dnl AMD64 mpn_addmul_1 and mpn_submul_1 optimised for Intel Haswell.
dnl Contributed to the GNU project by Torbjörn Granlund.
dnl Copyright 2013 Free Software Foundation, Inc.
dnl This file is part of the GNU MP Library.
dnl
dnl The GNU MP Library is free software; you can redistribute it and/or modify
dnl it under the terms of either:
dnl
dnl * the GNU Lesser General Public License as published by the Free
dnl Software Foundation; either version 3 of the License, or (at your
dnl option) any later version.
dnl
dnl or
dnl
dnl * the GNU General Public License as published by the Free Software
dnl Foundation; either version 2 of the License, or (at your option) any
dnl later version.
dnl
dnl or both in parallel, as here.
dnl
dnl The GNU MP 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 General Public License
dnl for more details.
dnl
dnl You should have received copies of the GNU General Public License and the
dnl GNU Lesser General Public License along with the GNU MP Library. If not,
dnl see https://www.gnu.org/licenses/.
include(`../config.m4')
C cycles/limb
C AMD K8,K9 n/a
C AMD K10 n/a
C AMD bull n/a
C AMD pile n/a
C AMD steam ?
C AMD bobcat n/a
C AMD jaguar ?
C Intel P4 n/a
C Intel core n/a
C Intel NHM n/a
C Intel SBR n/a
C Intel IBR n/a
C Intel HWL 2.32
C Intel BWL ?
C Intel atom n/a
C VIA nano n/a
C The loop of this code is the result of running a code generation and
C optimisation tool suite written by David Harvey and Torbjörn Granlund.
C TODO
C * Handle small n separately, for lower overhead.
define(`rp', `%rdi') C rcx
define(`up', `%rsi') C rdx
define(`n_param', `%rdx') C r8
define(`v0_param',`%rcx') C r9
define(`n', `%rbp')
define(`v0', `%rdx')
ifdef(`OPERATION_addmul_1',`
define(`ADDSUB', `add')
define(`ADCSBB', `adc')
define(`func', `mpn_addmul_1')
')
ifdef(`OPERATION_submul_1',`
define(`ADDSUB', `sub')
define(`ADCSBB', `sbb')
define(`func', `mpn_submul_1')
')
ABI_SUPPORT(DOS64)
ABI_SUPPORT(STD64)
MULFUNC_PROLOGUE(mpn_addmul_1 mpn_submul_1)
ASM_START()
TEXT
ALIGN(16)
PROLOGUE(func)
FUNC_ENTRY(4)
push %rbx
push %rbp
push %r12
push %r13
mov n_param, n
mov v0_param, v0
test $1, R8(n)
jnz L(bx1)
L(bx0): shr $2, n
jc L(b10)
L(b00): mulx( (up), %r13, %r12)
mulx( 8,(up), %rbx, %rax)
add %r12, %rbx
adc $0, %rax
mov (rp), %r12
mov 8(rp), %rcx
mulx( 16,(up), %r9, %r8)
lea -16(rp), rp
lea 16(up), up
ADDSUB %r13, %r12
jmp L(lo0)
L(bx1): shr $2, n
jc L(b11)
L(b01): mulx( (up), %r11, %r10)
jnz L(gt1)
L(n1): ADDSUB %r11, (rp)
mov $0, R32(%rax)
adc %r10, %rax
jmp L(ret)
L(gt1): mulx( 8,(up), %r13, %r12)
mulx( 16,(up), %rbx, %rax)
lea 24(up), up
add %r10, %r13
adc %r12, %rbx
adc $0, %rax
mov (rp), %r10
mov 8(rp), %r12
mov 16(rp), %rcx
lea -8(rp), rp
ADDSUB %r11, %r10
jmp L(lo1)
L(b11): mulx( (up), %rbx, %rax)
mov (rp), %rcx
mulx( 8,(up), %r9, %r8)
lea 8(up), up
lea -24(rp), rp
inc n C adjust n
ADDSUB %rbx, %rcx
jmp L(lo3)
L(b10): mulx( (up), %r9, %r8)
mulx( 8,(up), %r11, %r10)
lea -32(rp), rp
mov $0, R32(%rax)
clc C clear cf
jz L(end) C depends on old shift
ALIGN(16)
L(top): adc %rax, %r9
lea 32(rp), rp
adc %r8, %r11
mulx( 16,(up), %r13, %r12)
mov (rp), %r8
mulx( 24,(up), %rbx, %rax)
lea 32(up), up
adc %r10, %r13
adc %r12, %rbx
adc $0, %rax
mov 8(rp), %r10
mov 16(rp), %r12
ADDSUB %r9, %r8
mov 24(rp), %rcx
mov %r8, (rp)
ADCSBB %r11, %r10
L(lo1): mulx( (up), %r9, %r8)
mov %r10, 8(rp)
ADCSBB %r13, %r12
L(lo0): mov %r12, 16(rp)
ADCSBB %rbx, %rcx
L(lo3): mulx( 8,(up), %r11, %r10)
mov %rcx, 24(rp)
dec n
jnz L(top)
L(end): adc %rax, %r9
adc %r8, %r11
mov 32(rp), %r8
mov %r10, %rax
adc $0, %rax
mov 40(rp), %r10
ADDSUB %r9, %r8
mov %r8, 32(rp)
ADCSBB %r11, %r10
mov %r10, 40(rp)
adc $0, %rax
L(ret): pop %r13
pop %r12
pop %rbp
pop %rbx
FUNC_EXIT()
ret
EPILOGUE()
CF_PROT