mpir/mpn/x86_64/k8/sqr_basecase.as

482 lines
9.2 KiB
ActionScript

; AMD64 mpn_sqr_basecase
; Copyright 2009 Jason Moxham
; This file is part of the MPIR Library.
; The MPIR 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 MPIR 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 MPIR Library; see the file COPYING.LIB. If not, write
; to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
; Boston, MA 02110-1301, USA.
; (rdi, 2*rdx) = (rsi,rdx)^2
; Version 1.0.5
; same as the addmul for now
; changes from standard mul
; change r8 to r12 and rcx to r13
; reemove ret and write last limb
%include 'yasm_mac.inc'
BITS 64
%macro mulloop 1
align 16
%%1:
mov r10, 0
mul r13
mov [rdi+r11*8], r12
add r9, rax
db 0x26
adc r10, rdx
mov rax, [rsi+r11*8+16]
mul r13
mov [rdi+r11*8+8], r9
add r10, rax
mov ebx, 0
adc rbx, rdx
mov rax, [rsi+r11*8+24]
mov r12, 0
mov r9, 0
mul r13
mov [rdi+r11*8+16], r10
db 0x26
add rbx, rax
db 0x26
adc r12, rdx
mov rax, [rsi+r11*8+32]
mul r13
mov [rdi+r11*8+24], rbx
db 0x26
add r12, rax
db 0x26
adc r9, rdx
add r11, 4
mov rax, [rsi+r11*8+8]
jnc %%1
%endmacro
%macro mulnext0 0
mov rax, [rsi+r11*8+16]
mul r13
mov [rdi+r11*8+8], r9
add r10, rax
mov ebx, 0
adc rbx, rdx
mov rax, [rsi+r11*8+24]
mov r12d, 0
mul r13
mov [rdi+r11*8+16], r10
add rbx, rax
adc r12, rdx
mov rax, [rsi+r11*8+32]
mul r13
mov [rdi+r11*8+24], rbx
add r12, rax
adc rdx, 0
mov [rdi+r11*8+32], r12
mov [rdi+r11*8+40], rdx
inc r14
lea rdi, [rdi+8]
%endmacro
%macro mulnext1 0
mov rax, [rsi+r11*8+16]
mul r13
mov [rdi+r11*8+8], r9
add r10, rax
mov r12d, 0
adc r12, rdx
mov rax, [rsi+r11*8+24]
mul r13
mov [rdi+r11*8+16], r10
add r12, rax
adc rdx, 0
mov [rdi+r11*8+24], r12
mov [rdi+r11*8+32], rdx
inc r14
lea rdi, [rdi+8]
%endmacro
%macro mulnext2 0
mov rax, [rsi+r11*8+16]
mul r13
mov [rdi+r11*8+8], r9
add r10, rax
mov ebx, 0
adc rbx, rdx
mov [rdi+r11*8+16], r10
mov [rdi+r11*8+24], rbx
inc r14
lea rdi, [rdi+8]
%endmacro
%macro mulnext3 0
mov [rdi+r11*8+8], r9
mov [rdi+r11*8+16], r10
inc r14
lea rdi, [rdi+8]
%endmacro
; changes from standard addmul
; change r8 to r12 and rcx to r13
; reemove ret and write last limb
%macro addmulloop 1
align 16
%%1:
mov r10, 0
mul r13
add [rdi+r11*8], r12
adc r9, rax
db 0x26
adc r10, rdx
mov rax, [rsi+r11*8+16]
mul r13
add [rdi+r11*8+8], r9
adc r10, rax
mov ebx, 0
adc rbx, rdx
mov rax, [rsi+r11*8+24]
mov r12, 0
mov r9, 0
