diff options
Diffstat (limited to 'sysdeps/i386/i586/add_n.S')
| -rw-r--r-- | sysdeps/i386/i586/add_n.S | 133 |
1 files changed, 63 insertions, 70 deletions
diff --git a/sysdeps/i386/i586/add_n.S b/sysdeps/i386/i586/add_n.S index f52f9c60bc..f214c8cb36 100644 --- a/sysdeps/i386/i586/add_n.S +++ b/sysdeps/i386/i586/add_n.S @@ -1,7 +1,7 @@ /* Pentium __mpn_add_n -- Add two limb vectors of the same length > 0 and store sum in a third limb vector. -Copyright (C) 1992, 1994, 1995 Free Software Foundation, Inc. +Copyright (C) 1992, 1994, 1995, 1996 Free Software Foundation, Inc. This file is part of the GNU MP Library. @@ -30,13 +30,6 @@ the Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ #include "sysdep.h" #include "asm-syntax.h" -#define t1 %eax -#define t2 %edx -#define src1 %esi -#define src2 %ebp -#define dst %edi -#define x %ebx - .text ALIGN (3) .globl C_SYMBOL_NAME(__mpn_add_n) @@ -46,85 +39,85 @@ C_SYMBOL_NAME(__mpn_add_n:) pushl %ebx pushl %ebp - movl 20(%esp),dst /* res_ptr */ - movl 24(%esp),src1 /* s1_ptr */ - movl 28(%esp),src2 /* s2_ptr */ + movl 20(%esp),%edi /* res_ptr */ + movl 24(%esp),%esi /* s1_ptr */ + movl 28(%esp),%ebp /* s2_ptr */ movl 32(%esp),%ecx /* size */ - movl (src2),x + movl (%ebp),%ebx decl %ecx - movl %ecx,t2 + movl %ecx,%edx shrl $3,%ecx - andl $7,t2 + andl $7,%edx testl %ecx,%ecx /* zero carry flag */ jz Lend - pushl t2 + pushl %edx ALIGN (3) -Loop: movl 28(dst),%eax /* fetch destination cache line */ - leal 32(dst),dst - -L1: movl (src1),t1 - movl 4(src1),t2 - adcl x,t1 - movl 4(src2),x - adcl x,t2 - movl 8(src2),x - movl t1,-32(dst) - movl t2,-28(dst) - -L2: movl 8(src1),t1 - movl 12(src1),t2 - adcl x,t1 - movl 12(src2),x - adcl x,t2 - movl 16(src2),x - movl t1,-24(dst) - movl t2,-20(dst) - -L3: movl 16(src1),t1 - movl 20(src1),t2 - adcl x,t1 - movl 20(src2),x - adcl x,t2 - movl 24(src2),x - movl t1,-16(dst) - movl t2,-12(dst) - -L4: movl 24(src1),t1 - movl 28(src1),t2 - adcl x,t1 - movl 28(src2),x - adcl x,t2 - movl 32(src2),x - movl t1,-8(dst) - movl t2,-4(dst) - - leal 32(src1),src1 - leal 32(src2),src2 +Loop: movl 28(%edi),%eax /* fetch destination cache line */ + leal 32(%edi),%edi + +L1: movl (%esi),%eax + movl 4(%esi),%edx + adcl %ebx,%eax + movl 4(%ebp),%ebx + adcl %ebx,%edx + movl 8(%ebp),%ebx + movl %eax,-32(%edi) + movl %edx,-28(%edi) + +L2: movl 8(%esi),%eax + movl 12(%esi),%edx + adcl %ebx,%eax + movl 12(%ebp),%ebx + adcl %ebx,%edx + movl 16(%ebp),%ebx + movl %eax,-24(%edi) + movl %edx,-20(%edi) + +L3: movl 16(%esi),%eax + movl 20(%esi),%edx + adcl %ebx,%eax + movl 20(%ebp),%ebx + adcl %ebx,%edx + movl 24(%ebp),%ebx + movl %eax,-16(%edi) + movl %edx,-12(%edi) + +L4: movl 24(%esi),%eax + movl 28(%esi),%edx + adcl %ebx,%eax + movl 28(%ebp),%ebx + adcl %ebx,%edx + movl 32(%ebp),%ebx + movl %eax,-8(%edi) + movl %edx,-4(%edi) + + leal 32(%esi),%esi + leal 32(%ebp),%ebp decl %ecx jnz Loop - popl t2 + popl %edx Lend: - decl t2 /* test t2 w/o clobbering carry */ + decl %edx /* test %edx w/o clobbering carry */ js Lend2 - incl t2 + incl %edx Loop2: - leal 4(dst),dst - movl (src1),t1 - adcl x,t1 - movl 4(src2),x - movl t1,-4(dst) - leal 4(src1),src1 - leal 4(src2),src2 - decl t2 + leal 4(%edi),%edi + movl (%esi),%eax + adcl %ebx,%eax + movl 4(%ebp),%ebx + movl %eax,-4(%edi) + leal 4(%esi),%esi + leal 4(%ebp),%ebp + decl %edx jnz Loop2 Lend2: - movl (src1),t1 - adcl x,t1 - movl t1,(dst) + movl (%esi),%eax + adcl %ebx,%eax + movl %eax,(%edi) sbbl %eax,%eax negl %eax |