[llvm] r316940 - [X86][SSE] computeKnownBits tests showing missing VSELECT demandedelts support

[Simon Pilgrim via llvm-commits]

Author: rksimon
Date: Mon Oct 30 11:48:31 2017
New Revision: 316940

URL: http://llvm.org/viewvc/llvm-project?rev=316940&view=rev
Log:
[X86][SSE] computeKnownBits tests showing missing VSELECT demandedelts support 

Modified:
    llvm/trunk/test/CodeGen/X86/known-bits-vector.ll

Modified: llvm/trunk/test/CodeGen/X86/known-bits-vector.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/known-bits-vector.ll?rev=316940&r1=316939&r2=316940&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/known-bits-vector.ll (original)
+++ llvm/trunk/test/CodeGen/X86/known-bits-vector.ll Mon Oct 30 11:48:31 2017
@@ -604,3 +604,51 @@ define <4 x float> @knownbits_or_abs_uit
   %6 = uitofp <4 x i32> %5 to <4 x float>
   ret <4 x float> %6
 }
+
+define <4 x float> @knownbits_lshr_and_select_shuffle_uitofp(<4 x i32> %a0, <4 x i32> %a1, <4 x i32> %a2, <4 x i32> %a3) nounwind {
+; X32-LABEL: knownbits_lshr_and_select_shuffle_uitofp:
+; X32:       # BB#0:
+; X32-NEXT:    pushl %ebp
+; X32-NEXT:    movl %esp, %ebp
+; X32-NEXT:    andl $-16, %esp
+; X32-NEXT:    subl $16, %esp
+; X32-NEXT:    vmovaps 8(%ebp), %xmm3
+; X32-NEXT:    vpsrld $1, %xmm2, %xmm4
+; X32-NEXT:    vpsrld $5, %xmm2, %xmm2
+; X32-NEXT:    vpblendw {{.*#+}} xmm2 = xmm2[0,1],xmm4[2,3],xmm2[4,5],xmm4[6,7]
+; X32-NEXT:    vandps {{\.LCPI.*}}, %xmm3, %xmm3
+; X32-NEXT:    vpcmpeqd %xmm1, %xmm0, %xmm0
+; X32-NEXT:    vblendvps %xmm0, %xmm2, %xmm3, %xmm0
+; X32-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,0,2,2]
+; X32-NEXT:    vpblendw {{.*#+}} xmm1 = xmm0[0],mem[1],xmm0[2],mem[3],xmm0[4],mem[5],xmm0[6],mem[7]
+; X32-NEXT:    vpsrld $16, %xmm0, %xmm0
+; X32-NEXT:    vpblendw {{.*#+}} xmm0 = xmm0[0],mem[1],xmm0[2],mem[3],xmm0[4],mem[5],xmm0[6],mem[7]
+; X32-NEXT:    vaddps {{\.LCPI.*}}, %xmm0, %xmm0
+; X32-NEXT:    vaddps %xmm0, %xmm1, %xmm0
+; X32-NEXT:    movl %ebp, %esp
+; X32-NEXT:    popl %ebp
+; X32-NEXT:    retl
+;
+; X64-LABEL: knownbits_lshr_and_select_shuffle_uitofp:
+; X64:       # BB#0:
+; X64-NEXT:    vpsrld $1, %xmm2, %xmm4
+; X64-NEXT:    vpsrld $5, %xmm2, %xmm2
+; X64-NEXT:    vpblendw {{.*#+}} xmm2 = xmm2[0,1],xmm4[2,3],xmm2[4,5],xmm4[6,7]
+; X64-NEXT:    vandps {{.*}}(%rip), %xmm3, %xmm3
+; X64-NEXT:    vpcmpeqd %xmm1, %xmm0, %xmm0
+; X64-NEXT:    vblendvps %xmm0, %xmm2, %xmm3, %xmm0
+; X64-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,0,2,2]
+; X64-NEXT:    vpblendw {{.*#+}} xmm1 = xmm0[0],mem[1],xmm0[2],mem[3],xmm0[4],mem[5],xmm0[6],mem[7]
+; X64-NEXT:    vpsrld $16, %xmm0, %xmm0
+; X64-NEXT:    vpblendw {{.*#+}} xmm0 = xmm0[0],mem[1],xmm0[2],mem[3],xmm0[4],mem[5],xmm0[6],mem[7]
+; X64-NEXT:    vaddps {{.*}}(%rip), %xmm0, %xmm0
+; X64-NEXT:    vaddps %xmm0, %xmm1, %xmm0
+; X64-NEXT:    retq
+  %1 = lshr <4 x i32> %a2, <i32 5, i32 1, i32 5, i32 1>
+  %2 = and <4 x i32> %a3, <i32 255, i32 -1, i32 65535, i32 -1>
+  %3 = icmp eq <4 x i32> %a0, %a1
+  %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> %2
+  %5 = shufflevector <4 x i32> %4, <4 x i32> undef, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
+  %6 = uitofp <4 x i32> %5 to <4 x float>
+  ret <4 x float> %6
+}