Skip to content

[RISCV] Cost @llvm.vector.{extract,insert} as free at index 0 #81818

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

[RISCV] Cost @llvm.vector.{extract,insert} as free at index 0 #81818

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
28 changes: 28 additions & 0 deletions llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -809,6 +809,34 @@ RISCVTTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA,
}
break;
}
case Intrinsic::vector_extract: {
// A vector extract at index 0 is a (free) subregister extract.
if (auto *CIdx = dyn_cast<ConstantInt>(ICA.getArgs()[1]);
CIdx && CIdx->isZero())
return TTI::TCC_Free;
break;
}
case Intrinsic::vector_insert: {
auto FitsSubreg = [this](Type *Ty) {
if (!isa<ScalableVectorType>(Ty))
return false;
// Any scalable vector LMUL >= 1 will fit exactly into a register group.
auto [_Cost, LT] = getTypeLegalizationCost(Ty);
auto [_Coeff, Fractional] =
RISCVVType::decodeVLMUL(RISCVTargetLowering::getLMUL(LT));
return !Fractional;
};

// A vector insert at index 0 is a (free) subregister insert if:
//
// - The subvec fits exactly into a register group or
// - The vector is undef
if (auto *CIdx = dyn_cast<ConstantInt>(ICA.getArgs()[2]);
CIdx && CIdx->isZero() &&
(FitsSubreg(ICA.getArgTypes()[1]) || isa<UndefValue>(ICA.getArgs()[0])))
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

isa(ICA.getArgs()[0])? Should it be isa(ICA.getArgs()[1])?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I believe we check if it's the vector that's undef, not the subvector, because if the vector is undef and the index is 0 we can just "replace" the entire vector with the subvector:

llvm-project/llvm/lib/Target/RISCV/RISCVISelLowering.cpp

Lines 9655 to 9668 in f01ed3b

// 1. If the Idx has been completely eliminated and this subvector's size is
// a vector register or a multiple thereof, or the surrounding elements are
// undef, then this is a subvector insert which naturally aligns to a vector
// register. These can easily be handled using subregister manipulation.
// 2. If the subvector is smaller than a vector register, then the insertion
// must preserve the undisturbed elements of the register. We do this by
// lowering to an EXTRACT_SUBVECTOR grabbing the nearest LMUL=1 vector type
// (which resolves to a subregister copy), performing a VSLIDEUP to place the
// subvector within the vector register, and an INSERT_SUBVECTOR of that
// LMUL=1 type back into the larger vector (resolving to another subregister
// operation). See below for how our VSLIDEUP works. We go via a LMUL=1 type
// to avoid allocating a large register group to hold our subvector.
if (RemIdx == 0 && (!IsSubVecPartReg || Vec.isUndef()))
return Op;

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ah, yes, just thought about it in shufflevector form, where the second argument is undef in a canonical form.

return TTI::TCC_Free;
break;
}
// TODO: add more intrinsic
case Intrinsic::experimental_stepvector: {
unsigned Cost = 1; // vid
Expand Down
12 changes: 6 additions & 6 deletions llvm/test/Analysis/CostModel/RISCV/rvv-shuffle.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -52,17 +52,17 @@ define void @vector_broadcast() {

