From 7ca219ae2d01c862ceac32522bd4e4e792a76be0 Mon Sep 17 00:00:00 2001 From: Luke Lau Date: Fri, 22 Nov 2024 18:10:35 +0800 Subject: [PATCH 1/5] Precommit tests --- .../Analysis/CostModel/RISCV/vp-intrinsics.ll | 106 ++++++++++++++++++ 1 file changed, 106 insertions(+) diff --git a/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll b/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll index 800ea223850d3..2b7f699e8b690 100644 --- a/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll +++ b/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll @@ -2125,6 +2125,112 @@ define void @vp_fdiv(){ ret void } +define void @splat() { +; CHECK-LABEL: 'splat' +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %1 = call <2 x i8> @llvm.experimental.vp.splat.v2i8(i8 undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %2 = call <4 x i8> @llvm.experimental.vp.splat.v4i8(i8 undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %3 = call <8 x i8> @llvm.experimental.vp.splat.v8i8(i8 undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %4 = call <16 x i8> @llvm.experimental.vp.splat.v16i8(i8 undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %5 = call <2 x i16> @llvm.experimental.vp.splat.v2i16(i16 undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %6 = call <4 x i16> @llvm.experimental.vp.splat.v4i16(i16 undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %7 = call <8 x i16> @llvm.experimental.vp.splat.v8i16(i16 undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %8 = call <16 x i16> @llvm.experimental.vp.splat.v16i16(i16 undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %9 = call <2 x i32> @llvm.experimental.vp.splat.v2i32(i32 undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %10 = call <4 x i32> @llvm.experimental.vp.splat.v4i32(i32 undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %11 = call <8 x i32> @llvm.experimental.vp.splat.v8i32(i32 undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %12 = call <16 x i32> @llvm.experimental.vp.splat.v16i32(i32 undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %13 = call <2 x i64> @llvm.experimental.vp.splat.v2i64(i64 undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %14 = call <4 x i64> @llvm.experimental.vp.splat.v4i64(i64 undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %15 = call <8 x i64> @llvm.experimental.vp.splat.v8i64(i64 undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %16 = call <16 x i64> @llvm.experimental.vp.splat.v16i64(i64 undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %17 = call @llvm.experimental.vp.splat.nxv2i8(i8 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %18 = call @llvm.experimental.vp.splat.nxv4i8(i8 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %19 = call @llvm.experimental.vp.splat.nxv8i8(i8 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %20 = call @llvm.experimental.vp.splat.nxv16i8(i8 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %21 = call @llvm.experimental.vp.splat.nxv2i16(i16 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %22 = call @llvm.experimental.vp.splat.nxv4i16(i16 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %23 = call @llvm.experimental.vp.splat.nxv8i16(i16 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %24 = call @llvm.experimental.vp.splat.nxv16i16(i16 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %25 = call @llvm.experimental.vp.splat.nxv2i32(i32 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %26 = call @llvm.experimental.vp.splat.nxv4i32(i32 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %27 = call @llvm.experimental.vp.splat.nxv8i32(i32 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %28 = call @llvm.experimental.vp.splat.nxv16i32(i32 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %29 = call @llvm.experimental.vp.splat.nxv2i64(i64 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %30 = call @llvm.experimental.vp.splat.nxv4i64(i64 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %31 = call @llvm.experimental.vp.splat.nxv8i64(i64 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %32 = call @llvm.experimental.vp.splat.nxv16i64(i64 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void +; +; TYPEBASED-LABEL: 'splat' +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %1 = call <2 x i8> @llvm.experimental.vp.splat.v2i8(i8 undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %2 = call <4 x i8> @llvm.experimental.vp.splat.v4i8(i8 undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %3 = call <8 x i8> @llvm.experimental.vp.splat.v8i8(i8 undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 95 for instruction: %4 = call <16 x i8> @llvm.experimental.vp.splat.v16i8(i8 undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %5 = call <2 x i16> @llvm.experimental.vp.splat.v2i16(i16 undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %6 = call <4 x i16> @llvm.experimental.vp.splat.v4i16(i16 undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %7 = call <8 x i16> @llvm.experimental.vp.splat.v8i16(i16 undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 110 for instruction: %8 = call <16 x i16> @llvm.experimental.vp.splat.v16i16(i16 undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %9 = call <2 x i32> @llvm.experimental.vp.splat.v2i32(i32 undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %10 = call <4 x i32> @llvm.experimental.vp.splat.v4i32(i32 undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %11 = call <8 x i32> @llvm.experimental.vp.splat.v8i32(i32 undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 110 for instruction: %12 = call <16 x i32> @llvm.experimental.vp.splat.v16i32(i32 undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %13 = call <2 x i64> @llvm.experimental.vp.splat.v2i64(i64 undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %14 = call <4 x i64> @llvm.experimental.vp.splat.v4i64(i64 undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %15 = call <8 x i64> @llvm.experimental.vp.splat.v8i64(i64 undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 110 for instruction: %16 = call <16 x i64> @llvm.experimental.vp.splat.v16i64(i64 undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %17 = call @llvm.experimental.vp.splat.nxv2i8(i8 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %18 = call @llvm.experimental.vp.splat.nxv4i8(i8 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %19 = call @llvm.experimental.vp.splat.nxv8i8(i8 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %20 = call @llvm.experimental.vp.splat.nxv16i8(i8 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %21 = call @llvm.experimental.vp.splat.nxv2i16(i16 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %22 = call @llvm.experimental.vp.splat.nxv4i16(i16 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %23 = call @llvm.experimental.vp.splat.nxv8i16(i16 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %24 = call @llvm.experimental.vp.splat.nxv16i16(i16 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %25 = call @llvm.experimental.vp.splat.nxv2i32(i32 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %26 = call @llvm.experimental.vp.splat.nxv4i32(i32 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %27 = call @llvm.experimental.vp.splat.nxv8i32(i32 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %28 = call @llvm.experimental.vp.splat.nxv16i32(i32 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %29 = call @llvm.experimental.vp.splat.nxv2i64(i64 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %30 = call @llvm.experimental.vp.splat.nxv4i64(i64 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %31 = call @llvm.experimental.vp.splat.nxv8i64(i64 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %32 = call @llvm.experimental.vp.splat.nxv16i64(i64 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void +; + call <2 x i8> @llvm.experimental.vp.splat.v2i8(i8 undef, <2 x i1> undef, i32 undef) + call <4 x i8> @llvm.experimental.vp.splat.v4i8(i8 undef, <4 x i1> undef, i32 undef) + call <8 x i8> @llvm.experimental.vp.splat.v8i8(i8 undef, <8 x i1> undef, i32 undef) + call <16 x i8> @llvm.experimental.vp.splat.v16i8(i8 undef, <16 x i1> undef, i32 undef) + call <2 x i16> @llvm.experimental.vp.splat.v2i16(i16 undef, <2 x i1> undef, i32 undef) + call <4 x i16> @llvm.experimental.vp.splat.v4i16(i16 undef, <4 x i1> undef, i32 undef) + call <8 x i16> @llvm.experimental.vp.splat.v8i16(i16 undef, <8 x i1> undef, i32 undef) + call <16 x i16> @llvm.experimental.vp.splat.v16i16(i16 undef, <16 x i1> undef, i32 undef) + call <2 x i32> @llvm.experimental.vp.splat.v2i32(i32 undef, <2 x i1> undef, i32 undef) + call <4 x i32> @llvm.experimental.vp.splat.v4i32(i32 undef, <4 x i1> undef, i32 undef) + call <8 x i32> @llvm.experimental.vp.splat.v8i32(i32 undef, <8 x i1> undef, i32 undef) + call <16 x i32> @llvm.experimental.vp.splat.v16i32(i32 undef, <16 x i1> undef, i32 undef) + call <2 x i64> @llvm.experimental.vp.splat.v2i64(i64 undef, <2 x i1> undef, i32 undef) + call <4 x i64> @llvm.experimental.vp.splat.v4i64(i64 undef, <4 x i1> undef, i32 undef) + call <8 x i64> @llvm.experimental.vp.splat.v8i64(i64 undef, <8 x i1> undef, i32 undef) + call <16 x i64> @llvm.experimental.vp.splat.v16i64(i64 undef, <16 x i1> undef, i32 undef) + call @llvm.experimental.vp.splat.nxv2i8(i8 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv4i8(i8 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv8i8(i8 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv16i8(i8 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv2i16(i16 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv4i16(i16 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv8i16(i16 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv16i16(i16 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv2i32(i32 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv4i32(i32 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv8i32(i32 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv16i32(i32 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv2i64(i64 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv4i64(i64 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv8i64(i64 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv16i64(i64 undef, undef, i32 undef) + ret void +} + declare <2 x i8> @llvm.vp.add.v2i8(<2 x i8>, <2 x i8>, <2 x i1>, i32) declare <4 x i8> @llvm.vp.add.v4i8(<4 x i8>, <4 x i8>, <4 x i1>, i32) declare <8 x i8> @llvm.vp.add.v8i8(<8 x i8>, <8 x i8>, <8 x i1>, i32) From eaac532672dc82c9cb6c325c5830e5d0bcfc91ad Mon Sep 17 00:00:00 2001 From: Luke Lau Date: Fri, 22 Nov 2024 18:13:15 +0800 Subject: [PATCH 2/5] [RISCV] Add cost for @llvm.experimental.vp.splat This is split off from #115274. There doesn't seem to be an easy way to share this with getShuffleCost since that requires passing in a real insert_element operand to get it to recognise it's a scalar splat. There's no tests for i1 vectors, since we can't lower vp splats of them yet and currently crash. Co-authored-by: Shih-Po Hung --- .../Target/RISCV/RISCVTargetTransformInfo.cpp | 5 + .../Analysis/CostModel/RISCV/vp-intrinsics.ll | 128 +++++++++--------- 2 files changed, 69 insertions(+), 64 deletions(-) diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp index 2b16dcbcd8695..ecdecd7edff07 100644 --- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp +++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp @@ -1155,6 +1155,11 @@ RISCVTTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA, return getCmpSelInstrCost(Instruction::Select, ICA.getReturnType(), ICA.getArgTypes()[0], CmpInst::BAD_ICMP_PREDICATE, CostKind); + case Intrinsic::experimental_vp_splat: { + auto LT = getTypeLegalizationCost(RetTy); + return LT.first * + getRISCVInstructionCost(RISCV::VMV_V_X, LT.second, CostKind); + } case Intrinsic::vp_reduce_add: case Intrinsic::vp_reduce_fadd: case Intrinsic::vp_reduce_mul: diff --git a/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll b/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll index 2b7f699e8b690..a7ce6f660ca4a 100644 --- a/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll +++ b/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll @@ -2127,73 +2127,73 @@ define void @vp_fdiv(){ define void @splat() { ; CHECK-LABEL: 'splat' -; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %1 = call <2 x i8> @llvm.experimental.vp.splat.v2i8(i8 undef, <2 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %2 = call <4 x i8> @llvm.experimental.vp.splat.v4i8(i8 undef, <4 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %3 = call <8 x i8> @llvm.experimental.vp.splat.v8i8(i8 undef, <8 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %4 = call <16 x i8> @llvm.experimental.vp.splat.v16i8(i8 undef, <16 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %5 = call <2 x i16> @llvm.experimental.vp.splat.v2i16(i16 undef, <2 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %6 = call <4 x i16> @llvm.experimental.vp.splat.v4i16(i16 undef, <4 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %7 = call <8 x i16> @llvm.experimental.vp.splat.v8i16(i16 undef, <8 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %8 = call <16 x i16> @llvm.experimental.vp.splat.v16i16(i16 undef, <16 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %9 = call <2 x i32> @llvm.experimental.vp.splat.v2i32(i32 undef, <2 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %10 = call <4 x i32> @llvm.experimental.vp.splat.v4i32(i32 undef, <4 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %11 = call <8 x i32> @llvm.experimental.vp.splat.v8i32(i32 undef, <8 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %12 = call <16 x i32> @llvm.experimental.vp.splat.v16i32(i32 undef, <16 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %13 = call <2 x i64> @llvm.experimental.vp.splat.v2i64(i64 undef, <2 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %14 = call <4 x i64> @llvm.experimental.vp.splat.v4i64(i64 undef, <4 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %15 = call <8 x i64> @llvm.experimental.vp.splat.v8i64(i64 undef, <8 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %16 = call <16 x i64> @llvm.experimental.vp.splat.v16i64(i64 undef, <16 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %17 = call @llvm.experimental.vp.splat.nxv2i8(i8 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %18 = call @llvm.experimental.vp.splat.nxv4i8(i8 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %19 = call @llvm.experimental.vp.splat.nxv8i8(i8 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %20 = call @llvm.experimental.vp.splat.nxv16i8(i8 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %21 = call @llvm.experimental.vp.splat.nxv2i16(i16 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %22 = call @llvm.experimental.vp.splat.nxv4i16(i16 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %23 = call @llvm.experimental.vp.splat.nxv8i16(i16 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %24 = call @llvm.experimental.vp.splat.nxv16i16(i16 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %25 = call @llvm.experimental.vp.splat.nxv2i32(i32 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %26 = call @llvm.experimental.vp.splat.nxv4i32(i32 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %27 = call @llvm.experimental.vp.splat.nxv8i32(i32 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %28 = call @llvm.experimental.vp.splat.nxv16i32(i32 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %29 = call @llvm.experimental.vp.splat.nxv2i64(i64 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %30 = call @llvm.experimental.vp.splat.nxv4i64(i64 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %31 = call @llvm.experimental.vp.splat.nxv8i64(i64 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Invalid cost for instruction: %32 = call @llvm.experimental.vp.splat.nxv16i64(i64 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call <2 x i8> @llvm.experimental.vp.splat.v2i8(i8 undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %2 = call <4 x i8> @llvm.experimental.vp.splat.v4i8(i8 undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %3 = call <8 x i8> @llvm.experimental.vp.splat.v8i8(i8 undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %4 = call <16 x i8> @llvm.experimental.vp.splat.v16i8(i8 undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %5 = call <2 x i16> @llvm.experimental.vp.splat.v2i16(i16 undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %6 = call <4 x i16> @llvm.experimental.vp.splat.v4i16(i16 undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <8 x i16> @llvm.experimental.vp.splat.v8i16(i16 undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %8 = call <16 x i16> @llvm.experimental.vp.splat.v16i16(i16 undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %9 = call <2 x i32> @llvm.experimental.vp.splat.v2i32(i32 undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %10 = call <4 x i32> @llvm.experimental.vp.splat.v4i32(i32 undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %11 = call <8 x i32> @llvm.experimental.vp.splat.v8i32(i32 undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %12 = call <16 x i32> @llvm.experimental.vp.splat.v16i32(i32 undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %13 = call <2 x i64> @llvm.experimental.vp.splat.v2i64(i64 undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %14 = call <4 x i64> @llvm.experimental.vp.splat.v4i64(i64 undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %15 = call <8 x i64> @llvm.experimental.vp.splat.v8i64(i64 undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %16 = call <16 x i64> @llvm.experimental.vp.splat.v16i64(i64 undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %17 = call @llvm.experimental.vp.splat.nxv2i8(i8 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %18 = call @llvm.experimental.vp.splat.nxv4i8(i8 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %19 = call @llvm.experimental.vp.splat.nxv8i8(i8 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %20 = call @llvm.experimental.vp.splat.nxv16i8(i8 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %21 = call @llvm.experimental.vp.splat.nxv2i16(i16 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %22 = call @llvm.experimental.vp.splat.nxv4i16(i16 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %23 = call @llvm.experimental.vp.splat.nxv8i16(i16 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %24 = call @llvm.experimental.vp.splat.nxv16i16(i16 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %25 = call @llvm.experimental.vp.splat.nxv2i32(i32 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %26 = call @llvm.experimental.vp.splat.nxv4i32(i32 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %27 = call @llvm.experimental.vp.splat.nxv8i32(i32 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %28 = call @llvm.experimental.vp.splat.nxv16i32(i32 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %29 = call @llvm.experimental.vp.splat.nxv2i64(i64 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %30 = call @llvm.experimental.vp.splat.nxv4i64(i64 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %31 = call @llvm.experimental.vp.splat.nxv8i64(i64 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %32 = call @llvm.experimental.vp.splat.nxv16i64(i64 undef, undef, i32 undef) ; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void ; ; TYPEBASED-LABEL: 'splat' -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %1 = call <2 x i8> @llvm.experimental.vp.splat.v2i8(i8 undef, <2 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %2 = call <4 x i8> @llvm.experimental.vp.splat.v4i8(i8 undef, <4 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %3 = call <8 x i8> @llvm.experimental.vp.splat.v8i8(i8 undef, <8 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 95 for instruction: %4 = call <16 x i8> @llvm.experimental.vp.splat.v16i8(i8 undef, <16 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %5 = call <2 x i16> @llvm.experimental.vp.splat.v2i16(i16 undef, <2 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %6 = call <4 x i16> @llvm.experimental.vp.splat.v4i16(i16 undef, <4 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %7 = call <8 x i16> @llvm.experimental.vp.splat.v8i16(i16 undef, <8 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 110 for instruction: %8 = call <16 x i16> @llvm.experimental.vp.splat.v16i16(i16 undef, <16 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %9 = call <2 x i32> @llvm.experimental.vp.splat.v2i32(i32 undef, <2 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %10 = call <4 x i32> @llvm.experimental.vp.splat.v4i32(i32 undef, <4 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %11 = call <8 x i32> @llvm.experimental.vp.splat.v8i32(i32 undef, <8 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 110 for instruction: %12 = call <16 x i32> @llvm.experimental.vp.splat.v16i32(i32 undef, <16 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %13 = call <2 x i64> @llvm.experimental.vp.splat.v2i64(i64 undef, <2 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %14 = call <4 x i64> @llvm.experimental.vp.splat.v4i64(i64 undef, <4 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %15 = call <8 x i64> @llvm.experimental.vp.splat.v8i64(i64 undef, <8 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 110 for instruction: %16 = call <16 x i64> @llvm.experimental.vp.splat.v16i64(i64 undef, <16 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %17 = call @llvm.experimental.vp.splat.nxv2i8(i8 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %18 = call @llvm.experimental.vp.splat.nxv4i8(i8 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %19 = call @llvm.experimental.vp.splat.nxv8i8(i8 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %20 = call @llvm.experimental.vp.splat.nxv16i8(i8 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %21 = call @llvm.experimental.vp.splat.nxv2i16(i16 