mpir/mpn/x86_64w/sqr_basecase.asm
2014-02-17 23:02:51 +00:00

646 lines
12 KiB
NASM

; void mpn_sqr_basecase (mp_ptr, mp_srcptr, mp_size_t)
; rax rdi rsi rdx
; rax rcx rdx r8
%include 'yasm_mac.inc'
%define reg_save_list rbx, rbp, rsi, rdi, r12, r13, r14
%macro jmp_tab 2
%ifdef PIC
dd %1 - %2
%else
dq %1
%endif
%endmacro
%macro do_jmp 0
lea rax, [rel L0]
%ifdef PIC
movzx r10, [rax+rcx*4]
add rax, r10
jmp rax
%else
jmp [rax+rcx*8]
%endif
%endmacro
text
xalign 16
WIN64_GCC_PROC mpn_sqr_basecase, 3
mov ecx, edx
mov r11d, edx
and ecx, 3
cmp edx, 4
lea r8, [rcx+4]
cmovg rcx, r8
do_jmp
xalign 8
L0: jmp_tab L4, L0
jmp_tab L1, L0
jmp_tab L2, L0
jmp_tab L3, L0
jmp_tab L5, L0
jmp_tab L8, L0
jmp_tab L12, L0
jmp_tab L15, L0
L1: mov rax, [rsi]
mul rax
mov [rdi], rax
mov [rdi+8], rdx
WIN64_GCC_EXIT
L2: mov rax, [rsi]
mov r8, rax
mul rax
mov r11, [rsi+8]
mov [rdi], rax
mov rax, r11
mov r9, rdx
mul rax
mov r10, rax
mov rax, r11
mov r11, rdx
mul r8
xor r8, r8
add r9, rax
adc r10, rdx
adc r11, r8
add r9, rax
mov [rdi+8], r9
adc r10, rdx
mov [rdi+16], r10
adc r11, r8
mov [rdi+24], r11
WIN64_GCC_EXIT
L3: mov rax, [rsi]
mov r10, rax
mul rax
mov r11, [rsi+8]
mov [rdi], rax
mov rax, r11
mov [rdi+8], rdx
mul rax
mov rcx, [rsi+16]
mov [rdi+16], rax
mov rax, rcx
mov [rdi+24], rdx
mul rax
mov [rdi+32], rax
mov [rdi+40], rdx
mov rax, r11
mul r10
mov r8, rax
mov rax, rcx
mov r9, rdx
mul r10
xor r10, r10
add r9, rax
mov rax, r11
mov r11, r10
adc r10, rdx
mul rcx
add r10, rax
adc rdx, r11
add r8, r8
adc r9, r9
adc r10, r10
adc rdx, rdx
adc r11, r11
add [rdi+8], r8
adc [rdi+16], r9
adc [rdi+24], r10
adc [rdi+32], rdx
adc [rdi+40], r11
WIN64_GCC_EXIT
L4: mov rax, [rsi]
mov r11, rax
mul rax
mov rbx, [rsi+8]
mov [rdi], rax
mov rax, rbx
mov [rdi+8], rdx
mul rax
mov [rdi+16], rax
mov [rdi+24], rdx
mov rax, [rsi+16]
mul rax
mov [rdi+32], rax
mov [rdi+40], rdx
mov rax, [rsi+24]
mul rax
mov [rdi+48], rax
mov rax, rbx
mov [rdi+56], rdx
mul r11
mov r8, rax
mov r9, rdx
mov rax, [rsi+16]
mul r11
xor r10, r10
add r9, rax
adc r10, rdx
mov rax, [rsi+24]
mul r11
xor r11, r11
add r10, rax
adc r11, rdx
mov rax, [rsi+16]
mul rbx
xor rcx, rcx
add r10, rax
adc r11, rdx
adc rcx, 0
mov rax, [rsi+24]
mul rbx
add r11, rax
adc rcx, rdx
mov rdx, [rsi+16]
mov rax, [rsi+24]
mul rdx
