[llvm] r326372 - [X86][AVX512] Improve support for signed saturation truncation stores

[Simon Pilgrim via llvm-commits]

Author: rksimon
Date: Wed Feb 28 13:42:19 2018
New Revision: 326372

URL: http://llvm.org/viewvc/llvm-project?rev=326372&view=rev
Log:
[X86][AVX512] Improve support for signed saturation truncation stores

Matches what we already manage for unsigned saturation truncation stores

Differential Revision: https://reviews.llvm.org/D43629

Modified:
    llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
    llvm/trunk/test/CodeGen/X86/vector-trunc-ssat.ll

Modified: llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86ISelLowering.cpp?rev=326372&r1=326371&r2=326372&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86ISelLowering.cpp (original)
+++ llvm/trunk/lib/Target/X86/X86ISelLowering.cpp Wed Feb 28 13:42:19 2018
@@ -34349,6 +34349,20 @@ static SDValue detectSSatPattern(SDValue
   return SDValue();
 }
 
+/// Detect a pattern of truncation with signed saturation.
+/// The types should allow to use VPMOVSS* instruction on AVX512.
+/// Return the source value to be truncated or SDValue() if the pattern was not
+/// matched.
+static SDValue detectAVX512SSatPattern(SDValue In, EVT VT,
+                                       const X86Subtarget &Subtarget,
+                                       const TargetLowering &TLI) {
+  if (!TLI.isTypeLegal(In.getValueType()))
+    return SDValue();
+  if (!isSATValidOnAVX512Subtarget(In.getValueType(), VT, Subtarget))
+    return SDValue();
+  return detectSSatPattern(In, VT);
+}
+
 /// Detect a pattern of truncation with saturation:
 /// (truncate (umin (x, unsigned_max_of_dest_type)) to dest_type).
 /// The types should allow to use VPMOVUS* instruction on AVX512.
@@ -34988,6 +35002,12 @@ static SDValue combineStore(SDNode *N, S
 
     const TargetLowering &TLI = DAG.getTargetLoweringInfo();
     if (SDValue Val =
+        detectAVX512SSatPattern(St->getValue(), St->getMemoryVT(), Subtarget,
+                                TLI))
+      return EmitTruncSStore(true /* Signed saturation */, St->getChain(),
+                             dl, Val, St->getBasePtr(),
+                             St->getMemoryVT(), St->getMemOperand(), DAG);
+    if (SDValue Val =
         detectAVX512USatPattern(St->getValue(), St->getMemoryVT(), Subtarget,
                                 TLI))
       return EmitTruncSStore(false /* Unsigned saturation */, St->getChain(),

Modified: llvm/trunk/test/CodeGen/X86/vector-trunc-ssat.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/vector-trunc-ssat.ll?rev=326372&r1=326371&r2=326372&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/vector-trunc-ssat.ll (original)
+++ llvm/trunk/test/CodeGen/X86/vector-trunc-ssat.ll Wed Feb 28 13:42:19 2018
@@ -2042,9 +2042,7 @@ define void @trunc_ssat_v8i64_v8i8_store
 ;
 ; AVX512-LABEL: trunc_ssat_v8i64_v8i8_store:
 ; AVX512:       # %bb.0:
-; AVX512-NEXT:    vpminsq {{.*}}(%rip){1to8}, %zmm0, %zmm0
-; AVX512-NEXT:    vpmaxsq {{.*}}(%rip){1to8}, %zmm0, %zmm0
-; AVX512-NEXT:    vpmovqb %zmm0, (%rdi)
+; AVX512-NEXT:    vpmovsqb %zmm0, (%rdi)
 ; AVX512-NEXT:    vzeroupper
 ; AVX512-NEXT:    retq
   %1 = icmp slt <8 x i64> %a0, <i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127>
@@ -2998,9 +2996,7 @@ define void @trunc_ssat_v8i32_v8i8_store
 ;
 ; AVX512VL-LABEL: trunc_ssat_v8i32_v8i8_store:
 ; AVX512VL:       # %bb.0:
-; AVX512VL-NEXT:    vpminsd {{.*}}(%rip){1to8}, %ymm0, %ymm0
-; AVX512VL-NEXT:    vpmaxsd {{.*}}(%rip){1to8}, %ymm0, %ymm0
-; AVX512VL-NEXT:    vpmovdb %ymm0, (%rdi)
+; AVX512VL-NEXT:    vpmovsdb %ymm0, (%rdi)
 ; AVX512VL-NEXT:    vzeroupper
 ; AVX512VL-NEXT:    retq
 ;
@@ -3018,9 +3014,7 @@ define void @trunc_ssat_v8i32_v8i8_store
 ;
 ; AVX512BWVL-LABEL: trunc_ssat_v8i32_v8i8_store:
 ; AVX512BWVL:       # %bb.0:
-; AVX512BWVL-NEXT:    vpminsd {{.*}}(%rip){1to8}, %ymm0, %ymm0
-; AVX512BWVL-NEXT:    vpmaxsd {{.*}}(%rip){1to8}, %ymm0, %ymm0
-; AVX512BWVL-NEXT:    vpmovdb %ymm0, (%rdi)
+; AVX512BWVL-NEXT:    vpmovsdb %ymm0, (%rdi)
 ; AVX512BWVL-NEXT:    vzeroupper
 ; AVX512BWVL-NEXT:    retq
   %1 = icmp slt <8 x i32> %a0, <i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127>