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %22 = call @llvm.experimental.vp.splat.nxv4i16(i16 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %23 = call @llvm.experimental.vp.splat.nxv8i16(i16 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %24 = call @llvm.experimental.vp.splat.nxv16i16(i16 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %25 = call @llvm.experimental.vp.splat.nxv2i32(i32 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %26 = call @llvm.experimental.vp.splat.nxv4i32(i32 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %27 = call @llvm.experimental.vp.splat.nxv8i32(i32 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %28 = call @llvm.experimental.vp.splat.nxv16i32(i32 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %29 = call @llvm.experimental.vp.splat.nxv2i64(i64 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %30 = call @llvm.experimental.vp.splat.nxv4i64(i64 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %31 = call @llvm.experimental.vp.splat.nxv8i64(i64 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %32 = call @llvm.experimental.vp.splat.nxv16i64(i64 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call <2 x i8> @llvm.experimental.vp.splat.v2i8(i8 undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %2 = call <4 x i8> @llvm.experimental.vp.splat.v4i8(i8 undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %3 = call <8 x i8> @llvm.experimental.vp.splat.v8i8(i8 undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %4 = call <16 x i8> @llvm.experimental.vp.splat.v16i8(i8 undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %5 = call <2 x i16> @llvm.experimental.vp.splat.v2i16(i16 undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %6 = call <4 x i16> @llvm.experimental.vp.splat.v4i16(i16 undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <8 x i16> @llvm.experimental.vp.splat.v8i16(i16 undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %8 = call <16 x i16> @llvm.experimental.vp.splat.v16i16(i16 undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %9 = call <2 x i32> @llvm.experimental.vp.splat.v2i32(i32 undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %10 = call <4 x i32> @llvm.experimental.vp.splat.v4i32(i32 undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %11 = call <8 x i32> @llvm.experimental.vp.splat.v8i32(i32 undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %12 = call <16 x i32> @llvm.experimental.vp.splat.v16i32(i32 undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %13 = call <2 x i64> @llvm.experimental.vp.splat.v2i64(i64 undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %14 = call <4 x i64> @llvm.experimental.vp.splat.v4i64(i64 undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %15 = call <8 x i64> @llvm.experimental.vp.splat.v8i64(i64 undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %16 = call <16 x i64> @llvm.experimental.vp.splat.v16i64(i64 undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %17 = call @llvm.experimental.vp.splat.nxv2i8(i8 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %18 = call @llvm.experimental.vp.splat.nxv4i8(i8 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %19 = call @llvm.experimental.vp.splat.nxv8i8(i8 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %20 = call @llvm.experimental.vp.splat.nxv16i8(i8 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %21 = call @llvm.experimental.vp.splat.nxv2i16(i16 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %22 = call @llvm.experimental.vp.splat.nxv4i16(i16 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %23 = call @llvm.experimental.vp.splat.nxv8i16(i16 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %24 = call @llvm.experimental.vp.splat.nxv16i16(i16 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %25 = call @llvm.experimental.vp.splat.nxv2i32(i32 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %26 = call @llvm.experimental.vp.splat.nxv4i32(i32 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %27 = call @llvm.experimental.vp.splat.nxv8i32(i32 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %28 = call @llvm.experimental.vp.splat.nxv16i32(i32 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %29 = call @llvm.experimental.vp.splat.nxv2i64(i64 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %30 = call @llvm.experimental.vp.splat.nxv4i64(i64 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %31 = call @llvm.experimental.vp.splat.nxv8i64(i64 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %32 = call @llvm.experimental.vp.splat.nxv16i64(i64 undef, undef, i32 undef) ; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void ; call <2 x i8> @llvm.experimental.vp.splat.v2i8(i8 undef, <2 x i1> undef, i32 undef) From b50a648108a92ff7d280150450ca208cd5b70c2a Mon Sep 17 00:00:00 2001 From: Luke Lau Date: Fri, 22 Nov 2024 23:41:25 +0800 Subject: [PATCH 3/5] Add tests for fp and i1, return invalid for i1 vectors --- .../Target/RISCV/RISCVTargetTransformInfo.cpp | 3 + .../Analysis/CostModel/RISCV/vp-intrinsics.ll | 252 +++++++++++++----- 2 files changed, 189 insertions(+), 66 deletions(-) diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp index ecdecd7edff07..da2ea526d7602 100644 --- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp +++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp @@ -1156,6 +1156,9 @@ RISCVTTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA, ICA.getArgTypes()[0], CmpInst::BAD_ICMP_PREDICATE, CostKind); case Intrinsic::experimental_vp_splat: { + // TODO: Lower i1 experimental_vp_splat + if (cast(RetTy)->getElementType()->isIntegerTy(1)) + return InstructionCost::getInvalid(); auto LT = getTypeLegalizationCost(RetTy); return LT.first * getRISCVInstructionCost(RISCV::VMV_V_X, LT.second, CostKind); diff --git a/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll b/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll index a7ce6f660ca4a..77094da2d4c18 100644 --- a/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll +++ b/llvm/test/Analysis/CostModel/RISCV/vp-intrinsics.ll @@ -1,6 +1,6 @@ ; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py -; RUN: opt < %s -passes="print" 2>&1 -disable-output -S -mtriple=riscv64 -mattr=+v | FileCheck %s -; RUN: opt < %s -passes="print" 2>&1 -disable-output -S -mtriple=riscv64 -mattr=+v --type-based-intrinsic-cost=true | FileCheck %s --check-prefixes=TYPEBASED +; RUN: opt < %s -passes="print" 2>&1 -disable-output -S -mtriple=riscv64 -mattr=+v,+zvfhmin,+zvfbfmin | FileCheck %s +; RUN: opt < %s -passes="print" 2>&1 -disable-output -S -mtriple=riscv64 -mattr=+v,+zvfhmin,+zvfbfmin --type-based-intrinsic-cost=true | FileCheck %s --check-prefixes=TYPEBASED define void @unsupported_fp_ops( %vec, i32 %extraarg) { ; CHECK-LABEL: 'unsupported_fp_ops' @@ -2127,75 +2127,159 @@ define void @vp_fdiv(){ define void @splat() { ; CHECK-LABEL: 'splat' -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call <2 x i8> @llvm.experimental.vp.splat.v2i8(i8 undef, <2 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %2 = call <4 x i8> @llvm.experimental.vp.splat.v4i8(i8 undef, <4 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %3 = call <8 x i8> @llvm.experimental.vp.splat.v8i8(i8 undef, <8 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %4 = call <16 x