mul r13
add [rdi+r11*8+16], r10
db 0x26
adc rbx, rax
db 0x26
adc r12, rdx
mov rax, [rsi+r11*8+32]
mul r13
add [rdi+r11*8+24], rbx
db 0x26
adc r12, rax
db 0x26
adc r9, rdx
add r11, 4
mov rax, [rsi+r11*8+8]
jnc %%1
%endmacro
%macro addmulnext0 0
mov r10d, 0
mul r13
add [rdi+r11*8], r12
adc r9, rax
adc r10, rdx
mov rax, [rsi+r11*8+16]
mul r13
add [rdi+r11*8+8], r9
adc r10, rax
mov ebx, 0
adc rbx, rdx
mov rax, [rsi+r11*8+24]
mov r12d, 0
mul r13
add [rdi+r11*8+16], r10
adc rbx, rax
adc r12, rdx
mov rax, [rsi+r11*8+32]
mul r13
add [rdi+r11*8+24], rbx
adc r12, rax
adc rdx, 0
add [rdi+r11*8+32], r12
adc rdx, 0
inc r14
mov [rdi+r11*8+40], rdx
lea rdi, [rdi+8]
%endmacro
%macro addmulnext1 0
mov r10d, 0
mul r13
add [rdi+r11*8], r12
adc r9, rax
adc r10, rdx
mov rax, [rsi+r11*8+16]
mul r13
add [rdi+r11*8+8], r9
adc r10, rax
mov r12d, 0
adc r12, rdx
mov rax, [rsi+r11*8+24]
mul r13
add [rdi+r11*8+16], r10
adc r12, rax
adc rdx, 0
add [rdi+r11*8+24], r12
adc rdx, 0
mov [rdi+r11*8+32], rdx
inc r14
lea rdi, [rdi+8]
%endmacro
%macro addmulnext2 0
mov r10d, 0
mul r13
add [rdi+r11*8], r12
adc r9, rax
adc r10, rdx
mov rax, [rsi+r11*8+16]
mul r13
mov ebx, 0
add [rdi+r11*8+8], r9
adc r10, rax
adc rbx, rdx
add [rdi+r11*8+16], r10
adc rbx, 0
mov [rdi+r11*8+24], rbx
inc r14
lea rdi, [rdi+8]
%endmacro
%macro addmulnext3 0
mul r13
add [rdi+r11*8], r12
adc r9, rax
mov r10d, 0
adc r10, rdx
add [rdi+r11*8+8], r9
adc r10, 0
mov [rdi+r11*8+16], r10
inc r14
lea rdi, [rdi+8]
cmp r14, 4
jnz theloop
%endmacro
GLOBAL_FUNC mpn_sqr_basecase
cmp rdx, 3
ja fourormore
jz three
jp two
one:
mov rax, [rsi]
mul rax
mov [rdi], rax
mov [rdi+8], rdx
ret
align 16
fourormore:
; this code can not handle cases 3,2,1
mov [rsp-8], r12
mov [rsp-16], r13
mov [rsp-24], r14
mov [rsp-32], rbx
; save data for later
mov [rsp-40], rdi
mov [rsp-48], rsi
mov [rsp-56], rdx
mov r13, [rsi]
mov rax, [rsi+8]
mov r14d, 6
sub r14, rdx
lea rdi, [rdi+rdx*8-40]
lea rsi, [rsi+rdx*8-40]
mov r11, r14
mul r13
mov r12, rax
mov rax, [rsi+r14*8+8]
mov r9, rdx
cmp r14, 0
jge mulskiploop1
mulloop 1
mulskiploop1:
mov r10d, 0
mul r13
mov [rdi+r11*8], r12
add r9, rax
adc r10, rdx
; could save r9 here
; could update here ie lea 8(rdi),rdi and inc r14
cmp r11, 2
je mcase2
ja mcase3
jp mcase1
mcase0:
mulnext0
jmp case1
align 16
mcase1:
mulnext1
jmp case2
align 16
mcase2:
mulnext2
jmp case3
align 16
mcase3:
