diff options
author | Roland McGrath <roland@gnu.org> | 1995-10-16 01:37:51 +0000 |
---|---|---|
committer | Roland McGrath <roland@gnu.org> | 1995-10-16 01:37:51 +0000 |
commit | 8f5ca04bc7fd53741d80117df992995ace8f6d2d (patch) | |
tree | e39c13fc198b22ec55647259a8080051988e8c69 /sysdeps/i386/i586/rshift.S | |
parent | 5d82cf5c55f56ae10d3b0a205d1fcc7de1cf56a0 (diff) | |
download | glibc-8f5ca04bc7fd53741d80117df992995ace8f6d2d.tar.gz glibc-8f5ca04bc7fd53741d80117df992995ace8f6d2d.tar.xz glibc-8f5ca04bc7fd53741d80117df992995ace8f6d2d.zip |
Sat Oct 14 02:52:36 1995 Ulrich Drepper <drepper@ipd.info.uni-karlsruhe.de>
* malloc/malloc.c (_malloc_internal): Performance fix. Move if statement out of loop. * stdio/_itoa.c, stdio/_itoa.h: Complete rewrite. Much faster implementation using GMP functions. Contributed by Torbjorn Granlund and Ulrich Drepper. * stdio/test_rdwr.c: Include <errno.h>. * sysdeps/i386/i586/Implies: New file. New highly optimized string functions for i[345]86. * sysdeps/i386/memchr.S, sysdeps/i386/memcmp.S: New files. * sysdeps/i386/stpcpy.S, sysdeps/i386/stpncpy.S: New files. * sysdeps/i386/strchr.S, sysdeps/i386/strcspn.S: New files. * sysdeps/i386/strpbrk.S, sysdeps/i386/strrchr.S: New files. * sysdeps/i386/strspn.S, sysdeps/i386/i486/strcat.S: New files. * sysdeps/i386/i486/strlen.S, sysdeps/i386/i586/strchr.S: New files. * sysdeps/i386/i586/strlen.S: New file. * sysdeps/i386/memchr.c: Removed. There is now an assembler version. * sysdeps/i386/i586/memcopy.h (WORD_COPY_BWD): Parameters did not correspond to used values. * sysdeps/unix/sysv/linux/nfs/nfs.h: New file. Simply a wrapper around a kernel header file. * sysdeps/unix/sysv/linux/Dist: Add it. * sysdeps/unix/sysv/linux/Makefile [$(subdir)=sunrpc] (headers): Likewise. * sysdeps/unix/sysv/linux/local_lim.h: Rewrite. Instead of defining ourself we use a kernel header file. * sysdeps/unix/sysv/linux/i386/sysdep.h (DO_CALL): Optimize system call handler for i586. * sysdeps/unix/sysv/linux/sys/param.h: Add copyright and clean up. Sat Oct 14 02:52:36 1995 Ulrich Drepper <drepper@ipd.info.uni-karlsruhe.de> * malloc/malloc.c (_malloc_internal): Performance fix. Move if statement out of loop. * stdio/_itoa.c, stdio/_itoa.h: Complete rewrite. Much faster implementation using GMP functions. Contributed by Torbjorn Granlund and Ulrich Drepper. * stdio/test_rdwr.c: Include <errno.h>. * sysdeps/i386/i586/Implies: New file. New highly optimized string functions for i[345]86. * sysdeps/i386/memchr.S, sysdeps/i386/memcmp.S: New files. * sysdeps/i386/stpcpy.S, sysdeps/i386/stpncpy.S: New files. * sysdeps/i386/strchr.S, sysdeps/i386/strcspn.S: New files. * sysdeps/i386/strpbrk.S, sysdeps/i386/strrchr.S: New files. * sysdeps/i386/strspn.S, sysdeps/i386/i486/strcat.S: New files. * sysdeps/i386/i486/strlen.S, sysdeps/i386/i586/strchr.S: New files. * sysdeps/i386/i586/strlen.S: New file. * sysdeps/i386/memchr.c: Removed. There is now an assembler version. * sysdeps/i386/i586/memcopy.h (WORD_COPY_BWD): Parameters did not correspond to used values. * sysdeps/unix/sysv/linux/nfs/nfs.h: New file. Simply a wrapper around a kernel header file. * sysdeps/unix/sysv/linux/Dist: Add it. * sysdeps/unix/sysv/linux/Makefile [$(subdir)=sunrpc] (headers): Likewise. * sysdeps/unix/sysv/linux/local_lim.h: Rewrite. Instead of defining ourself we use a kernel header file. * sysdeps/unix/sysv/linux/i386/sysdep.h (DO_CALL): Optimize system call handler for i586. * sysdeps/unix/sysv/linux/sys/param.h: Add copyright and clean up.