define void @vector_insert_extract(<vscale x 4 x i32> %v0, <vscale x 16 x i32> %v1, <16 x i32> %v2) {
; CHECK-LABEL: 'vector_insert_extract'
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %extract_fixed_from_scalable = call <16 x i32> @llvm.vector.extract.v16i32.nxv4i32(<vscale x 4 x i32> %v0, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %extract_fixed_from_scalable = call <16 x i32> @llvm.vector.extract.v16i32.nxv4i32(<vscale x 4 x i32> %v0, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %insert_fixed_into_scalable = call <vscale x 4 x i32> @llvm.vector.insert.nxv4i32.v16i32(<vscale x 4 x i32> %v0, <16 x i32> %v2, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %extract_scalable_from_scalable = call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> %v1, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %insert_scalable_into_scalable = call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> %v1, <vscale x 4 x i32> %v0, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %extract_scalable_from_scalable = call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> %v1, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %insert_scalable_into_scalable = call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> %v1, <vscale x 4 x i32> %v0, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SIZE-LABEL: 'vector_insert_extract'
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %extract_fixed_from_scalable = call <16 x i32> @llvm.vector.extract.v16i32.nxv4i32(<vscale x 4 x i32> %v0, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %extract_fixed_from_scalable = call <16 x i32> @llvm.vector.extract.v16i32.nxv4i32(<vscale x 4 x i32> %v0, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %insert_fixed_into_scalable = call <vscale x 4 x i32> @llvm.vector.insert.nxv4i32.v16i32(<vscale x 4 x i32> %v0, <16 x i32> %v2, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %extract_scalable_from_scalable = call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> %v1, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %insert_scalable_into_scalable = call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> %v1, <vscale x 4 x i32> %v0, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %extract_scalable_from_scalable = call <vscale x 4 x i32> @llvm.vector.extract.nxv4i32.nxv16i32(<vscale x 16 x i32> %v1, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %insert_scalable_into_scalable = call <vscale x 16 x i32> @llvm.vector.insert.nxv16i32.nxv4i32(<vscale x 16 x i32> %v1, <vscale x 4 x i32> %v0, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
;
%extract_fixed_from_scalable = call <16 x i32> @llvm.vector.extract.v16i32.nxv4i32(<vscale x 4 x i32> %v0, i64 0)
Expand Down
84 changes: 42 additions & 42 deletions llvm/test/Analysis/CostModel/RISCV/rvv-vectorextract.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,37 +4,37 @@

define void @vector_extract_nxv128i8_0(<vscale x 128 x i8> %v) {
; CHECK-LABEL: 'vector_extract_nxv128i8_0'
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_mf8 = call <vscale x 1 x i8> @llvm.vector.extract.nxv1i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_mf4 = call <vscale x 2 x i8> @llvm.vector.extract.nxv2i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_mf2 = call <vscale x 4 x i8> @llvm.vector.extract.nxv4i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_m1 = call <vscale x 8 x i8> @llvm.vector.extract.nxv8i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_m2 = call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_m4 = call <vscale x 32 x i8> @llvm.vector.extract.nxv32i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_m8 = call <vscale x 64 x i8> @llvm.vector.extract.nxv64i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %fixed_mf8 = call <2 x i8> @llvm.vector.extract.v2i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %fixed_mf4 = call <4 x i8> @llvm.vector.extract.v4i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %fixed_mf2 = call <8 x i8> @llvm.vector.extract.v8i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %fixed_m1 = call <16 x i8> @llvm.vector.extract.v16i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %fixed_m2 = call <32 x i8> @llvm.vector.extract.v32i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %fixed_m4 = call <64 x i8> @llvm.vector.extract.v64i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %fixed_m8 = call <128 x i8> @llvm.vector.extract.v128i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_mf8 = call <vscale x 1 x i8> @llvm.vector.extract.nxv1i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_mf4 = call <vscale x 2 x i8> @llvm.vector.extract.nxv2i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_mf2 = call <vscale x 4 x i8> @llvm.vector.extract.nxv4i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_m1 = call <vscale x 8 x i8> @llvm.vector.extract.nxv8i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_m2 = call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_m4 = call <vscale x 32 x i8> @llvm.vector.extract.nxv32i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_m8 = call <vscale x 64 x i8> @llvm.vector.extract.nxv64i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_mf8 = call <2 x i8> @llvm.vector.extract.v2i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_mf4 = call <4 x i8> @llvm.vector.extract.v4i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_mf2 = call <8 x i8> @llvm.vector.extract.v8i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m1 = call <16 x i8> @llvm.vector.extract.v16i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m2 = call <32 x i8> @llvm.vector.extract.v32i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m4 = call <64 x i8> @llvm.vector.extract.v64i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m8 = call <128 x i8> @llvm.vector.extract.v128i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SIZE-LABEL: 'vector_extract_nxv128i8_0'
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_mf8 = call <vscale x 1 x i8> @llvm.vector.extract.nxv1i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_mf4 = call <vscale x 2 x i8> @llvm.vector.extract.nxv2i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_mf2 = call <vscale x 4 x i8> @llvm.vector.extract.nxv4i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_m1 = call <vscale x 8 x i8> @llvm.vector.extract.nxv8i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_m2 = call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_m4 = call <vscale x 32 x i8> @llvm.vector.extract.nxv32i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %scalable_m8 = call <vscale x 64 x i8> @llvm.vector.extract.nxv64i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %fixed_mf8 = call <2 x i8> @llvm.vector.extract.v2i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %fixed_mf4 = call <4 x i8> @llvm.vector.extract.v4i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %fixed_mf2 = call <8 x i8> @llvm.vector.extract.v8i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %fixed_m1 = call <16 x i8> @llvm.vector.extract.v16i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %fixed_m2 = call <32 x i8> @llvm.vector.extract.v32i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %fixed_m4 = call <64 x i8> @llvm.vector.extract.v64i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %fixed_m8 = call <128 x i8> @llvm.vector.extract.v128i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_mf8 = call <vscale x 1 x i8> @llvm.vector.extract.nxv1i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_mf4 = call <vscale x 2 x i8> @llvm.vector.extract.nxv2i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_mf2 = call <vscale x 4 x i8> @llvm.vector.extract.nxv4i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_m1 = call <vscale x 8 x i8> @llvm.vector.extract.nxv8i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_m2 = call <vscale x 16 x i8> @llvm.vector.extract.nxv16i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_m4 = call <vscale x 32 x i8> @llvm.vector.extract.nxv32i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %scalable_m8 = call <vscale x 64 x i8> @llvm.vector.extract.nxv64i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_mf8 = call <2 x i8> @llvm.vector.extract.v2i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_mf4 = call <4 x i8> @llvm.vector.extract.v4i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_mf2 = call <8 x i8> @llvm.vector.extract.v8i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m1 = call <16 x i8> @llvm.vector.extract.v16i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m2 = call <32 x i8> @llvm.vector.extract.v32i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m4 = call <64 x i8> @llvm.vector.extract.v64i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m8 = call <128 x i8> @llvm.vector.extract.v128i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
;
%scalable_mf8 = call <vscale x 1 x i8> @llvm.vector.extract.nxv1i8.nxv128i8(<vscale x 128 x i8> %v, i64 0)
Expand Down Expand Up @@ -110,23 +110,23 @@ define void @vector_extract_nxv128i8_1(<vscale x 128 x i8> %v) {