mulnext3
; jmp case0 just fall thru
align 16
theloop:
case0:
mov rax, [rsi+r14*8]
mov r13, [rsi+r14*8-8]
mov r11, r14
mul r13
mov r12, rax
mov rax, [rsi+r14*8+8]
mov r9, rdx
cmp r14, 0
jge addmulskiploop0
addmulloop 0
addmulskiploop0:
addmulnext0
case1:
mov rax, [rsi+r14*8]
mov r13, [rsi+r14*8-8]
mov r11, r14
mul r13
mov r12, rax
mov rax, [rsi+r14*8+8]
mov r9, rdx
cmp r14, 0
jge addmulskiploop1
addmulloop 1
addmulskiploop1:
addmulnext1
case2:
mov rax, [rsi+r14*8]
mov r13, [rsi+r14*8-8]
mov r11, r14
mul r13
mov r12, rax
mov rax, [rsi+r14*8+8]
mov r9, rdx
cmp r14, 0
jge addmulskiploop2
addmulloop 2
addmulskiploop2:
addmulnext2
case3:
mov rax, [rsi+r14*8]
mov r13, [rsi+r14*8-8]
mov r11, r14
mul r13
mov r12, rax
mov rax, [rsi+r14*8+8]
mov r9, rdx
cmp r14, 0
jge addmulskiploop3
addmulloop 3
addmulskiploop3:
addmulnext3
; only need to add one more line here
mov rax, [rsi+r14*8]
mov r13, [rsi+r14*8-8]
mul r13
add [rdi+r14*8], rax
adc rdx, 0
mov [rdi+r14*8+8], rdx
; now lsh by 1 and add in the diagonal
mov rdi, [rsp-40]
mov rsi, [rsp-48]
mov rcx, [rsp-56]
mov r12, [rsp-8]
mov r13, [rsp-16]
xor rbx, rbx
xor r11, r11
lea rsi, [rsi+rcx*8]
mov [rdi], r11
lea r10, [rdi+rcx*8]
mov [r10+rcx*8-8], r11
neg rcx
align 16
dialoop:
mov rax, [rsi+rcx*8]
mul rax
mov r8, [rdi]
mov r9, [rdi+8]
add rbx, 1
adc r8, r8
adc r9, r9
sbb rbx, rbx
add r11, 1
adc r8, rax
adc r9, rdx
sbb r11, r11
mov [rdi], r8
mov [rdi+8], r9
inc rcx
lea rdi, [rdi+16]
jnz dialoop
mov rbx, [rsp-32]
mov r14, [rsp-24]
ret
align 16
two:
mov rax, [rsi]
mov r9, [rsi+8]
mov r8, rax
mul rax
mov [rdi], rax
mov rax, r9
mov [rdi+8], rdx
mul rax
mov [rdi+16], rax
mov rax, r8
mov r10, rdx
mul r9
add rax, rax
adc rdx, rdx
adc r10, 0
add [rdi+8], rax
adc [rdi+16], rdx
adc r10, 0
mov [rdi+24], r10
ret
align 16
three:
mov r8, [rsi]
mov rax, [rsi+8]
mul r8
mov r11d, 0
mov [rdi+8], rax
mov rax, [rsi+16]
mov r9, rdx
mul r8
mov r8, [rsi+8]
add r9, rax
mov rax, [rsi+16]
adc r11, rdx
mul r8
mov [rdi+16], r9
add r11, rax
mov r9d, 0
mov [rdi+24], r11
adc r9, rdx
mov [rdi+32], r9
mov rcx, -3
xor r10, r10
xor r11, r11
lea rsi, [rsi+24]
mov [rdi], r11
mov [rdi+40], r11
dialoop1:
mov rax, [rsi+rcx*8]
mul rax
mov r8, [rdi]
mov r9, [rdi+8]
add r10, 1
adc r8, r8
adc r9, r9
sbb r10, r10
add r11, 1
adc r8, rax
adc r9, rdx
sbb r11, r11
mov [rdi], r8
mov [rdi+8], r9
inc rcx
lea rdi, [rdi+16]
jnz dialoop1
nop
ret