RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-02-01 13:11:39 +01:00
Code Issues Projects Releases Wiki Activity

[X86][SSE] Regenerate vsplit and tests

Browse Source 
To make it more obvious how bad some of that truncation code is....

llvm-svn: 283880
This commit is contained in:
Simon Pilgrim 2016-10-11 13:51:44 +00:00
parent edf39dd95a
commit e5236e5c4f
1 changed files with 53 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

59
test/CodeGen/X86/vsplit-and.ll
View File

@ -1,9 +1,17 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-linux -mcpu=penryn | FileCheck %s
define void @t0(<2 x i64>* %dst, <2 x i64> %src1, <2 x i64> %src2) nounwind readonly {
; CHECK: t0
; CHECK: pand
; CHECK: ret
; CHECK-LABEL: t0:
; CHECK: # BB#0:
; CHECK-NEXT: pxor %xmm2, %xmm2
; CHECK-NEXT: pcmpeqq %xmm2, %xmm0
; CHECK-NEXT: pcmpeqd %xmm3, %xmm3
; CHECK-NEXT: pxor %xmm0, %xmm3
; CHECK-NEXT: pcmpeqq %xmm2, %xmm1
; CHECK-NEXT: pandn %xmm3, %xmm1
; CHECK-NEXT: movdqa %xmm1, (%rdi)
; CHECK-NEXT: retq
%cmp1 = icmp ne <2 x i64> %src1, zeroinitializer
%cmp2 = icmp ne <2 x i64> %src2, zeroinitializer
%t1 = and <2 x i1> %cmp1, %cmp2
@ -13,9 +21,48 @@ define void @t0(<2 x i64>* %dst, <2 x i64> %src1, <2 x i64> %src2) nounwind read
}
define void @t2(<3 x i64>* %dst, <3 x i64> %src1, <3 x i64> %src2) nounwind readonly {
; CHECK: t2
; CHECK: pand
; CHECK: ret
; CHECK-LABEL: t2:
; CHECK: # BB#0:
; CHECK-NEXT: movd %r9, %xmm1
; CHECK-NEXT: movd %r8, %xmm0
; CHECK-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; CHECK-NEXT: movd %rdx, %xmm1
; CHECK-NEXT: movd %rsi, %xmm2
; CHECK-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm1[0]
; CHECK-NEXT: movd %rcx, %xmm1
; CHECK-NEXT: movq {{.*#+}} xmm3 = mem[0],zero
; CHECK-NEXT: pxor %xmm4, %xmm4
; CHECK-NEXT: pcmpeqq %xmm4, %xmm1
; CHECK-NEXT: pcmpeqd %xmm5, %xmm5
; CHECK-NEXT: pxor %xmm5, %xmm1
; CHECK-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,1,0,2]
; CHECK-NEXT: pcmpeqq %xmm4, %xmm2
; CHECK-NEXT: pxor %xmm5, %xmm2
; CHECK-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
; CHECK-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3],xmm1[4,5,6,7]
; CHECK-NEXT: pslld $31, %xmm2
; CHECK-NEXT: psrad $31, %xmm2
; CHECK-NEXT: pcmpeqq %xmm4, %xmm3
; CHECK-NEXT: pxor %xmm5, %xmm3
; CHECK-NEXT: pshufd {{.*#+}} xmm1 = xmm3[0,1,0,2]
; CHECK-NEXT: pcmpeqq %xmm4, %xmm0
; CHECK-NEXT: pxor %xmm5, %xmm0
; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; CHECK-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
; CHECK-NEXT: pslld $31, %xmm0
; CHECK-NEXT: psrad $31, %xmm0
; CHECK-NEXT: pand %xmm2, %xmm0
; CHECK-NEXT: pmovzxdq {{.*#+}} xmm1 = xmm0[0],zero,xmm0[1],zero
; CHECK-NEXT: psllq $63, %xmm1
; CHECK-NEXT: psrad $31, %xmm1
; CHECK-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,2,3,3]
; CHECK-NEXT: psllq $63, %xmm0
; CHECK-NEXT: psrad $31, %xmm0
; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
; CHECK-NEXT: movq %xmm0, 16(%rdi)
; CHECK-NEXT: movdqa %xmm1, (%rdi)
; CHECK-NEXT: retq
%cmp1 = icmp ne <3 x i64> %src1, zeroinitializer
%cmp2 = icmp ne <3 x i64> %src2, zeroinitializer
%t1 = and <3 x i1> %cmp1, %cmp2