define void @vector_extract_v128i8_0(<128 x i8> %v) {
; CHECK-LABEL: 'vector_extract_v128i8_0'
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %fixed_mf8 = call <2 x i8> @llvm.vector.extract.v2i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %fixed_mf4 = call <4 x i8> @llvm.vector.extract.v4i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %fixed_mf2 = call <8 x i8> @llvm.vector.extract.v8i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %fixed_m1 = call <16 x i8> @llvm.vector.extract.v16i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %fixed_m2 = call <32 x i8> @llvm.vector.extract.v32i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %fixed_m4 = call <64 x i8> @llvm.vector.extract.v64i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %fixed_m8 = call <128 x i8> @llvm.vector.extract.v128i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_mf8 = call <2 x i8> @llvm.vector.extract.v2i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_mf4 = call <4 x i8> @llvm.vector.extract.v4i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_mf2 = call <8 x i8> @llvm.vector.extract.v8i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m1 = call <16 x i8> @llvm.vector.extract.v16i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m2 = call <32 x i8> @llvm.vector.extract.v32i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m4 = call <64 x i8> @llvm.vector.extract.v64i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m8 = call <128 x i8> @llvm.vector.extract.v128i8.v128i8(<128 x i8> %v, i64 0)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SIZE-LABEL: 'vector_extract_v128i8_0'
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %fixed_mf8 = call <2 x i8> @llvm.vector.extract.v2i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %fixed_mf4 = call <4 x i8> @llvm.vector.extract.v4i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %fixed_mf2 = call <8 x i8> @llvm.vector.extract.v8i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %fixed_m1 = call <16 x i8> @llvm.vector.extract.v16i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %fixed_m2 = call <32 x i8> @llvm.vector.extract.v32i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %fixed_m4 = call <64 x i8> @llvm.vector.extract.v64i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %fixed_m8 = call <128 x i8> @llvm.vector.extract.v128i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_mf8 = call <2 x i8> @llvm.vector.extract.v2i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_mf4 = call <4 x i8> @llvm.vector.extract.v4i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_mf2 = call <8 x i8> @llvm.vector.extract.v8i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m1 = call <16 x i8> @llvm.vector.extract.v16i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m2 = call <32 x i8> @llvm.vector.extract.v32i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m4 = call <64 x i8> @llvm.vector.extract.v64i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %fixed_m8 = call <128 x i8> @llvm.vector.extract.v128i8.v128i8(<128 x i8> %v, i64 0)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret void
;
%fixed_mf8 = call <2 x i8> @llvm.vector.extract.v2i8.v128i8(<128 x i8> %v, i64 0)
Expand Down
Loading