diff options
Diffstat (limited to 'llvm/lib/Transforms/InstCombine')
-rw-r--r-- | llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp | 12 | ||||
-rw-r--r-- | llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp | 10 |
2 files changed, 21 insertions, 1 deletions
diff --git a/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp b/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp index 2ef82ba3ed8..ec2ebaaed88 100644 --- a/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp +++ b/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp @@ -2315,10 +2315,20 @@ Instruction *InstCombiner::visitCallInst(CallInst &CI) { case Intrinsic::x86_ssse3_pshuf_b_128: case Intrinsic::x86_avx2_pshuf_b: - case Intrinsic::x86_avx512_pshuf_b_512: + case Intrinsic::x86_avx512_pshuf_b_512: { if (Value *V = simplifyX86pshufb(*II, *Builder)) return replaceInstUsesWith(*II, V); + + unsigned VWidth = II->getType()->getVectorNumElements(); + APInt UndefElts(VWidth, 0); + APInt DemandedElts = APInt::getAllOnesValue(VWidth); + if (Value *V = SimplifyDemandedVectorElts(II, DemandedElts, UndefElts)) { + if (V != II) + return replaceInstUsesWith(*II, V); + return II; + } break; + } case Intrinsic::x86_avx_vpermilvar_ps: case Intrinsic::x86_avx_vpermilvar_ps_256: diff --git a/llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp b/llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp index 8b930bd95df..95100d074b4 100644 --- a/llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp +++ b/llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp @@ -1472,6 +1472,16 @@ Value *InstCombiner::SimplifyDemandedVectorElts(Value *V, APInt DemandedElts, break; } + case Intrinsic::x86_ssse3_pshuf_b_128: + case Intrinsic::x86_avx2_pshuf_b: + case Intrinsic::x86_avx512_pshuf_b_512: { + Value *Op1 = II->getArgOperand(1); + TmpV = SimplifyDemandedVectorElts(Op1, DemandedElts, UndefElts, + Depth + 1); + if (TmpV) { II->setArgOperand(1, TmpV); MadeChange = true; } + break; + } + // SSE4A instructions leave the upper 64-bits of the 128-bit result // in an undefined state. case Intrinsic::x86_sse4a_extrq: |