mpir/mpn/x86w/p6/divexact_1.asm
2010-08-13 12:54:25 +00:00

193 lines
4.9 KiB
NASM

; Copyright 2001, 2002 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.
;
; Translation of AT&T syntax code by Brian Gladman
%include "..\x86i.inc"
extern ___gmp_modlimb_invert_table
global ___gmpn_divexact_1
%ifdef DLL
export ___gmpn_divexact_1
%endif
%define PARAM_DIVISOR esp+frame+16
%define PARAM_SIZE esp+frame+12
%define PARAM_SRC esp+frame+8
%define PARAM_DST esp+frame+4
%define SAVE_EBX esp+frame-4
%define SAVE_ESI esp+frame-8
%define SAVE_EDI esp+frame-12
%define SAVE_EBP esp+frame-16
%define VAR_INVERSE esp+frame-20
%define STACK_SPACE 20
%define frame 0
section .text
align 16
___gmpn_divexact_1:
mov eax,[PARAM_DIVISOR]
sub esp,STACK_SPACE
FR_sesp STACK_SPACE
mov [SAVE_ESI],esi
mov esi,[PARAM_SRC]
mov [SAVE_EBX],ebx
mov ebx,[PARAM_SIZE]
bsf ecx,eax ; trailing twos
mov [SAVE_EBP],ebp
shr eax,cl ; d without twos
mov edx,eax
shr eax,1 ; d/2 without twos
mov [PARAM_DIVISOR],edx
and eax,127
%ifdef PIC
call Lmovl_eip_ebp
add ebp,_GLOBAL_OFFSET_TABLE_
mov ebp,[___gmp_modlimb_invert_table+edx+ebp]
movzx ebp,byte [eax+ebp] ; inv 8 bits
%else
movzx ebp,byte [___gmp_modlimb_invert_table+eax] ; inv 8 bits
%endif
lea eax,[ebp+ebp] ; 2*inv
imul ebp,ebp ; inv*inv
mov [SAVE_EDI],edi
mov edi,[PARAM_DST]
lea esi,[esi+ebx*4] ; src end
imul ebp,[PARAM_DIVISOR] ; inv*inv*d
sub eax,ebp ; inv = 2*inv - inv*inv*d
lea ebp,[eax+eax] ; 2*inv
imul eax,eax ; inv*inv
lea edi,[edi+ebx*4] ; dst end
neg ebx ; -size
mov [PARAM_DST],edi
imul eax,[PARAM_DIVISOR] ; inv*inv*d
sub ebp,eax ; inv = 2*inv - inv*inv*d
mov [VAR_INVERSE],ebp
mov eax,[esi+ebx*4] ; src[0]
or ecx,ecx
jnz Leven
jmp Lodd_entry ; ecx initial carry is zero
; The dependent chain here is
;
; subl %edx,%eax 1
; imull %ebp,%eax 4
; mull PARAM_DIVISOR 5
; ----
; total 10
;
; and this is the measured speed. No special scheduling is necessary,out
; of order execution hides the load latency.
;
; eax scratch (src limb)
; ebx counter,limbs,negative
; ecx carry bit
; edx carry limb,high of last product
; esi &src[size]
; edi &dst[size]
Lodd_top:
mul dword [PARAM_DIVISOR]
mov eax,[esi+ebx*4]
sub eax,ecx
sbb ecx,ecx
sub eax,edx
sbb ecx,0
Lodd_entry:
imul eax,[VAR_INVERSE]
mov [edi+ebx*4],eax
neg ecx
inc ebx
jnz Lodd_top
mov esi,[SAVE_ESI]
mov edi,[SAVE_EDI]
mov ebp,[SAVE_EBP]
mov ebx,[SAVE_EBX]
add esp,STACK_SPACE
ret
; eax src[0]
; ebx counter,limbs,negative
; ecx shift
Leven:
xor ebp,ebp ; initial carry bit
xor edx,edx ; initial carry limb (for size==1)
inc ebx
jz Leven_one
mov edi,[esi+ebx*4] ; src[1]
shrd eax,edi,cl
jmp Leven_entry
; eax scratch
; ebx counter,limbs,negative
; ecx shift
; edx scratch
; esi &src[size]
; edi &dst[size] and scratch
; ebp carry bit
Leven_top:
mov edi,[esi+ebx*4]
mul dword [PARAM_DIVISOR]
mov eax,[-4+esi+ebx*4]
shrd eax,edi,cl
sub eax,ebp
sbb ebp,ebp
sub eax,edx
sbb ebp,0
Leven_entry:
imul eax,[VAR_INVERSE]
mov edi,[PARAM_DST]
neg ebp
mov [-4+edi+ebx*4],eax
inc ebx
jnz Leven_top
mul dword [PARAM_DIVISOR]
mov eax,[-4+esi]
Leven_one:
shr eax,cl
mov esi,[SAVE_ESI]
sub eax,ebp
mov ebp,[SAVE_EBP]
sub eax,edx
mov ebx,[SAVE_EBX]
imul eax,[VAR_INVERSE]
mov [-4+edi],eax
mov edi,[SAVE_EDI]
add esp,STACK_SPACE
ret
%ifdef PIC
Lmovl_eip_ebp:
mov ebp,[esp]
ret
%endif
end