i8> @llvm.experimental.vp.splat.v16i8(i8 undef, <16 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %5 = call <2 x i16> @llvm.experimental.vp.splat.v2i16(i16 undef, <2 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %6 = call <4 x i16> @llvm.experimental.vp.splat.v4i16(i16 undef, <4 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <8 x i16> @llvm.experimental.vp.splat.v8i16(i16 undef, <8 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %8 = call <16 x i16> @llvm.experimental.vp.splat.v16i16(i16 undef, <16 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %9 = call <2 x i32> @llvm.experimental.vp.splat.v2i32(i32 undef, <2 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %10 = call <4 x i32> @llvm.experimental.vp.splat.v4i32(i32 undef, <4 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %11 = call <8 x i32> @llvm.experimental.vp.splat.v8i32(i32 undef, <8 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %12 = call <16 x i32> @llvm.experimental.vp.splat.v16i32(i32 undef, <16 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %13 = call <2 x i64> @llvm.experimental.vp.splat.v2i64(i64 undef, <2 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %14 = call <4 x i64> @llvm.experimental.vp.splat.v4i64(i64 undef, <4 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %15 = call <8 x i64> @llvm.experimental.vp.splat.v8i64(i64 undef, <8 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %16 = call <16 x i64> @llvm.experimental.vp.splat.v16i64(i64 undef, <16 x i1> undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %17 = call @llvm.experimental.vp.splat.nxv2i8(i8 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %18 = call @llvm.experimental.vp.splat.nxv4i8(i8 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %19 = call @llvm.experimental.vp.splat.nxv8i8(i8 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %20 = call @llvm.experimental.vp.splat.nxv16i8(i8 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %21 = call @llvm.experimental.vp.splat.nxv2i16(i16 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %22 = call @llvm.experimental.vp.splat.nxv4i16(i16 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %23 = call @llvm.experimental.vp.splat.nxv8i16(i16 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %24 = call @llvm.experimental.vp.splat.nxv16i16(i16 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %25 = call @llvm.experimental.vp.splat.nxv2i32(i32 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %26 = call @llvm.experimental.vp.splat.nxv4i32(i32 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %27 = call @llvm.experimental.vp.splat.nxv8i32(i32 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %28 = call @llvm.experimental.vp.splat.nxv16i32(i32 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %29 = call @llvm.experimental.vp.splat.nxv2i64(i64 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %30 = call @llvm.experimental.vp.splat.nxv4i64(i64 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %31 = call @llvm.experimental.vp.splat.nxv8i64(i64 undef, undef, i32 undef) -; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %32 = call @llvm.experimental.vp.splat.nxv16i64(i64 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %1 = call <2 x i1> @llvm.experimental.vp.splat.v2i1(i1 undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %2 = call <4 x i1> @llvm.experimental.vp.splat.v4i1(i1 undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %3 = call <8 x i1> @llvm.experimental.vp.splat.v8i1(i1 undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %4 = call <16 x i1> @llvm.experimental.vp.splat.v16i1(i1 undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %5 = call <2 x i8> @llvm.experimental.vp.splat.v2i8(i8 undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %6 = call <4 x i8> @llvm.experimental.vp.splat.v4i8(i8 undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <8 x i8> @llvm.experimental.vp.splat.v8i8(i8 undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %8 = call <16 x i8> @llvm.experimental.vp.splat.v16i8(i8 undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %9 = call <2 x i16> @llvm.experimental.vp.splat.v2i16(i16 undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %10 = call <4 x i16> @llvm.experimental.vp.splat.v4i16(i16 undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %11 = call <8 x i16> @llvm.experimental.vp.splat.v8i16(i16 undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %12 = call <16 x i16> @llvm.experimental.vp.splat.v16i16(i16 undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %13 = call <2 x i32> @llvm.experimental.vp.splat.v2i32(i32 undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %14 = call <4 x i32> @llvm.experimental.vp.splat.v4i32(i32 undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %15 = call <8 x i32> @llvm.experimental.vp.splat.v8i32(i32 undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %16 = call <16 x i32> @llvm.experimental.vp.splat.v16i32(i32 undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %17 = call <2 x i64> @llvm.experimental.vp.splat.v2i64(i64 undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %18 = call <4 x i64> @llvm.experimental.vp.splat.v4i64(i64 undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %19 = call <8 x i64> @llvm.experimental.vp.splat.v8i64(i64 undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %20 = call <16 x i64> @llvm.experimental.vp.splat.v16i64(i64 undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %21 = call <2 x bfloat> @llvm.experimental.vp.splat.v2bf16(bfloat undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %22 = call <4 x bfloat> @llvm.experimental.vp.splat.v4bf16(bfloat undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %23 = call <8 x bfloat> @llvm.experimental.vp.splat.v8bf16(bfloat undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %24 = call <16 x bfloat> @llvm.experimental.vp.splat.v16bf16(bfloat undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %25 = call <2 x half> @llvm.experimental.vp.splat.v2f16(half undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %26 = call <4 x half> @llvm.experimental.vp.splat.v4f16(half undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %27 = call <8 x half> @llvm.experimental.vp.splat.v8f16(half undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %28 = call <16 x half> @llvm.experimental.vp.splat.v16f16(half undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %29 = call <2 x float> @llvm.experimental.vp.splat.v2f32(float undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %30 = call <4 x float> @llvm.experimental.vp.splat.v4f32(float undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %31 = call <8 x float> @llvm.experimental.vp.splat.v8f32(float undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %32 = call <16 x float> @llvm.experimental.vp.splat.v16f32(float undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %33 = call <2 x double> @llvm.experimental.vp.splat.v2f64(double undef, <2 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %34 = call <4 x double> @llvm.experimental.vp.splat.v4f64(double undef, <4 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %35 = call <8 x double> @llvm.experimental.vp.splat.v8f64(double undef, <8 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %36 = call <16 x