|
Packit |
5c3484 |
dnl AMD64 mpn_add_n, mpn_sub_n optimised for bobcat.
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
dnl Copyright 2003-2005, 2007, 2008, 2010-2013 Free Software Foundation, Inc.
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
dnl This file is part of the GNU MP Library.
|
|
Packit |
5c3484 |
dnl
|
|
Packit |
5c3484 |
dnl The GNU MP Library is free software; you can redistribute it and/or modify
|
|
Packit |
5c3484 |
dnl it under the terms of either:
|
|
Packit |
5c3484 |
dnl
|
|
Packit |
5c3484 |
dnl * the GNU Lesser General Public License as published by the Free
|
|
Packit |
5c3484 |
dnl Software Foundation; either version 3 of the License, or (at your
|
|
Packit |
5c3484 |
dnl option) any later version.
|
|
Packit |
5c3484 |
dnl
|
|
Packit |
5c3484 |
dnl or
|
|
Packit |
5c3484 |
dnl
|
|
Packit |
5c3484 |
dnl * the GNU General Public License as published by the Free Software
|
|
Packit |
5c3484 |
dnl Foundation; either version 2 of the License, or (at your option) any
|
|
Packit |
5c3484 |
dnl later version.
|
|
Packit |
5c3484 |
dnl
|
|
Packit |
5c3484 |
dnl or both in parallel, as here.
|
|
Packit |
5c3484 |
dnl
|
|
Packit |
5c3484 |
dnl The GNU MP Library is distributed in the hope that it will be useful, but
|
|
Packit |
5c3484 |
dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
|
|
Packit |
5c3484 |
dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
|
|
Packit |
5c3484 |
dnl for more details.
|
|
Packit |
5c3484 |
dnl
|
|
Packit |
5c3484 |
dnl You should have received copies of the GNU General Public License and the
|
|
Packit |
5c3484 |
dnl GNU Lesser General Public License along with the GNU MP Library. If not,
|
|
Packit |
5c3484 |
dnl see https://www.gnu.org/licenses/.
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
include(`../config.m4')
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
C cycles/limb
|
|
Packit |
5c3484 |
C AMD K8,K9
|
|
Packit |
5c3484 |
C AMD K10
|
|
Packit |
5c3484 |
C AMD bd1
|
|
Packit |
5c3484 |
C AMD bobcat 2.28
|
|
Packit |
5c3484 |
C Intel P4
|
|
Packit |
5c3484 |
C Intel core2
|
|
Packit |
5c3484 |
C Intel NHM
|
|
Packit |
5c3484 |
C Intel SBR
|
|
Packit |
5c3484 |
C Intel IBR
|
|
Packit |
5c3484 |
C Intel atom
|
|
Packit |
5c3484 |
C VIA nano
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
C The loop of this code is the result of running a code generation and
|
|
Packit |
5c3484 |
C optimization tool suite written by David Harvey and Torbjorn Granlund.
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
C INPUT PARAMETERS
|
|
Packit |
5c3484 |
define(`rp', `%rdi') C rcx
|
|
Packit |
5c3484 |
define(`up', `%rsi') C rdx
|
|
Packit |
5c3484 |
define(`vp', `%rdx') C r8
|
|
Packit |
5c3484 |
define(`n', `%rcx') C r9
|
|
Packit |
5c3484 |
define(`cy', `%r8') C rsp+40 (mpn_add_nc and mpn_sub_nc)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
ifdef(`OPERATION_add_n', `
|
|
Packit |
5c3484 |
define(ADCSBB, adc)
|
|
Packit |
5c3484 |
define(func, mpn_add_n)
|
|
Packit |
5c3484 |
define(func_nc, mpn_add_nc)')
|
|
Packit |
5c3484 |
ifdef(`OPERATION_sub_n', `
|
|
Packit |
5c3484 |
define(ADCSBB, sbb)
|
|
Packit |
5c3484 |
define(func, mpn_sub_n)
|
|
Packit |
5c3484 |
define(func_nc, mpn_sub_nc)')
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
