6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2025-05-22 03:39:03 +00:00
Code Issues Projects Releases Wiki Activity

[SSE] Fix for MUL v16i8 on pre-SSE41 targets (PR23369).

Browse Source 
Sign extension of i8 to i16 was placing the unpacked bytes in the lower byte instead of the upper byte.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@236209 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Simon Pilgrim 2015-04-30 08:23:16 +00:00
parent 21092d8a7c
commit c8ee30be4f
2 changed files with 16 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
lib/Target/X86/X86ISelLowering.cpp
View File

@ -16044,8 +16044,8 @@ static SDValue LowerMUL(SDValue Op, const X86Subtarget *Subtarget,
ALo = DAG.getNode(X86ISD::VSEXT, dl, ExVT, A);
BLo = DAG.getNode(X86ISD::VSEXT, dl, ExVT, B);
} else {
const int ShufMask[] = {0, -1, 1, -1, 2, -1, 3, -1,
4, -1, 5, -1, 6, -1, 7, -1};
const int ShufMask[] = {-1, 0, -1, 1, -1, 2, -1, 3,
-1, 4, -1, 5, -1, 6, -1, 7};
ALo = DAG.getVectorShuffle(VT, dl, A, A, ShufMask);
BLo = DAG.getVectorShuffle(VT, dl, B, B, ShufMask);
ALo = DAG.getNode(ISD::BITCAST, dl, ExVT, ALo);
@ -16064,8 +16064,8 @@ static SDValue LowerMUL(SDValue Op, const X86Subtarget *Subtarget,
AHi = DAG.getNode(X86ISD::VSEXT, dl, ExVT, AHi);
BHi = DAG.getNode(X86ISD::VSEXT, dl, ExVT, BHi);
} else {
const int ShufMask[] = {8, -1, 9, -1, 10, -1, 11, -1,
12, -1, 13, -1, 14, -1, 15, -1};
const int ShufMask[] = {-1, 8, -1, 9, -1, 10, -1, 11,
-1, 12, -1, 13, -1, 14, -1, 15};
AHi = DAG.getVectorShuffle(VT, dl, A, A, ShufMask);
BHi = DAG.getVectorShuffle(VT, dl, B, B, ShufMask);
AHi = DAG.getNode(ISD::BITCAST, dl, ExVT, AHi);

25
test/CodeGen/X86/pmul.ll
View File

@ -92,23 +92,22 @@ entry:
define <16 x i8> @mul8(<16 x i8> %i, <16 x i8> %j) nounwind {
; SSE2-LABEL: mul8:
; SSE2: # BB#0: # %entry
; SSE2-NEXT: movdqa %xmm1, %xmm3
; SSE2-NEXT: punpcklbw {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1],xmm3[2],xmm0[2],xmm3[3],xmm0[3],xmm3[4],xmm0[4],xmm3[5],xmm0[5],xmm3[6],xmm0[6],xmm3[7],xmm0[7]
; SSE2-NEXT: psraw $8, %xmm3
; SSE2-NEXT: movdqa %xmm0, %xmm2
; SSE2-NEXT: punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3],xmm2[4],xmm0[4],xmm2[5],xmm0[5],xmm2[6],xmm0[6],xmm2[7],xmm0[7]
; SSE2-NEXT: movdqa %xmm1, %xmm2
; SSE2-NEXT: punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT: psraw $8, %xmm2
; SSE2-NEXT: pmullw %xmm3, %xmm2
; SSE2-NEXT: movdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255]
; SSE2-NEXT: pand %xmm3, %xmm2
; SSE2-NEXT: punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT: movdqa %xmm0, %xmm3
; SSE2-NEXT: punpckhbw {{.*#+}} xmm3 = xmm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT: psraw $8, %xmm3
; SSE2-NEXT: pmullw %xmm2, %xmm3
; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
; SSE2-NEXT: pand %xmm2, %xmm3
; SSE2-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT: psraw $8, %xmm1
; SSE2-NEXT: punpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT: psraw $8, %xmm0
; SSE2-NEXT: pmullw %xmm1, %xmm0
; SSE2-NEXT: pand %xmm3, %xmm0
; SSE2-NEXT: packuswb %xmm0, %xmm2
; SSE2-NEXT: movdqa %xmm2, %xmm0
; SSE2-NEXT: pand %xmm2, %xmm0
; SSE2-NEXT: packuswb %xmm3, %xmm0
; SSE2-NEXT: retq
;
; SSE41-LABEL: mul8: