Optimizations from GMP.

* sysdeps/x86_64/add_n.S: New file.
	* sysdeps/x86_64/addmul_1.S: New file.
	* sysdeps/x86_64/lshift.S: New file.
	* sysdeps/x86_64/mul_1.S: New file.
	* sysdeps/x86_64/rshift.S: New file.
	* sysdeps/x86_64/sub_n.S: New file.
	* sysdeps/x86_64/submul_1.S: New file.
This commit is contained in:
Ulrich Drepper 2009-04-14 22:26:05 +00:00
parent 7fd23f1f3b
commit 893a5fd440
7 changed files with 309 additions and 0 deletions

View File

@ -1,3 +1,14 @@
2009-04-14 Ulrich Drepper <drepper@redhat.com>
Optimizations from GMP.
* sysdeps/x86_64/add_n.S: New file.
* sysdeps/x86_64/addmul_1.S: New file.
* sysdeps/x86_64/lshift.S: New file.
* sysdeps/x86_64/mul_1.S: New file.
* sysdeps/x86_64/rshift.S: New file.
* sysdeps/x86_64/sub_n.S: New file.
* sysdeps/x86_64/submul_1.S: New file.
2009-04-14 Jakub Jelinek <jakub@redhat.com>
* sysdeps/powerpc/powerpc32/elf/start.S: Ensure .data is non-empty.

46
sysdeps/x86_64/addmul_1.S Normal file
View File

@ -0,0 +1,46 @@
/* AMD64 __mpn_addmul_1 -- Multiply a limb vector with a limb and add
the result to a second limb vector.
Copyright (C) 2004 Free Software Foundation, Inc.
This file is part of the GNU MP Library.
The GNU MP 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 GNU MP 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 GNU MP Library; see the file COPYING.LIB. If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
MA 02111-1307, USA. */
#include "sysdep.h"
#include "asm-syntax.h"
.text
ENTRY (__mpn_addmul_1)
movq %rdx, %r11
leaq (%rsi,%rdx,8), %rsi
leaq (%rdi,%rdx,8), %rdi
negq %r11
xorl %r8d, %r8d
xorl %r10d, %r10d
.p2align 2
L(loop):
movq (%rsi,%r11,8), %rax
mulq %rcx
addq (%rdi,%r11,8), %rax
adcq %r10, %rdx
addq %r8, %rax
movq %r10, %r8
movq %rax, (%rdi,%r11,8)
adcq %rdx, %r8
incq %r11
jne L(loop)
movq %r8, %rax
ret
END (__mpn_addmul_1)

60
sysdeps/x86_64/lshift.S Normal file
View File

@ -0,0 +1,60 @@
/* AMD64 __mpn_lshift --
Copyright 2004, 2006 Free Software Foundation, Inc.
This file is part of the GNU MP Library.
The GNU MP 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 GNU MP 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 GNU MP Library; see the file COPYING.LIB. If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
MA 02111-1307, USA. */
#include "sysdep.h"
#include "asm-syntax.h"
.text
ENTRY (__mpn_lshift)
movq -8(%rsi,%rdx,8), %mm7
movd %ecx, %mm1
movl $64, %eax
subl %ecx, %eax
movd %eax, %mm0
movq %mm7, %mm3
psrlq %mm0, %mm7
movd %mm7, %rax
subq $2, %rdx
jl L(endo)
.p2align 2
L(loop):
movq (%rsi,%rdx,8), %mm6
movq %mm6, %mm2
psrlq %mm0, %mm6
psllq %mm1, %mm3
por %mm6, %mm3
movq %mm3, 8(%rdi,%rdx,8)
je L(ende)
movq -8(%rsi,%rdx,8), %mm7
movq %mm7, %mm3
psrlq %mm0, %mm7
psllq %mm1, %mm2
por %mm7, %mm2
movq %mm2, (%rdi,%rdx,8)
subq $2, %rdx
jge L(loop)
L(endo):
movq %mm3, %mm2
L(ende):
psllq %mm1, %mm2
movq %mm2, (%rdi)
emms
ret
END (__mpn_lshift)