Diffstat (limited to 'sysdeps/i386/i586/rshift.S')
-rw-r--r-- | sysdeps/i386/i586/rshift.S | 213 |
1 files changed, 213 insertions, 0 deletions
diff --git a/sysdeps/i386/i586/rshift.S b/sysdeps/i386/i586/rshift.S new file mode 100644 index 0000000000..51cde8f07f --- /dev/null +++ b/sysdeps/i386/i586/rshift.S @@ -0,0 +1,213 @@ +/* Pentium optimized __mpn_rshift -- + +Copyright (C) 1992, 1994, 1995 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 Library General Public License as published by +the Free Software Foundation; either version 2 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 Library General Public +License for more details. + +You should have received a copy of the GNU Library General Public License +along with the GNU MP Library; see the file COPYING.LIB. If not, write to +the Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* + INPUT PARAMETERS + res_ptr (sp + 4) + s_ptr (sp + 8) + size (sp + 12) + cnt (sp + 16) +*/ + +#include "sysdep.h" +#include "asm-syntax.h" + +.text + ALIGN (3) + .globl C_SYMBOL_NAME(__mpn_rshift) +C_SYMBOL_NAME(__mpn_rshift:) + pushl %edi + pushl %esi + pushl %ebx + pushl %ebp + + movl 20(%esp),%edi /* res_ptr */ + movl 24(%esp),%esi /* s_ptr */ + movl 28(%esp),%ebp /* size */ + movl 32(%esp),%ecx /* cnt */ + + cmp $1,%ecx + jne Lnormal + movl %edi,%eax + subl %esi,%eax + cmpl %ebp,%eax + jnc Lspecial + +Lnormal: + movl (%esi),%edx + addl $4,%esi + xorl %eax,%eax + shrdl %cl,%edx,%eax /* compute carry limb */ + pushl %eax /* push carry limb onto stack */ + + decl %ebp + pushl %ebp + shrl $3,%ebp + jz Lend + + movl (%edi),%eax /* fetch destination cache line */ + + ALIGN (2) +Loop: movl 28(%edi),%eax /* fetch destination cache line */ + movl %edx,%ebx + + movl (%esi),%eax + movl 4(%esi),%edx + shrdl %cl,%eax,%ebx + shrdl %cl,%edx,%eax + movl %ebx,(%edi) + movl %eax,4(%edi) + + movl 8(%esi),%ebx + movl 12(%esi),%eax + shrdl %cl,%ebx,%edx + shrdl %cl,%eax,%ebx + movl %edx,8(%edi) + movl %ebx,12(%edi) + + movl 16(%esi),%edx + movl 20(%esi),%ebx + shrdl %cl,%edx,%eax + shrdl %cl,%ebx,%edx + movl %eax,16(%edi) + movl %edx,20(%edi) + + movl 24(%esi),%eax + movl 28(%esi),%edx + shrdl %cl,%eax,%ebx + shrdl %cl,%edx,%eax + movl %ebx,24(%edi) + movl %eax,28(%edi) + + addl $32,%esi + addl $32,%edi + decl %ebp + jnz Loop + +Lend: popl %ebp + andl $7,%ebp + jz Lend2 +Loop2: movl (%esi),%eax + shrdl %cl,%eax,%edx /* compute result limb */ + movl %edx,(%edi) + movl %eax,%edx + addl $4,%esi + addl $4,%edi + decl %ebp + jnz Loop2 + +Lend2: shrl %cl,%edx /* compute most significant limb */ + movl %edx,(%edi) /* store it */ + + popl %eax /* pop carry limb */ + + popl %ebp + popl %ebx + popl %esi + popl %edi + ret + +/* We loop from least significant end of the arrays, which is only + permissable if the source and destination don't overlap, since the + function is documented to work for overlapping source and destination. +*/ + +Lspecial: + leal -4(%edi,%ebp,4),%edi + leal -4(%esi,%ebp,4),%esi + + movl (%esi),%edx + subl $4,%esi + + decl %ebp + pushl %ebp + shrl $3,%ebp + + shrl $1,%edx + incl %ebp + decl %ebp + jz LLend + + movl (%edi),%eax /* fetch destination cache line */ + + ALIGN (2) +LLoop: movl -28(%edi),%eax /* fetch destination cache line */ + movl %edx,%ebx + + movl (%esi),%eax + movl -4(%esi),%edx + rcrl $1,%eax + movl %ebx,(%edi) + rcrl $1,%edx + movl %eax,-4(%edi) + + movl -8(%esi),%ebx + movl -12(%esi),%eax + rcrl $1,%ebx + movl %edx,-8(%edi) + rcrl $1,%eax + movl %ebx,-12(%edi) + + movl -16(%esi),%edx + movl -20(%esi),%ebx + rcrl $1,%edx + movl %eax,-16(%edi) + rcrl $1,%ebx + movl %edx,-20(%edi) + + movl -24(%esi),%eax + movl -28(%esi),%edx + rcrl $1,%eax + movl %ebx,-24(%edi) + rcrl $1,%edx + movl %eax,-28(%edi) + + leal -32(%esi),%esi /* use leal not to clobber carry */ + leal -32(%edi),%edi + decl %ebp + jnz LLoop + +LLend: popl %ebp + sbbl %eax,%eax /* save carry in %eax */ + andl $7,%ebp + jz LLend2 + addl %eax,%eax /* restore carry from eax */ +LLoop2: movl %edx,%ebx + movl (%esi),%edx + rcrl $1,%edx + movl %ebx,(%edi) + + leal -4(%esi),%esi /* use leal not to clobber carry */ + leal -4(%edi),%edi + decl %ebp + jnz LLoop2 + + jmp LL1 +LLend2: addl %eax,%eax /* restore carry from eax */ +LL1: movl %edx,(%edi) /* store last limb */ + + movl $0,%eax + rcrl $1,%eax + + popl %ebp + popl %ebx + popl %esi + popl %edi + ret |