[llvm] r321951 - [X86] Add EVEX vcvtph2ps to the load folding tables.

[Craig Topper via llvm-commits]

Author: ctopper
Date: Sat Jan  6 22:24:24 2018
New Revision: 321951

URL: http://llvm.org/viewvc/llvm-project?rev=321951&view=rev
Log:
[X86] Add EVEX vcvtph2ps to the load folding tables.

Modified:
    llvm/trunk/lib/Target/X86/X86InstrInfo.cpp
    llvm/trunk/test/CodeGen/X86/stack-folding-fp-avx512.ll

Modified: llvm/trunk/lib/Target/X86/X86InstrInfo.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrInfo.cpp?rev=321951&r1=321950&r2=321951&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86InstrInfo.cpp (original)
+++ llvm/trunk/lib/Target/X86/X86InstrInfo.cpp Sat Jan  6 22:24:24 2018
@@ -1099,8 +1099,11 @@ X86InstrInfo::X86InstrInfo(X86Subtarget
     { X86::VPSRLWZ128ri,         X86::VPSRLWZ128mi,         0 },
 
     // F16C foldable instructions
-    { X86::VCVTPH2PSrr,        X86::VCVTPH2PSrm,            0 },
+    { X86::VCVTPH2PSrr,        X86::VCVTPH2PSrm,            TB_NO_REVERSE },
     { X86::VCVTPH2PSYrr,       X86::VCVTPH2PSYrm,           0 },
+    { X86::VCVTPH2PSZ128rr,    X86::VCVTPH2PSZ128rm,        TB_NO_REVERSE },
+    { X86::VCVTPH2PSZ256rr,    X86::VCVTPH2PSZ256rm,        0 },
+    { X86::VCVTPH2PSZrr,       X86::VCVTPH2PSZrm,           0 },
 
     // AES foldable instructions
     { X86::AESIMCrr,              X86::AESIMCrm,              TB_ALIGN_16 },

Modified: llvm/trunk/test/CodeGen/X86/stack-folding-fp-avx512.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/stack-folding-fp-avx512.ll?rev=321951&r1=321950&r2=321951&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/stack-folding-fp-avx512.ll (original)
+++ llvm/trunk/test/CodeGen/X86/stack-folding-fp-avx512.ll Sat Jan  6 22:24:24 2018
@@ -208,6 +208,15 @@ define <8 x float> @stack_fold_cvtpd2ps(
   ret <8 x float> %2
 }
 
+define <16 x float> @stack_fold_cvtph2ps(<16 x i16> %a0) {
+  ;CHECK-LABEL: stack_fold_cvtph2ps
+  ;CHECK:   vcvtph2ps {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <16 x float> @llvm.x86.avx512.mask.vcvtph2ps.512(<16 x i16> %a0, <16 x float> undef, i16 -1, i32 4)
+  ret <16 x float> %2
+}
+declare <16 x float> @llvm.x86.avx512.mask.vcvtph2ps.512(<16 x i16>, <16 x float>, i16, i32) nounwind readonly
+
 define <16 x i16> @stack_fold_cvtps2ph(<16 x float> %a0) {
   ;CHECK-LABEL: stack_fold_cvtps2ph
   ;CHECK:   vcvtps2ph $0, {{%zmm[0-9][0-9]*}}, {{-?[0-9]*}}(%rsp) {{.*#+}} 32-byte Folded Spill