double> @llvm.experimental.vp.splat.v16f64(double undef, <16 x i1> undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %37 = call @llvm.experimental.vp.splat.nxv2i1(i1 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %38 = call @llvm.experimental.vp.splat.nxv4i1(i1 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %39 = call @llvm.experimental.vp.splat.nxv8i1(i1 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Invalid cost for instruction: %40 = call @llvm.experimental.vp.splat.nxv16i1(i1 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %41 = call @llvm.experimental.vp.splat.nxv2i8(i8 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %42 = call @llvm.experimental.vp.splat.nxv4i8(i8 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %43 = call @llvm.experimental.vp.splat.nxv8i8(i8 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %44 = call @llvm.experimental.vp.splat.nxv16i8(i8 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %45 = call @llvm.experimental.vp.splat.nxv2i16(i16 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %46 = call @llvm.experimental.vp.splat.nxv4i16(i16 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %47 = call @llvm.experimental.vp.splat.nxv8i16(i16 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %48 = call @llvm.experimental.vp.splat.nxv16i16(i16 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %49 = call @llvm.experimental.vp.splat.nxv2i32(i32 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %50 = call @llvm.experimental.vp.splat.nxv4i32(i32 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %51 = call @llvm.experimental.vp.splat.nxv8i32(i32 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %52 = call @llvm.experimental.vp.splat.nxv16i32(i32 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %53 = call @llvm.experimental.vp.splat.nxv2i64(i64 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %54 = call @llvm.experimental.vp.splat.nxv4i64(i64 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %55 = call @llvm.experimental.vp.splat.nxv8i64(i64 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %56 = call @llvm.experimental.vp.splat.nxv16i64(i64 undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %57 = call @llvm.experimental.vp.splat.nxv2bf16(bfloat undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %58 = call @llvm.experimental.vp.splat.nxv4bf16(bfloat undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %59 = call @llvm.experimental.vp.splat.nxv8bf16(bfloat undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %60 = call @llvm.experimental.vp.splat.nxv16bf16(bfloat undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %61 = call @llvm.experimental.vp.splat.nxv2f16(half undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %62 = call @llvm.experimental.vp.splat.nxv4f16(half undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %63 = call @llvm.experimental.vp.splat.nxv8f16(half undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %64 = call @llvm.experimental.vp.splat.nxv16f16(half undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %65 = call @llvm.experimental.vp.splat.nxv2f32(float undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %66 = call @llvm.experimental.vp.splat.nxv4f32(float undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %67 = call @llvm.experimental.vp.splat.nxv8f32(float undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %68 = call @llvm.experimental.vp.splat.nxv16f32(float undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %69 = call @llvm.experimental.vp.splat.nxv2f64(double undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %70 = call @llvm.experimental.vp.splat.nxv4f64(double undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %71 = call @llvm.experimental.vp.splat.nxv8f64(double undef, undef, i32 undef) +; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %72 = call @llvm.experimental.vp.splat.nxv16f64(double undef, undef, i32 undef) ; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void ; ; TYPEBASED-LABEL: 'splat' -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %1 = call <2 x i8> @llvm.experimental.vp.splat.v2i8(i8 undef, <2 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %2 = call <4 x i8> @llvm.experimental.vp.splat.v4i8(i8 undef, <4 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %3 = call <8 x i8> @llvm.experimental.vp.splat.v8i8(i8 undef, <8 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %4 = call <16 x i8> @llvm.experimental.vp.splat.v16i8(i8 undef, <16 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %5 = call <2 x i16> @llvm.experimental.vp.splat.v2i16(i16 undef, <2 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %6 = call <4 x i16> @llvm.experimental.vp.splat.v4i16(i16 undef, <4 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <8 x i16> @llvm.experimental.vp.splat.v8i16(i16 undef, <8 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %8 = call <16 x i16> @llvm.experimental.vp.splat.v16i16(i16 undef, <16 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %9 = call <2 x i32> @llvm.experimental.vp.splat.v2i32(i32 undef, <2 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %10 = call <4 x i32> @llvm.experimental.vp.splat.v4i32(i32 undef, <4 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %11 = call <8 x i32> @llvm.experimental.vp.splat.v8i32(i32 undef, <8 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %12 = call <16 x i32> @llvm.experimental.vp.splat.v16i32(i32 undef, <16 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %13 = call <2 x i64> @llvm.experimental.vp.splat.v2i64(i64 undef, <2 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %14 = call <4 x i64> @llvm.experimental.vp.splat.v4i64(i64 undef, <4 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %15 = call <8 x i64> @llvm.experimental.vp.splat.v8i64(i64 undef, <8 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %16 = call <16 x i64> @llvm.experimental.vp.splat.v16i64(i64 undef, <16 x i1> undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %17 = call @llvm.experimental.vp.splat.nxv2i8(i8 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %18 = call @llvm.experimental.vp.splat.nxv4i8(i8 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %19 = call @llvm.experimental.vp.splat.nxv8i8(i8 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %20 = call @llvm.experimental.vp.splat.nxv16i8(i8 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %21 = call @llvm.experimental.vp.splat.nxv2i16(i16 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %22 = call @llvm.experimental.vp.splat.nxv4i16(i16 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %23 = call @llvm.experimental.vp.splat.nxv8i16(i16 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %24 = call @llvm.experimental.vp.splat.nxv16i16(i16 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %25 = call @llvm.experimental.vp.splat.nxv2i32(i32 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %26 = call @llvm.experimental.vp.splat.nxv4i32(i32 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %27 = call @llvm.experimental.vp.splat.nxv8i32(i32 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %28 = call @llvm.experimental.vp.splat.nxv16i32(i32 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %29 = call @llvm.experimental.vp.splat.nxv2i64(i64 