[llvm] r369200 - [X86] Add a one use check to the combineStore code that handles v16i16->v16i8 truncate+store by extending to v16i32 and then emitting a v16i32->v16i8 truncstore.

[Craig Topper via llvm-commits]

Author: ctopper
Date: Sat Aug 17 15:46:15 2019
New Revision: 369200

URL: http://llvm.org/viewvc/llvm-project?rev=369200&view=rev
Log:
[X86] Add a one use check to the combineStore code that handles v16i16->v16i8 truncate+store by extending to v16i32 and then emitting a v16i32->v16i8 truncstore.

This prevent us from emitting a separate truncate and a truncating
store instruction.

Modified:
    llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
    llvm/trunk/test/CodeGen/X86/avx512-trunc.ll

Modified: llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86ISelLowering.cpp?rev=369200&r1=369199&r2=369200&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86ISelLowering.cpp (original)
+++ llvm/trunk/lib/Target/X86/X86ISelLowering.cpp Sat Aug 17 15:46:15 2019
@@ -39581,7 +39581,7 @@ static SDValue combineStore(SDNode *N, S
       St->getValue().getOpcode() == ISD::TRUNCATE &&
       St->getValue().getOperand(0).getValueType() == MVT::v16i16 &&
       TLI.isTruncStoreLegal(MVT::v16i32, MVT::v16i8) &&
-      !DCI.isBeforeLegalizeOps()) {
+      St->getValue().hasOneUse() && !DCI.isBeforeLegalizeOps()) {
     SDValue Ext = DAG.getNode(ISD::ANY_EXTEND, dl, MVT::v16i32, St->getValue());
     return DAG.getTruncStore(St->getChain(), dl, Ext, St->getBasePtr(),
                              MVT::v16i8, St->getMemOperand());

Modified: llvm/trunk/test/CodeGen/X86/avx512-trunc.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx512-trunc.ll?rev=369200&r1=369199&r2=369200&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/avx512-trunc.ll (original)
+++ llvm/trunk/test/CodeGen/X86/avx512-trunc.ll Sat Aug 17 15:46:15 2019
@@ -514,6 +514,26 @@ define void @trunc_wb_256_mem(<16 x i16>
     ret void
 }
 
+define <16 x i8> @trunc_wb_256_mem_and_ret(<16 x i16> %i, <16 x i8>* %res) #0 {
+; KNL-LABEL: trunc_wb_256_mem_and_ret:
+; KNL:       ## %bb.0:
+; KNL-NEXT:    vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
+; KNL-NEXT:    vpmovdb %zmm0, %xmm0
+; KNL-NEXT:    vmovdqa %xmm0, (%rdi)
+; KNL-NEXT:    vzeroupper
+; KNL-NEXT:    retq
+;
+; SKX-LABEL: trunc_wb_256_mem_and_ret:
+; SKX:       ## %bb.0:
+; SKX-NEXT:    vpmovwb %ymm0, %xmm0
+; SKX-NEXT:    vmovdqa %xmm0, (%rdi)
+; SKX-NEXT:    vzeroupper
+; SKX-NEXT:    retq
+    %x = trunc <16 x i16> %i to <16 x i8>
+    store <16 x i8> %x, <16 x i8>* %res
+    ret <16 x i8> %x
+}
+
 define <8 x i8> @trunc_wb_128(<8 x i16> %i) #0 {
 ; ALL-LABEL: trunc_wb_128:
 ; ALL:       ## %bb.0: