[llvm] r288801 - [X86] Add tests to show missed opportunities to calculate knownbits in SMAX/SMIN/UMAX/UMIN

[Simon Pilgrim via llvm-commits]

Author: rksimon
Date: Tue Dec  6 06:12:20 2016
New Revision: 288801

URL: http://llvm.org/viewvc/llvm-project?rev=288801&view=rev
Log:
[X86] Add tests to show missed opportunities to calculate knownbits in SMAX/SMIN/UMAX/UMIN

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=288801&r1=288800&r2=288801&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/known-bits-vector.ll (original)
+++ llvm/trunk/test/CodeGen/X86/known-bits-vector.ll Tue Dec  6 06:12:20 2016
@@ -379,3 +379,97 @@ define <4 x float> @knownbits_lshr_bitca
   %4 = uitofp <4 x i32> %3 to <4 x float>
   ret <4 x float> %4
 }
+
+define <4 x float> @knownbits_smax_smin_shuffle_uitofp(<4 x i32> %a0) {
+; X32-LABEL: knownbits_smax_smin_shuffle_uitofp:
+; X32:       # BB#0:
+; X32-NEXT:    vpminsd {{\.LCPI.*}}, %xmm0, %xmm0
+; X32-NEXT:    vpmaxsd {{\.LCPI.*}}, %xmm0, %xmm0
+; X32-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,0,3,3]
+; 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:    retl
+;
+; X64-LABEL: knownbits_smax_smin_shuffle_uitofp:
+; X64:       # BB#0:
+; X64-NEXT:    vpminsd {{.*}}(%rip), %xmm0, %xmm0
+; X64-NEXT:    vpmaxsd {{.*}}(%rip), %xmm0, %xmm0
+; X64-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,0,3,3]
+; 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 = call <4 x i32> @llvm.x86.sse41.pminsd(<4 x i32> %a0, <4 x i32> <i32 0, i32 -65535, i32 -65535, i32 0>)
+  %2 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %1, <4 x i32> <i32 65535, i32 -1, i32 -1, i32 131071>)
+  %3 = shufflevector <4 x i32> %2, <4 x i32> undef, <4 x i32> <i32 0, i32 0, i32 3, i32 3>
+  %4 = uitofp <4 x i32> %3 to <4 x float>
+  ret <4 x float> %4
+}
+declare <4 x i32> @llvm.x86.sse41.pminsd(<4 x i32>, <4 x i32>) nounwind readnone
+declare <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32>, <4 x i32>) nounwind readnone
+
+define <4 x float> @knownbits_umax_shuffle_uitofp(<4 x i32> %a0) {
+; X32-LABEL: knownbits_umax_shuffle_uitofp:
+; X32:       # BB#0:
+; X32-NEXT:    vpmaxud {{\.LCPI.*}}, %xmm0, %xmm0
+; X32-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,0,3,3]
+; 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:    retl
+;
+; X64-LABEL: knownbits_umax_shuffle_uitofp:
+; X64:       # BB#0:
+; X64-NEXT:    vpmaxud {{.*}}(%rip), %xmm0, %xmm0
+; X64-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,0,3,3]
+; 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 = call <4 x i32> @llvm.x86.sse41.pmaxud(<4 x i32> %a0, <4 x i32> <i32 65535, i32 -1, i32 -1, i32 262143>)
+  %2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 0, i32 0, i32 3, i32 3>
+  %3 = uitofp <4 x i32> %2 to <4 x float>
+  ret <4 x float> %3
+}
+declare <4 x i32> @llvm.x86.sse41.pmaxud(<4 x i32>, <4 x i32>) nounwind readnone
+
+define <4 x float> @knownbits_umin_shl_shuffle_uitofp(<4 x i32> %a0) {
+; X32-LABEL: knownbits_umin_shl_shuffle_uitofp:
+; X32:       # BB#0:
+; X32-NEXT:    vpminud {{\.LCPI.*}}, %xmm0, %xmm0
+; X32-NEXT:    vpslld $16, %xmm0, %xmm0
+; X32-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,0,3,3]
+; 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:    retl
+;
+; X64-LABEL: knownbits_umin_shl_shuffle_uitofp:
+; X64:       # BB#0:
+; X64-NEXT:    vpminud {{.*}}(%rip), %xmm0, %xmm0
+; X64-NEXT:    vpslld $16, %xmm0, %xmm0
+; X64-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,0,3,3]
+; 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 = call <4 x i32> @llvm.x86.sse41.pminud(<4 x i32> %a0, <4 x i32> <i32 65536, i32 -1, i32 -1, i32 262143>)
+  %2 = shl <4 x i32> %1, <i32 16, i32 16, i32 16, i32 16>
+  %3 = shufflevector <4 x i32> %2, <4 x i32> undef, <4 x i32> <i32 0, i32 0, i32 3, i32 3>
+  %4 = uitofp <4 x i32> %3 to <4 x float>
+  ret <4 x float> %4
+}
+declare <4 x i32> @llvm.x86.sse41.pminud(<4 x i32>, <4 x i32>) nounwind readnone