mpir/mpn/x86_64/k8/divrem_hensel_qr_1_2.asm
2009-09-30 23:41:09 +00:00

161 lines
3.7 KiB
NASM

dnl X86_64 mpn_divrem_hensel_qr_1_2
dnl Copyright 2009 Jason Moxham
dnl This file is part of the MPIR Library.
dnl The MPIR Library is free software; you can redistribute it and/or modify
dnl it under the terms of the GNU Lesser General Public License as published
dnl by the Free Software Foundation; either version 2.1 of the License, or (at
dnl your option) any later version.
dnl The MPIR Library is distributed in the hope that it will be useful, but
dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
dnl License for more details.
dnl You should have received a copy of the GNU Lesser General Public License
dnl along with the MPIR Library; see the file COPYING.LIB. If not, write
dnl to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
dnl Boston, MA 02110-1301, USA.
include(`../config.m4')
C (rdi,rdx)=(rsi,rdx) / rcx rdx>=1
C rax=hensel remainder from div
C This is divrem_hensel_1 with shifting on the output of the quotient
C On k8/k10 the shifting comes for free so no need to have different
C fn for that. And on K8/K10 this runs at 10c/l which is optimal
C This function "replaces" divexact_1 and modexact_1_odd
C This is same as the shifting version but with no shifting
ASM_START()
PROLOGUE(mpn_divrem_hensel_qr_1_2)
mov $1,%r9
sub %rdx,%r9
lea -8(%rdi,%rdx,8),%rdi
lea -8(%rsi,%rdx,8),%rsi #// last use of rdx
push %r12
push %r13
push %r14
mov %rcx,%rdx #// rdx is 3 bit inverse
mov %rdx,%rax
imul %ecx,%edx
mov $2,%r11
sub %rdx,%r11
imul %eax,%r11d #//r11 has 4 bits
mov %r11,%rax
imul %ecx,%r11d
mov $2,%rdx
sub %r11,%rdx
imul %eax,%edx #//rdx has 8 bits
mov %rdx,%rax
imul %ecx,%edx
mov $2,%r11
sub %rdx,%r11
imul %eax,%r11d #//r11 has 16 bits
mov %r11,%rax
imul %ecx,%r11d
mov $2,%rdx
sub %r11,%rdx
imul %eax,%edx #// rdx has 32 bits
mov %rdx,%rax
imul %rcx,%rdx
mov $2,%r11
sub %rdx,%r11
imul %rax,%r11 #//r11 has 64 bits
mov %r11,%rax
mov %r11,%r12
mul %rcx
neg %rdx
imul %rdx,%r12 #// r12,r11 has 128 bits
mov %r11,%r13
mov %r12,%r14
mov (%rsi,%r9,8),%r11
mov 8(%rsi,%r9,8),%r12
mov $0,%r10
add $2,%r9
jc skiplp
ALIGN(16)
lp:
mov %r12,%r8
mov %r13,%rax #// r13 low inverse
mul %r11
mov %rax,-16(%rdi,%r9,8) #// store low quotient
imul %r14,%r11 #// r14 high inverse
imul %r13,%r12
add %r11,%rdx
add %r12,%rdx
mov (%rsi,%r9,8),%r11
mov 8(%rsi,%r9,8),%r12
mov %rdx,-8(%rdi,%r9,8) #// store high quotient
mov %rcx,%rax #// rcx is divisor
mul %rdx #// (*) divisor*high quotient
add %r10,%r10
sbb $0,%r11
sbb $0,%r12
sbb %r10,%r10
cmp %rax,%r8 #// (*) and corrrection
sbb %rdx,%r11 #// (*) to give our "carry" not including old r10
sbb $0,%r12
sbb $0,%r10
add $2,%r9
jnc lp
skiplp:
mov %r12,%r8
mov %r13,%rax #// r13 low inverse
mul %r11
mov %rax,-16(%rdi,%r9,8) #// store low quotient
imul %r14,%r11 #// r14 high inverse
imul %r13,%r12
add %r11,%rdx
add %r12,%rdx
cmp $0,%r9
jne case0
case1:
mov (%rsi,%r9,8),%r11
mov %rdx,-8(%rdi,%r9,8) #// store high quotient
mov %rcx,%rax #// rcx is divisor
mul %rdx #// (*) divisor*high quotient
add %r10,%r10
sbb $0,%r11
sbb %r10,%r10
cmp %rax,%r8 #// (*) and corrrection
sbb %rdx,%r11 #// (*) to give our "carry" not including old r10
sbb $0,%r10
mov %r11,%rax
imul %r13,%rax
mov %rax,(%rdi,%r9,8)
mul %rcx
add %r10,%r10
mov $0,%rax
adc %rdx,%rax
pop %r14
pop %r13
pop %r12
ret
case0:
mov %rdx,-8(%rdi,%r9,8) #// store high quotient
mov %rcx,%rax #// rcx is divisor
mul %rdx #// (*) divisor*high quotient
cmp %rax,%r8 #// (*) and corrrection
mov $0,%rax
adc %rdx,%rax #// (*) to give our "carry" not including old r10
sub %r10,%rax
pop %r14
pop %r13
pop %r12
ret
EPILOGUE()