42
sysdeps/x86_64/mul_1.S Normal file
View File

@ -0,0 +1,42 @@
/* AMD64 __mpn_mul_1 -- Multiply a limb vector with a limb and store
the result in a second limb vector.
Copyright (C) 2004 Free Software Foundation, Inc.
This file is part of the GNU MP Library.
The GNU MP 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 GNU MP 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 GNU MP Library; see the file COPYING.LIB. If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
MA 02111-1307, USA. */
#include <sysdep.h>
#include "asm-syntax.h"
.text
ENTRY (__mpn_mul_1)
movq %rdx, %r11
leaq (%rsi,%rdx,8), %rsi
leaq (%rdi,%rdx,8), %rdi
negq %r11
xorl %r8d, %r8d
L(loop):
movq (%rsi,%r11,8), %rax
mulq %rcx
addq %r8, %rax
movl $0, %r8d
adcq %rdx, %r8
movq %rax, (%rdi,%r11,8)
incq %r11
jne L(loop)
movq %r8, %rax
ret
END (__mpn_mul_1)

62
sysdeps/x86_64/rshift.S Normal file
View File

@ -0,0 +1,62 @@
/* AMD64 __mpn_rshift --
Copyright (C) 2004, 2006 Free Software Foundation, Inc.
This file is part of the GNU MP Library.
The GNU MP 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 GNU MP 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 GNU MP Library; see the file COPYING.LIB. If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
MA 02111-1307, USA. */
#include "sysdep.h"
#include "asm-syntax.h"
.text
ENTRY (__mpn_rshift)
movq (%rsi), %mm7
movd %ecx, %mm1
movl $64, %eax
subl %ecx, %eax
movd %eax, %mm0
movq %mm7, %mm3
psllq %mm0, %mm7
movd %mm7, %rax
leaq (%rsi,%rdx,8), %rsi
leaq (%rdi,%rdx,8), %rdi
negq %rdx
addq $2, %rdx
jg L(endo)
.p2align 2
L(loop):
movq -8(%rsi,%rdx,8), %mm6
movq %mm6, %mm2
psllq %mm0, %mm6
psrlq %mm1, %mm3
por %mm6, %mm3
movq %mm3, -16(%rdi,%rdx,8)
je L(ende)
movq (%rsi,%rdx,8), %mm7
movq %mm7, %mm3
psllq %mm0, %mm7
psrlq %mm1, %mm2
por %mm7, %mm2
movq %mm2, -8(%rdi,%rdx,8)
addq $2, %rdx
jle L(loop)
L(endo):
movq %mm3, %mm2
L(ende):
psrlq %mm1, %mm2
movq %mm2, -8(%rdi)
emms
ret
END (__mpn_rshift)

42
sysdeps/x86_64/sub_n.S Normal file
View File

@ -0,0 +1,42 @@
/* AMD64 __mpn_sub_n -- Add two limb vectors of the same length > 0 and store
sum in a third limb vector.
Copyright (C) 2004 Free Software Foundation, Inc.
This file is part of the GNU MP Library.
The GNU MP 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 GNU MP 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 GNU MP Library; see the file COPYING.LIB. If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
MA 02111-1307, USA. */
#include "sysdep.h"
#include "asm-syntax.h"
.text
ENTRY (__mpn_sub_n)
leaq (%rsi,%rcx,8), %rsi
leaq (%rdi,%rcx,8), %rdi
leaq (%rdx,%rcx,8), %rdx
negq %rcx
xorl %eax, %eax # clear cy
.p2align 2
L(loop):
movq (%rsi,%rcx,8), %rax
movq (%rdx,%rcx,8), %r10
sbbq %r10, %rax
movq %rax, (%rdi,%rcx,8)
incq %rcx
jne L(loop)
movq %rcx, %rax # zero %rax
adcq %rax, %rax
ret
END (__mpn_sub_n)

46
sysdeps/x86_64/submul_1.S Normal file
View File

@ -0,0 +1,46 @@
/* AMD64 __mpn_submul_1 -- Multiply a limb vector with a limb and subtract
the result from a second limb vector.
Copyright (C) 2004 Free Software Foundation, Inc.
This file is part of the GNU MP Library.
The GNU MP 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 GNU MP 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 GNU MP Library; see the file COPYING.LIB. If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
MA 02111-1307, USA. */
#include "sysdep.h"
#include "asm-syntax.h"
.text
ENTRY (__mpn_submul_1)
movq %rdx, %r11
leaq (%rsi,%r11,8), %rsi
leaq (%rdi,%r11,8), %rdi
negq %r11
xorl %r8d, %r8d
.p2align 3
L(loop):
movq (%rsi,%r11,8), %rax
movq (%rdi,%r11,8), %r10
mulq %rcx
subq %r8, %r10
movl $0, %r8d
adcl %r8d, %r8d
subq %rax, %r10
adcq %rdx, %r8
movq %r10, (%rdi,%r11,8)
incq %r11
jne L(loop)
movq %r8, %rax
ret
END (__mpn_submul_1)