add rcx, rax
adc rdx, 0
add r8, r8
adc r9, r9
adc r10, r10
adc r11, r11
adc rcx, rcx
mov eax, 0
adc rdx, rdx
adc rax, rax
add [rdi+8], r8
adc [rdi+16], r9
adc [rdi+24], r10
adc [rdi+32], r11
adc [rdi+40], rcx
adc [rdi+48], rdx
adc [rdi+56], rax
WIN64_GCC_EXIT
L5: lea r12, [rdi+r11*8-16]
mov r13, [rsi]
mov rax, [rsi+8]
lea rsi, [rsi+r11*8]
lea r8, [r11-4]
xor r9d, r9d
sub r9, r11
mul r13
xor ebp, ebp
mov rbx, rax
mov rax, [rsi+r9*8+16]
mov r10, rdx
jmp .7
xalign 16
.6: add rbp, rax
mov [r12+r9*8], r10
mov rax, [rsi+r9*8]
adc rcx, rdx
xor ebx, ebx
mul r13
xor r10d, r10d
mov [r12+r9*8+8], rbp
add rcx, rax
adc rbx, rdx
mov rax, [rsi+r9*8+8]
mov [r12+r9*8+16], rcx
xor ebp, ebp
mul r13
add rbx, rax
mov rax, [rsi+r9*8+16]
adc r10, rdx
.7: xor ecx, ecx
mul r13
add r10, rax
mov rax, [rsi+r9*8+24]
adc rbp, rdx
mov [r12+r9*8+24], rbx
mul r13
add r9, 4
js .6
add rbp, rax
mov [r12], r10
adc rcx, rdx
mov [r12+8], rbp
mov [r12+16], rcx
lea r12, [r12+16]
lea rsi, [rsi-8]
jmp L18
L8: lea r12, [rdi+r11*8+8]
mov r13, [rsi]
mov rax, [rsi+8]
lea rsi, [rsi+r11*8+8]
lea r8, [r11-3]
lea r9, [r11-3]
neg r9
mov r14, rax
mul r13
mov rcx, rdx
xor ebp, ebp
mov [rdi+8], rax
jmp .10
xalign 16
.9: mul r14
add rbx, rax
adc rcx, rdx
mov rax, [rsi+r9*8-24]
mov ebp, 0
mul r13
add rbx, rax
mov rax, [rsi+r9*8-24]
adc rcx, rdx
adc ebp, 0
mul r14
add rcx, rax
mov [r12+r9*8-24], rbx
adc rbp, rdx
.10:mov rax, [rsi+r9*8-16]
mul r13
mov r10d, 0
add rcx, rax
adc rbp, rdx
mov rax, [rsi+r9*8-16]
adc r10d, 0
mov ebx, 0
mov [r12+r9*8-16], rcx
mul r14
add rbp, rax
mov rax, [rsi+r9*8-8]
adc r10, rdx
mov ecx, 0
mul r13
add rbp, rax
mov rax, [rsi+r9*8-8]
adc r10, rdx
adc ebx, 0
mul r14
add r10, rax
mov [r12+r9*8-8], rbp
adc rbx, rdx
.11:mov rax, [rsi+r9*8]
mul r13
add r10, rax
adc rbx, rdx
adc ecx, 0
add r9, 4
mov rax, [rsi+r9*8-32]
mov [r12+r9*8-32], r10
js .9
mul r14
add rbx, rax
adc rcx, rdx
mov [r12-8], rbx
mov [r12], rcx
lea rsi, [rsi-16]
jmp L24
L12:lea r12, [rdi+r11*8-16]
mov r13, [rsi]
mov rax, [rsi+8]
lea rsi, [rsi+r11*8]
lea r8, [r11-4]
lea r9, [r11-2]
neg r9
mul r13
mov rbp, rax
mov rax, [rsi+r9*8]
mov rcx, rdx
jmp .14
xalign 16
.13:add rbp, rax
mov [r12+r9*8], r10
mov rax, [rsi+r9*8]
adc rcx, rdx
.14:xor ebx, ebx
mul r13
xor r10d, r10d
mov [r12+r9*8+8], rbp