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %30 = call @llvm.experimental.vp.splat.nxv4i64(i64 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %31 = call @llvm.experimental.vp.splat.nxv8i64(i64 undef, undef, i32 undef) -; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %32 = call @llvm.experimental.vp.splat.nxv16i64(i64 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %1 = call <2 x i1> @llvm.experimental.vp.splat.v2i1(i1 undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %2 = call <4 x i1> @llvm.experimental.vp.splat.v4i1(i1 undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %3 = call <8 x i1> @llvm.experimental.vp.splat.v8i1(i1 undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %4 = call <16 x i1> @llvm.experimental.vp.splat.v16i1(i1 undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %5 = call <2 x i8> @llvm.experimental.vp.splat.v2i8(i8 undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %6 = call <4 x i8> @llvm.experimental.vp.splat.v4i8(i8 undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %7 = call <8 x i8> @llvm.experimental.vp.splat.v8i8(i8 undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %8 = call <16 x i8> @llvm.experimental.vp.splat.v16i8(i8 undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %9 = call <2 x i16> @llvm.experimental.vp.splat.v2i16(i16 undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %10 = call <4 x i16> @llvm.experimental.vp.splat.v4i16(i16 undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %11 = call <8 x i16> @llvm.experimental.vp.splat.v8i16(i16 undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %12 = call <16 x i16> @llvm.experimental.vp.splat.v16i16(i16 undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %13 = call <2 x i32> @llvm.experimental.vp.splat.v2i32(i32 undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %14 = call <4 x i32> @llvm.experimental.vp.splat.v4i32(i32 undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %15 = call <8 x i32> @llvm.experimental.vp.splat.v8i32(i32 undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %16 = call <16 x i32> @llvm.experimental.vp.splat.v16i32(i32 undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %17 = call <2 x i64> @llvm.experimental.vp.splat.v2i64(i64 undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %18 = call <4 x i64> @llvm.experimental.vp.splat.v4i64(i64 undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %19 = call <8 x i64> @llvm.experimental.vp.splat.v8i64(i64 undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %20 = call <16 x i64> @llvm.experimental.vp.splat.v16i64(i64 undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %21 = call <2 x bfloat> @llvm.experimental.vp.splat.v2bf16(bfloat undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %22 = call <4 x bfloat> @llvm.experimental.vp.splat.v4bf16(bfloat undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %23 = call <8 x bfloat> @llvm.experimental.vp.splat.v8bf16(bfloat undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %24 = call <16 x bfloat> @llvm.experimental.vp.splat.v16bf16(bfloat undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %25 = call <2 x half> @llvm.experimental.vp.splat.v2f16(half undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %26 = call <4 x half> @llvm.experimental.vp.splat.v4f16(half undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %27 = call <8 x half> @llvm.experimental.vp.splat.v8f16(half undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %28 = call <16 x half> @llvm.experimental.vp.splat.v16f16(half undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %29 = call <2 x float> @llvm.experimental.vp.splat.v2f32(float undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %30 = call <4 x float> @llvm.experimental.vp.splat.v4f32(float undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %31 = call <8 x float> @llvm.experimental.vp.splat.v8f32(float undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %32 = call <16 x float> @llvm.experimental.vp.splat.v16f32(float undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %33 = call <2 x double> @llvm.experimental.vp.splat.v2f64(double undef, <2 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %34 = call <4 x double> @llvm.experimental.vp.splat.v4f64(double undef, <4 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %35 = call <8 x double> @llvm.experimental.vp.splat.v8f64(double undef, <8 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %36 = call <16 x double> @llvm.experimental.vp.splat.v16f64(double undef, <16 x i1> undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %37 = call @llvm.experimental.vp.splat.nxv2i1(i1 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %38 = call @llvm.experimental.vp.splat.nxv4i1(i1 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %39 = call @llvm.experimental.vp.splat.nxv8i1(i1 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %40 = call @llvm.experimental.vp.splat.nxv16i1(i1 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %41 = call @llvm.experimental.vp.splat.nxv2i8(i8 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %42 = call @llvm.experimental.vp.splat.nxv4i8(i8 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %43 = call @llvm.experimental.vp.splat.nxv8i8(i8 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %44 = call @llvm.experimental.vp.splat.nxv16i8(i8 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %45 = call @llvm.experimental.vp.splat.nxv2i16(i16 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %46 = call @llvm.experimental.vp.splat.nxv4i16(i16 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %47 = call @llvm.experimental.vp.splat.nxv8i16(i16 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %48 = call @llvm.experimental.vp.splat.nxv16i16(i16 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %49 = call @llvm.experimental.vp.splat.nxv2i32(i32 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %50 = call @llvm.experimental.vp.splat.nxv4i32(i32 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %51 = call @llvm.experimental.vp.splat.nxv8i32(i32 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %52 = call @llvm.experimental.vp.splat.nxv16i32(i32 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %53 = call @llvm.experimental.vp.splat.nxv2i64(i64 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %54 = call @llvm.experimental.vp.splat.nxv4i64(i64 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %55 = call @llvm.experimental.vp.splat.nxv8i64(i64 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %56 = call @llvm.experimental.vp.splat.nxv16i64(i64 undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %57 = call @llvm.experimental.vp.splat.nxv2bf16(bfloat undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %58 = call @llvm.experimental.vp.splat.nxv4bf16(bfloat undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %59 = call @llvm.experimental.vp.splat.nxv8bf16(bfloat undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %60 = call @llvm.experimental.vp.splat.nxv16bf16(bfloat undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %61 = call @llvm.experimental.vp.splat.nxv2f16(half undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %62 = call @llvm.experimental.vp.splat.nxv4f16(half undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %63 = call @llvm.experimental.vp.splat.nxv8f16(half undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %64 = call @llvm.experimental.vp.splat.nxv16f16(half undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %65 = call @llvm.experimental.vp.splat.nxv2f32(float undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %66 = call @llvm.experimental.vp.splat.nxv4f32(float undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %67 = call @llvm.experimental.vp.splat.nxv8f32(float undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %68 = call @llvm.experimental.vp.splat.nxv16f32(float undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %69 = call @llvm.experimental.vp.splat.nxv2f64(double undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %70 = call @llvm.experimental.vp.splat.nxv4f64(double undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %71 = call @llvm.experimental.vp.splat.nxv8f64(double undef, undef, i32 undef) +; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %72 = call @llvm.experimental.vp.splat.nxv16f64(double undef, undef, i32 undef) ; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void ; + call <2 x i1> @llvm.experimental.vp.splat.v2i1(i1 undef, <2 x i1> undef, i32 undef) + call <4 x i1> @llvm.experimental.vp.splat.v4i1(i1 undef, <4 x i1> undef, i32 undef) + call <8 x i1> @llvm.experimental.vp.splat.v8i1(i1 undef, <8 x i1> undef, i32 undef) + call <16 x i1> @llvm.experimental.vp.splat.v16i1(i1 undef, <16 x i1> undef, i32 undef) call <2 x i8> @llvm.experimental.vp.splat.v2i8(i8 undef, <2 x i1> undef, i32 undef) call <4 x i8> @llvm.experimental.vp.splat.v4i8(i8 undef, <4 x i1> undef, i32 undef) call <8 x i8> @llvm.experimental.vp.splat.v8i8(i8 undef, <8 x i1> undef, i32 undef) @@ -2212,6 +2296,26 @@ define void @splat() { call <4 x i64> @llvm.experimental.vp.splat.v4i64(i64 undef, <4 x i1> undef, i32 undef) call <8 x i64> @llvm.experimental.vp.splat.v8i64(i64 undef, <8 x i1> undef, i32 undef) call <16 x i64> @llvm.experimental.vp.splat.v16i64(i64 undef, <16 x i1> undef, i32 undef) + call <2 x bfloat> @llvm.experimental.vp.splat.v2bf16(bfloat undef, <2 x i1> undef, i32 undef) + call <4 x bfloat> @llvm.experimental.vp.splat.v4bf16(bfloat undef, <4 x i1> undef, i32 undef) + call <8 x bfloat> @llvm.experimental.vp.splat.v8bf16(bfloat undef, <8 x i1> undef, i32 undef) + call <16 x bfloat> @llvm.experimental.vp.splat.v16bf16(bfloat undef, <16 x i1> undef, i32 undef) + call <2 x half> @llvm.experimental.vp.splat.v2f16(half undef, <2 x i1> undef, i32 undef) + call <4 x half> @llvm.experimental.vp.splat.v4f16(half undef, <4 x i1> undef, i32 undef) + call <8 x half> @llvm.experimental.vp.splat.v8f16(half undef, <8 x i1> undef, i32 undef) + call <16 x half> @llvm.experimental.vp.splat.v16f16(half undef, <16 x i1> undef, i32 undef) + call <2 x float> @llvm.experimental.vp.splat.v2f32(float undef, <2 x i1> undef, i32 undef) + call <4 x float> @llvm.experimental.vp.splat.v4f32(float undef, <4 x i1> undef, i32 undef) + call <8 x float> @llvm.experimental.vp.splat.v8f32(float undef, <8 x i1> undef, i32 undef) + call <16 x float> @llvm.experimental.vp.splat.v16f32(float undef, <16 x i1> undef, i32 undef) + call <2 x double> @llvm.experimental.vp.splat.v2f64(double undef, <2 x i1> undef, i32 undef) + call <4 x double> @llvm.experimental.vp.splat.v4f64(double undef, <4 x i1> undef, i32 undef) + call <8 x double> @llvm.experimental.vp.splat.v8f64(double undef, <8 x i1> undef, i32 undef) + call <16 x double> @llvm.experimental.vp.splat.v16f64(double undef, <16 x i1> undef, i32 undef) + call @llvm.experimental.vp.splat.nxv2i1(i1 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv4i1(i1 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv8i1(i1 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv16i1(i1 undef, undef, i32 undef) call @llvm.experimental.vp.splat.nxv2i8(i8 undef, undef, i32 undef) call @llvm.experimental.vp.splat.nxv4i8(i8 undef, undef, i32 undef) call @llvm.experimental.vp.splat.nxv8i8(i8 undef, undef, i32 undef) @@ -2228,6 +2332,22 @@ define void @splat() { call @llvm.experimental.vp.splat.nxv4i64(i64 undef, undef, i32 undef) call @llvm.experimental.vp.splat.nxv8i64(i64 undef, undef, i32 undef) call @llvm.experimental.vp.splat.nxv16i64(i64 undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv2bf16(bfloat undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv4bf16(bfloat undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv8bf16(bfloat undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv16bf16(bfloat undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv2f16(half undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv4f16(half undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv8f16(half undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv16f16(half undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv2f32(float undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv4f32(float undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv8f32(float undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv16f32(float undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv2f64(double undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv4f64(double undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv8f64(double undef, undef, i32 undef) + call @llvm.experimental.vp.splat.nxv16f64(double undef, undef, i32 undef) ret void } From ec89b5cd2c3b1d9c5c4c63157051b41593a44a6e Mon Sep 17 00:00:00 2001 From: Luke Lau Date: Fri, 22 Nov 2024 23:48:52 +0800 Subject: [PATCH 4/5] Explicitly use VFMV_V_F --- llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp index da2ea526d7602..4b3a26ae82427 100644 --- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp +++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp @@ -1156,12 +1156,14 @@ RISCVTTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA, ICA.getArgTypes()[0], CmpInst::BAD_ICMP_PREDICATE, CostKind); case Intrinsic::experimental_vp_splat: { + auto LT = getTypeLegalizationCost(RetTy); // TODO: Lower i1 experimental_vp_splat - if (cast(RetTy)->getElementType()->isIntegerTy(1)) + if (LT.second.getScalarType() == MVT::i1) return InstructionCost::getInvalid(); - auto LT = getTypeLegalizationCost(RetTy); - return LT.first * - getRISCVInstructionCost(RISCV::VMV_V_X, LT.second, CostKind); + return LT.first * getRISCVInstructionCost(LT.second.isFloatingPoint() + ? RISCV::VFMV_V_F + : RISCV::VMV_V_X, + LT.second, CostKind); } case Intrinsic::vp_reduce_add: case Intrinsic::vp_reduce_fadd: From 2d02fa5527ad7370e375dd0541a341357514119e Mon Sep 17 00:00:00 2001 From: Luke Lau Date: Sat, 23 Nov 2024 00:22:31 +0800 Subject: [PATCH 5/5] Return invalid without V instructions --- llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp index 4b3a26ae82427..722bb79b893b4 100644 --- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp +++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp @@ -1158,7 +1158,7 @@ RISCVTTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA, case Intrinsic::experimental_vp_splat: { auto LT = getTypeLegalizationCost(RetTy); // TODO: Lower i1 experimental_vp_splat - if (LT.second.getScalarType() == MVT::i1) + if (!ST->hasVInstructions() || LT.second.getScalarType() == MVT::i1) return InstructionCost::getInvalid(); return LT.first * getRISCVInstructionCost(LT.second.isFloatingPoint() ? RISCV::VFMV_V_F