cheng/mpir
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Browse Source
This commit is contained in:
gladman 2009-07-28 16:29:28 +00:00
parent 558c10cba3
commit 40d6c79399
1 changed files with 65 additions and 65 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

130
mpn/x86_64w/core2/lshift.asm
View File

@ -20,9 +20,9 @@
; to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
; Boston, MA 02110-1301, USA.
;
; mpn_lshift(mp_ptr rdi, mp_ptr rsi, mp_size_t rdx, mp_limb_t rcx)
; mpn_lshift(mp_ptr rdi, mp_ptr rsi, mp_size_t rdx, mp_limb_t rcx)
; rcx rdx r8d r9
; rax=carry
; rax=carry
%include "..\yasm_mac.inc"
@ -33,78 +33,78 @@
FRAME_PROC mpn_lshift, 0, reg_save_list
; odd and even n seem to have different runtimes
movsxd rbx, r8d
lea rsi, [rdx+24]
lea rdi, [rcx+24]
mov rcx, r9
mov rdx, [rsi+rbx*8-32]
xor rax, rax
shld rax, rdx, cl
sub rbx, 5
js L_skiplp
xalign 16
movsxd rbx, r8d
lea rsi, [rdx+24]
lea rdi, [rcx+24]
mov rcx, r9
mov rdx, [rsi+rbx*8-32]
xor rax, rax
shld rax, rdx, cl
sub rbx, 5
js L_skiplp
xalign 16
L_lp:
mov r8, [rsi+rbx*8]
mov r11, [rsi+rbx*8-24]
mov r9, [rsi+rbx*8-8]
shld rdx, r8, cl
mov [rdi+rbx*8+8], rdx
mov rdx, r11
mov r10, [rsi+rbx*8-16]
shld r8, r9, cl
shld r9, r10, cl
mov [rdi+rbx*8], r8
mov [rdi+rbx*8-8], r9
shld r10, r11, cl
sub rbx, 4
mov [rdi+rbx*8+16], r10
jns L_lp
mov r8, [rsi+rbx*8]
mov r11, [rsi+rbx*8-24]
mov r9, [rsi+rbx*8-8]
shld rdx, r8, cl
mov [rdi+rbx*8+8], rdx
mov rdx, r11
mov r10, [rsi+rbx*8-16]
shld r8, r9, cl
shld r9, r10, cl
mov [rdi+rbx*8], r8
mov [rdi+rbx*8-8], r9
shld r10, r11, cl
sub rbx, 4
mov [rdi+rbx*8+16], r10
jns L_lp
L_skiplp:
cmp rbx, -2
ja L_case3
je L_case2
jp L_case1
cmp rbx, -2
ja L_case3
je L_case2
jp L_case1
; ALIGN(16)
L_case0:
shl rdx, cl
mov [rdi+rbx*8+8], rdx
jmp L_xit
shl rdx, cl
mov [rdi+rbx*8+8], rdx
jmp L_xit
xalign 16
xalign 16
L_case3:
mov r8, [rsi+rbx*8]
mov r9, [rsi+rbx*8-8]
shld rdx, r8, cl
mov [rdi+rbx*8+8], rdx
mov r10, [rsi+rbx*8-16]
shld r8, r9, cl
shld r9, r10, cl
mov [rdi+rbx*8], r8
mov [rdi+rbx*8-8], r9
shl r10, cl
mov [rdi+rbx*8-16], r10
jmp L_xit
mov r8, [rsi+rbx*8]
mov r9, [rsi+rbx*8-8]
shld rdx, r8, cl
mov [rdi+rbx*8+8], rdx
mov r10, [rsi+rbx*8-16]
shld r8, r9, cl
shld r9, r10, cl
mov [rdi+rbx*8], r8
mov [rdi+rbx*8-8], r9
shl r10, cl
mov [rdi+rbx*8-16], r10
jmp L_xit
xalign 16
xalign 16
L_case2:
mov r8, [rsi+rbx*8]
mov r9, [rsi+rbx*8-8]
shld rdx, r8, cl
mov [rdi+rbx*8+8], rdx
shld r8, r9, cl
shl r9, cl
mov [rdi+rbx*8], r8
mov [rdi+rbx*8-8], r9
jmp L_xit
xalign 16
mov r8, [rsi+rbx*8]
mov r9, [rsi+rbx*8-8]
shld rdx, r8, cl
mov [rdi+rbx*8+8], rdx
shld r8, r9, cl
shl r9, cl
mov [rdi+rbx*8], r8
mov [rdi+rbx*8-8], r9
jmp L_xit
xalign 16
L_case1:
mov r8, [rsi+rbx*8]
shld rdx, r8, cl
mov [rdi+rbx*8+8], rdx
shl r8, cl
mov [rdi+rbx*8], r8
mov r8, [rsi+rbx*8]
shld rdx, r8, cl
mov [rdi+rbx*8+8], rdx
shl r8, cl
mov [rdi+rbx*8], r8
L_xit:
END_PROC reg_save_list