MULFUNC_PROLOGUE(mpn_add_n mpn_add_nc mpn_sub_n mpn_sub_nc)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
ABI_SUPPORT(DOS64)
|
|
Packit |
5c3484 |
ABI_SUPPORT(STD64)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
ASM_START()
|
|
Packit |
5c3484 |
TEXT
|
|
Packit |
5c3484 |
ALIGN(16)
|
|
Packit |
5c3484 |
PROLOGUE(func)
|
|
Packit |
5c3484 |
FUNC_ENTRY(4)
|
|
Packit |
5c3484 |
xor %r8, %r8
|
|
Packit |
5c3484 |
L(ent): test $1, R8(n)
|
|
Packit |
5c3484 |
jnz L(bx1)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(bx0): test $2, R8(n)
|
|
Packit |
5c3484 |
jnz L(b10)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(b00): shr $2, n
|
|
Packit |
5c3484 |
neg %r8
|
|
Packit |
5c3484 |
mov $3, R32(%rax)
|
|
Packit |
5c3484 |
mov (up), %r10
|
|
Packit |
5c3484 |
mov 8(up), %r11
|
|
Packit |
5c3484 |
jmp L(lo0)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(b10): shr $2, n
|
|
Packit |
5c3484 |
neg %r8
|
|
Packit |
5c3484 |
mov $1, R32(%rax)
|
|
Packit |
5c3484 |
mov (up), %r8
|
|
Packit |
5c3484 |
mov 8(up), %r9
|
|
Packit |
5c3484 |
jrcxz L(cj2)
|
|
Packit |
5c3484 |
jmp L(top)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(bx1): test $2, R8(n)
|
|
Packit |
5c3484 |
jnz L(b11)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(b01): shr $2, n
|
|
Packit |
5c3484 |
neg %r8
|
|
Packit |
5c3484 |
mov $0, R32(%rax)
|
|
Packit |
5c3484 |
mov (up), %r9
|
|
Packit |
5c3484 |
jrcxz L(cj1)
|
|
Packit |
5c3484 |
mov 8(up), %r10
|
|
Packit |
5c3484 |
jmp L(lo1)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
ALIGN(8)
|
|
Packit |
5c3484 |
L(b11): inc n
|
|
Packit |
5c3484 |
shr $2, n
|
|
Packit |
5c3484 |
neg %r8
|
|
Packit |
5c3484 |
mov $2, R32(%rax)
|
|
Packit |
5c3484 |
mov (up), %r11
|
|
Packit |
5c3484 |
jmp L(lo3)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
ALIGN(4)
|
|
Packit |
5c3484 |
L(top): mov 8(up,%rax,8), %r10
|
|
Packit |
5c3484 |
ADCSBB -8(vp,%rax,8), %r8
|
|
Packit |
5c3484 |
mov %r8, -8(rp,%rax,8)
|
|
Packit |
5c3484 |
L(lo1): mov 16(up,%rax,8), %r11
|
|
Packit |
5c3484 |
ADCSBB (vp,%rax,8), %r9
|
|
Packit |
5c3484 |
lea 4(%rax), %rax
|
|
Packit |
5c3484 |
mov %r9, -32(rp,%rax,8)
|
|
Packit |
5c3484 |
L(lo0): ADCSBB -24(vp,%rax,8), %r10
|
|
Packit |
5c3484 |
mov %r10, -24(rp,%rax,8)
|
|
Packit |
5c3484 |
L(lo3): ADCSBB -16(vp,%rax,8), %r11
|
|
Packit |
5c3484 |
dec n
|
|
Packit |
5c3484 |
mov -8(up,%rax,8), %r8
|
|
Packit |
5c3484 |
mov %r11, -16(rp,%rax,8)
|
|
Packit |
5c3484 |
L(lo2): mov (up,%rax,8), %r9
|
|
Packit |
5c3484 |
jnz L(top)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(cj2): ADCSBB -8(vp,%rax,8), %r8
|
|
Packit |
5c3484 |
mov %r8, -8(rp,%rax,8)
|
|
Packit |
5c3484 |
L(cj1): ADCSBB (vp,%rax,8), %r9
|
|
Packit |
5c3484 |
mov %r9, (rp,%rax,8)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
mov $0, R32(%rax)
|
|
Packit |
5c3484 |
adc $0, R32(%rax)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
FUNC_EXIT()
|
|
Packit |
5c3484 |
ret
|
|
Packit |
5c3484 |
EPILOGUE()
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
ALIGN(16)
|
|
Packit |
5c3484 |
PROLOGUE(func_nc)
|
|
Packit |
5c3484 |
FUNC_ENTRY(4)
|
|
Packit |
5c3484 |
IFDOS(` mov 56(%rsp), %r8 ')
|
|
Packit |
5c3484 |
jmp L(ent)
|
|
Packit |
5c3484 |
EPILOGUE()
|
|
rpm-build |
01f633 |
CF_PROT
|