|
Packit |
5c3484 |
dnl AMD64 mpn_mul_1 optimised for Intel Atom.
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
dnl Copyright 2003-2005, 2007, 2008, 2012, 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 best
|
|
Packit |
5c3484 |
C AMD K8,K9
|
|
Packit |
5c3484 |
C AMD K10
|
|
Packit |
5c3484 |
C AMD bd1
|
|
Packit |
5c3484 |
C AMD bd2
|
|
Packit |
5c3484 |
C AMD bobcat
|
|
Packit |
5c3484 |
C AMD jaguar
|
|
Packit |
5c3484 |
C Intel P4
|
|
Packit |
5c3484 |
C Intel PNR
|
|
Packit |
5c3484 |
C Intel NHM
|
|
Packit |
5c3484 |
C Intel SBR
|
|
Packit |
5c3484 |
C Intel IBR
|
|
Packit |
5c3484 |
C Intel HWL
|
|
Packit |
5c3484 |
C Intel BWL
|
|
Packit |
5c3484 |
C Intel atom 17.3 this
|
|
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 optimisation tool suite written by David Harvey and Torbjorn Granlund.
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
define(`rp', `%rdi') C rcx
|
|
Packit |
5c3484 |
define(`up', `%rsi') C rdx
|
|
Packit |
5c3484 |
define(`n_param', `%rdx') C r8
|
|
Packit |
5c3484 |
define(`v0', `%rcx') C r9
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
define(`n', `%r11')
|
|
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(mpn_mul_1)
|
|
Packit |
5c3484 |
FUNC_ENTRY(4)
|
|
Packit |
5c3484 |
xor %r8, %r8
|
|
Packit |
5c3484 |
L(com): mov (up), %rax
|
|
Packit |
5c3484 |
lea -16(up,n_param,8), up
|
|
Packit |
5c3484 |
lea -8(rp,n_param,8), rp
|
|
Packit |
5c3484 |
test $1, R8(n_param)
|
|
Packit |
5c3484 |
jnz L(bx1)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(bx0): mov %r8, %r9
|
|
Packit |
5c3484 |
test $2, R8(n_param)
|
|
Packit |
5c3484 |
jnz L(b10)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(b00): mov $2, R32(n)
|
|
Packit |
5c3484 |
sub n_param, n
|
|
Packit |
5c3484 |
jmp L(lo0)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(bx1): test $2, R8(n_param)
|
|
Packit |
5c3484 |
jnz L(b11)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(b01): mov $3, R32(n)
|
|
Packit |
5c3484 |
sub n_param, n
|
|
Packit |
5c3484 |
mul v0
|
|
Packit |
5c3484 |
cmp $2, n
|
|
Packit |
5c3484 |
jnz L(lo1)
|
|
Packit |
5c3484 |
jmp L(cj1)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(b11): mov $1, R32(n)
|
|
Packit |
5c3484 |
sub n_param, n
|
|
Packit |
5c3484 |
jmp L(lo3)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(b10): xor R32(n), R32(n)
|
|
Packit |
5c3484 |
sub n_param, n
|
|
Packit |
5c3484 |
jmp L(lo2)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(top): mul v0
|
|
Packit |
5c3484 |
mov %r9, -24(rp,n,8)
|
|
Packit |
5c3484 |
L(lo1): xor %r9d, %r9d
|
|
Packit |
5c3484 |
add %rax, %r8
|
|
Packit |
5c3484 |
mov (up,n,8), %rax
|
|
Packit |
5c3484 |
adc %rdx, %r9
|
|
Packit |
5c3484 |
mov %r8, -16(rp,n,8)
|
|
Packit |
5c3484 |
L(lo0): xor %r8d, %r8d
|
|
Packit |
5c3484 |
mul v0
|
|
Packit |
5c3484 |
add %rax, %r9
|
|
Packit |
5c3484 |
mov 8(up,n,8), %rax
|
|
Packit |
5c3484 |
adc %rdx, %r8
|
|
Packit |
5c3484 |
mov %r9, -8(rp,n,8)
|
|
Packit |
5c3484 |
L(lo3): xor %r9d, %r9d
|
|
Packit |
5c3484 |
mul v0
|
|
Packit |
5c3484 |
add %rax, %r8
|
|
Packit |
5c3484 |
mov 16(up,n,8), %rax
|
|
Packit |
5c3484 |
adc %rdx, %r9
|
|
Packit |
5c3484 |
mov %r8, (rp,n,8)
|
|
Packit |
5c3484 |
L(lo2): xor %r8d, %r8d
|
|
Packit |
5c3484 |
mul v0
|
|
Packit |
5c3484 |
add %rax, %r9
|
|
Packit |
5c3484 |
mov 24(up,n,8), %rax
|
|
Packit |
5c3484 |
adc %rdx, %r8
|
|
Packit |
5c3484 |
add $4, n
|
|
Packit |
5c3484 |
js L(top)
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
L(end): mul v0
|
|
Packit |
5c3484 |
mov %r9, -8(rp)
|
|
Packit |
5c3484 |
L(cj1): add %rax, %r8
|
|
Packit |
5c3484 |
mov $0, R32(%rax)
|
|
Packit |
5c3484 |
adc %rdx, %rax
|
|
Packit |
5c3484 |
mov %r8, (rp)
|
|
Packit |
5c3484 |
FUNC_EXIT()
|
|
Packit |
5c3484 |
ret
|
|
Packit |
5c3484 |
EPILOGUE()
|
|
Packit |
5c3484 |
|
|
Packit |
5c3484 |
PROLOGUE(mpn_mul_1c)
|
|
Packit |
5c3484 |
FUNC_ENTRY(4)
|
|
Packit |
5c3484 |
IFDOS(` mov 56(%rsp), %r8 ')
|
|
Packit |
5c3484 |
jmp L(com)
|
|
Packit |
5c3484 |
EPILOGUE()
|
|
Packit |
5c3484 |
ASM_END()
|
|
rpm-build |
01f633 |
CF_PROT
|