:
mov rax, [rdx+r8*8]
mov r11, rax
add r9, 1
adc rax, [rcx+r8*8]
sbb r9, r9
add r10, 1
sbb r11, [rcx+r8*8]
sbb r10, r10
mov [rdi+r8*8], rax
mov [rsi+r8*8], r11
inc r8
test r8, 3
jnz lp1
next:
cmp r8, 0
jz skiploop
mov [rsp-8], rbx
mov [rsp-16], rbp
mov [rsp-24], r12
mov [rsp-32], r13
mov [rsp-40], r14
mov [rsp-48], r15
align 16
loop1:
mov rax, [rdx+r8*8]
mov rbx, [rdx+r8*8+8]
mov rbp, [rdx+r8*8+16]
mov r12, [rdx+r8*8+24]
mov r11, rax
mov r13, rbx
mov r14, rbp
mov r15, r12
add r9, 1
adc rax, [rcx+r8*8]
adc rbx, [rcx+r8*8+8]
adc rbp, [rcx+r8*8+16]
adc r12, [rcx+r8*8+24]
sbb r9, r9
add r10, 1
sbb r11, [rcx+r8*8]
sbb r13, [rcx+r8*8+8]
sbb r14, [rcx+r8*8+16]
sbb r15, [rcx+r8*8+24]
sbb r10, r10
mov [rdi+r8*8], rax
mov [rdi+r8*8+8], rbx
mov [rdi+r8*8+16], rbp
mov [rdi+r8*8+24], r12
mov [rsi+r8*8], r11
mov [rsi+r8*8+8], r13
mov [rsi+r8*8+16], r14
mov [rsi+r8*8+24], r15
add r8, 4
jnz loop1
mov rbx, [rsp-8]
mov rbp, [rsp-16]
mov r12, [rsp-24]
mov r13, [rsp-32]
mov r14, [rsp-40]
mov r15, [rsp-48]
skiploop:
lea rax, [r10+r9*2]
neg rax
ret