add rcx, rax
adc rbx, rdx
mov rax, [rsi+r9*8+8]
mov [r12+r9*8+16], rcx
xor ebp, ebp
mul r13
add rbx, rax
mov rax, [rsi+r9*8+16]
adc r10, rdx
xor ecx, ecx
mul r13
add r10, rax
mov rax, [rsi+r9*8+24]
adc rbp, rdx
mov [r12+r9*8+24], rbx
mul r13
add r9, 4
js .13
add rbp, rax
mov [r12], r10
adc rcx, rdx
mov [r12+8], rbp
mov [r12+16], rcx
lea r12, [r12+16]
lea rsi, [rsi-8]
jmp L21
L15:lea r12, [rdi+r11*8+8]
mov r13, [rsi]
mov rax, [rsi+8]
lea rsi, [rsi+r11*8+8]
lea r8, [r11-5]
lea r9, [r11-1]
neg r9
mov r14, rax
mul r13
mov r10, rdx
xor ebx, ebx
xor ecx, ecx
mov [rdi+8], rax
jmp .17
xalign 16
.16:mul r14
add rbx, rax
adc rcx, rdx
mov rax, [rsi+r9*8-24]
mov ebp, 0
mul r13
add rbx, rax
mov rax, [rsi+r9*8-24]
adc rcx, rdx
adc ebp, 0
mul r14
add rcx, rax
mov [r12+r9*8-24], rbx
adc rbp, rdx
mov rax, [rsi+r9*8-16]
mul r13
mov r10d, 0
add rcx, rax
adc rbp, rdx
mov rax, [rsi+r9*8-16]
adc r10d, 0
mov ebx, 0
mov [r12+r9*8-16], rcx
mul r14
add rbp, rax
mov rax, [rsi+r9*8-8]
adc r10, rdx
mov ecx, 0
mul r13
add rbp, rax
mov rax, [rsi+r9*8-8]
adc r10, rdx
adc ebx, 0
mul r14
add r10, rax
mov [r12+r9*8-8], rbp
adc rbx, rdx
.17:mov rax, [rsi+r9*8]
mul r13
add r10, rax
adc rbx, rdx
adc ecx, 0
add r9, 4
mov rax, [rsi+r9*8-32]
mov [r12+r9*8-32], r10
js .16
mul r14
add rbx, rax
adc rcx, rdx
mov [r12-8], rbx
mov [r12], rcx
lea rsi, [rsi-16]
jmp L21
L18:lea r9, [r8+4]
neg r9
mov r13, [rsi+r9*8+16]
mov r14, [rsi+r9*8+24]
mov rax, [rsi+r9*8+24]
mul r13
xor r10d, r10d
add [r12+r9*8+24], rax
adc r10, rdx
xor ebx, ebx
xor ecx, ecx
jmp .20
xalign 16
.19:add [r12+r9*8], r10
adc rbx, rax
mov rax, [rsi+r9*8+8]
adc rcx, rdx
mov ebp, 0
mul r13
add rbx, rax
mov rax, [rsi+r9*8+8]
adc rcx, rdx
adc ebp, 0
mul r14
add [r12+r9*8+8], rbx
adc rcx, rax
adc rbp, rdx
mov rax, [rsi+r9*8+16]
mov r10d, 0
mul r13
add rcx, rax
mov rax, [rsi+r9*8+16]
adc rbp, rdx
adc r10d, 0
mul r14
add [r12+r9*8+16], rcx
adc rbp, rax
mov rax, [rsi+r9*8+24]
adc r10, rdx
mul r13
mov ebx, 0
add rbp, rax
adc r10, rdx
mov ecx, 0
mov rax, [rsi+r9*8+24]
adc ebx, 0
mul r14
add [r12+r9*8+24], rbp
adc r10, rax
adc rbx, rdx
.20:mov rax, [rsi+r9*8+32]
mul r13
add r10, rax
mov rax, [rsi+r9*8+32]
adc rbx, rdx
adc ecx, 0
mul r14
add r9, 4
js .19
add [r12], r10
adc rbx, rax
adc rcx, rdx
mov [r12+8], rbx
mov [r12+16], rcx
lea r12, [r12+16]
add r8d, -2
L21:lea r9, [r8+2]
neg r9
mov r13, [rsi+r9*8]
mov r14, [rsi+r9*8+8]
mov rax, [rsi+r9*8+8]
mul r13
xor ecx, ecx
add [r12+r9*8+8], rax
adc rcx, rdx
xor ebp, ebp
jmp .23
xalign 16
.22:add [r12+r9*8], r10
adc rbx, rax
mov rax, [rsi+r9*8+8]
adc rcx, rdx
mov ebp, 0
mul r13
add rbx, rax
mov rax, [rsi+r9*8+8]
adc rcx, rdx
adc ebp, 0
mul r14
add [r12+r9*8+8], rbx
adc rcx, rax
adc rbp, rdx
.23:mov rax, [rsi+r9*8+16]
mov r10d, 0
mul r13
add rcx, rax
mov rax, [rsi+r9*8+16]
adc rbp, rdx
adc r10d, 0
mul r14
add [r12+r9*8+16], rcx
adc rbp, rax
mov rax, [rsi+r9*8+24]
adc r10, rdx
mul r13
mov ebx, 0
add rbp, rax
adc r10, rdx
mov ecx, 0
mov rax, [rsi+r9*8+24]
adc ebx, 0
mul r14
add [r12+r9*8+24], rbp
adc r10, rax
adc rbx, rdx
mov rax, [rsi+r9*8+32]
mul r13
add r10, rax
mov rax, [rsi+r9*8+32]
adc rbx, rdx
adc ecx, 0
mul r14
add r9, 4
js .22
add [r12], r10
adc rbx, rax
adc rcx, rdx
mov [r12+8], rbx
mov [r12+16], rcx
lea r12, [r12+16]
L24:add r8d, -2
jne L18
mov r13, [rsi-16]
mov r14, [rsi-8]
mov rax, [rsi-8]
mul r13
xor r10d, r10d
add [r12-8], rax
adc r10, rdx
xor ebx, ebx
xor ecx, ecx
mov rax, [rsi]
mul r13
add r10, rax
mov rax, [rsi]
adc rbx, rdx
mul r14
add [r12], r10
adc rbx, rax
adc rcx, rdx
mov [r12+8], rbx
mov [r12+16], rcx
lea r9, [r11+r11-4]
mov r11, [rdi+8]
lea rsi, [rsi-8]
lea rdi, [rdi+r9*8]
neg r9
mov rax, [rsi+r9*4]
mul rax
test r9b, 2
jnz .26
.25:add r11, r11
sbb ebx, ebx
add r11, rdx
mov [rdi+r9*8], rax
jmp .28
.26:add r11, r11
sbb ebp, ebp
add r11, rdx
mov [rdi+r9*8], rax
lea r9, [r9-2]
jmp .29
xalign 16
.27:mov rax, [rsi+r9*4]
mul rax
add ebp, ebp
adc r10, rax
adc r11, rdx
mov [rdi+r9*8], r10
.28:mov [rdi+r9*8+8], r11
mov r10, [rdi+r9*8+16]
adc r10, r10
mov r11, [rdi+r9*8+24]
adc r11, r11
nop
sbb ebp, ebp
mov rax, [rsi+r9*4+8]
mul rax
add ebx, ebx
adc r10, rax
adc r11, rdx
mov [rdi+r9*8+16], r10
.29:mov [rdi+r9*8+24], r11
mov r10, [rdi+r9*8+32]
adc r10, r10
mov r11, [rdi+r9*8+40]
adc r11, r11
sbb ebx, ebx
add r9, 4
js .27
mov rax, [rsi]
mul rax
add ebp, ebp
adc r10, rax
adc r11, rdx
mov [rdi], r10
mov [rdi+8], r11
mov r10, [rdi+16]
adc r10, r10
sbb ebp, ebp
neg ebp
mov rax, [rsi+8]
mul rax
add ebx, ebx
adc r10, rax
adc rdx, rbp
mov [rdi+16], r10
mov [rdi+24], rdx
WIN64_GCC_END
end