Skip to content

Commit edc71e2

Browse files
ElvisWang123ElvisWang123
authored
[RISCV][TTI] Add instruction cost for vp.load/store. (#109245)
This patch makes the instruction cost of vp.load/store same as their non-vp counterpart.
1 parent 7183771 commit edc71e2

File tree

2 files changed

+50
-32
lines changed

2 files changed

+50
-32
lines changed

‎llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp

Lines changed: 18 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1048,6 +1048,24 @@ RISCVTTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA,
10481048
return getCmpSelInstrCost(*FOp, ICA.getArgTypes()[0], ICA.getReturnType(),
10491049
UI->getPredicate(), CostKind);
10501050
}
1051+
// vp load/store
1052+
case Intrinsic::vp_load:
1053+
case Intrinsic::vp_store: {
1054+
Intrinsic::ID IID = ICA.getID();
1055+
std::optional<unsigned> FOp = VPIntrinsic::getFunctionalOpcodeForVP(IID);
1056+
auto *UI = dyn_cast<VPIntrinsic>(ICA.getInst());
1057+
1058+
if (!UI)
1059+
break;
1060+
assert(FOp.has_value());
1061+
if (ICA.getID() == Intrinsic::vp_load)
1062+
return getMemoryOpCost(
1063+
*FOp, ICA.getReturnType(), UI->getPointerAlignment(),
1064+
UI->getOperand(0)->getType()->getPointerAddressSpace(), CostKind);
1065+
return getMemoryOpCost(
1066+
*FOp, ICA.getArgTypes()[0], UI->getPointerAlignment(),
1067+
UI->getOperand(1)->getType()->getPointerAddressSpace(), CostKind);
1068+
}
10511069
}
10521070

10531071
if (ST->hasVInstructions() && RetTy->isVectorTy()) {

‎llvm/test/Analysis/CostModel/RISCV/rvv-intrinsics.ll

Lines changed: 32 additions & 32 deletions
Original file line numberDiff line numberDiff line change
@@ -852,37 +852,37 @@ define void @load() {
852852
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
853853
;
854854
; TYPEBASED-LABEL: 'load'
855-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %t0 = call <2 x i8> @llvm.vp.load.v2i8.p0(ptr undef, <2 x i1> undef, i32 undef)
855+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t0 = call <2 x i8> @llvm.vp.load.v2i8.p0(ptr undef, <2 x i1> undef, i32 undef)
856856
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t1 = load <2 x i8>, ptr undef, align 2
857-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %t2 = call <4 x i8> @llvm.vp.load.v4i8.p0(ptr undef, <4 x i1> undef, i32 undef)
857+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t2 = call <4 x i8> @llvm.vp.load.v4i8.p0(ptr undef, <4 x i1> undef, i32 undef)
858858
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t3 = load <4 x i8>, ptr undef, align 4
859-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %t4 = call <8 x i8> @llvm.vp.load.v8i8.p0(ptr undef, <8 x i1> undef, i32 undef)
859+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t4 = call <8 x i8> @llvm.vp.load.v8i8.p0(ptr undef, <8 x i1> undef, i32 undef)
860860
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t5 = load <8 x i8>, ptr undef, align 8
861-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 110 for instruction: %t6 = call <16 x i8> @llvm.vp.load.v16i8.p0(ptr undef, <16 x i1> undef, i32 undef)
861+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t6 = call <16 x i8> @llvm.vp.load.v16i8.p0(ptr undef, <16 x i1> undef, i32 undef)
862862
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t7 = load <16 x i8>, ptr undef, align 16
863-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %t8 = call <2 x i64> @llvm.vp.load.v2i64.p0(ptr undef, <2 x i1> undef, i32 undef)
863+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t8 = call <2 x i64> @llvm.vp.load.v2i64.p0(ptr undef, <2 x i1> undef, i32 undef)
864864
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t9 = load <2 x i64>, ptr undef, align 16
865-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %t10 = call <4 x i64> @llvm.vp.load.v4i64.p0(ptr undef, <4 x i1> undef, i32 undef)
865+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %t10 = call <4 x i64> @llvm.vp.load.v4i64.p0(ptr undef, <4 x i1> undef, i32 undef)
866866
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %t12 = load <4 x i64>, ptr undef, align 32
867-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %t13 = call <8 x i64> @llvm.vp.load.v8i64.p0(ptr undef, <8 x i1> undef, i32 undef)
867+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %t13 = call <8 x i64> @llvm.vp.load.v8i64.p0(ptr undef, <8 x i1> undef, i32 undef)
868868
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %t14 = load <8 x i64>, ptr undef, align 64
869-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 110 for instruction: %t15 = call <16 x i64> @llvm.vp.load.v16i64.p0(ptr undef, <16 x i1> undef, i32 undef)
869+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %t15 = call <16 x i64> @llvm.vp.load.v16i64.p0(ptr undef, <16 x i1> undef, i32 undef)
870870
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %t16 = load <16 x i64>, ptr undef, align 128
871-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %t17 = call <vscale x 2 x i8> @llvm.vp.load.nxv2i8.p0(ptr undef, <vscale x 2 x i1> undef, i32 undef)
871+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t17 = call <vscale x 2 x i8> @llvm.vp.load.nxv2i8.p0(ptr undef, <vscale x 2 x i1> undef, i32 undef)
872872
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t18 = load <vscale x 2 x i8>, ptr undef, align 2
873-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %t19 = call <vscale x 4 x i8> @llvm.vp.load.nxv4i8.p0(ptr undef, <vscale x 4 x i1> undef, i32 undef)
873+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t19 = call <vscale x 4 x i8> @llvm.vp.load.nxv4i8.p0(ptr undef, <vscale x 4 x i1> undef, i32 undef)
874874
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t20 = load <vscale x 4 x i8>, ptr undef, align 4
875-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %t21 = call <vscale x 8 x i8> @llvm.vp.load.nxv8i8.p0(ptr undef, <vscale x 8 x i1> undef, i32 undef)
875+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t21 = call <vscale x 8 x i8> @llvm.vp.load.nxv8i8.p0(ptr undef, <vscale x 8 x i1> undef, i32 undef)
876876
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %t22 = load <vscale x 8 x i8>, ptr undef, align 8
877-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %t23 = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr undef, <vscale x 16 x i1> undef, i32 undef)
877+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %t23 = call <vscale x 16 x i8> @llvm.vp.load.nxv16i8.p0(ptr undef, <vscale x 16 x i1> undef, i32 undef)
878878
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %t24 = load <vscale x 16 x i8>, ptr undef, align 16
879-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %t25 = call <vscale x 2 x i64> @llvm.vp.load.nxv2i64.p0(ptr undef, <vscale x 2 x i1> undef, i32 undef)
879+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %t25 = call <vscale x 2 x i64> @llvm.vp.load.nxv2i64.p0(ptr undef, <vscale x 2 x i1> undef, i32 undef)
880880
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %t26 = load <vscale x 2 x i64>, ptr undef, align 16
881-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %t27 = call <vscale x 4 x i64> @llvm.vp.load.nxv4i64.p0(ptr undef, <vscale x 4 x i1> undef, i32 undef)
881+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %t27 = call <vscale x 4 x i64> @llvm.vp.load.nxv4i64.p0(ptr undef, <vscale x 4 x i1> undef, i32 undef)
882882
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %t28 = load <vscale x 4 x i64>, ptr undef, align 32
883-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %t29 = call <vscale x 8 x i64> @llvm.vp.load.nxv8i64.p0(ptr undef, <vscale x 8 x i1> undef, i32 undef)
883+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %t29 = call <vscale x 8 x i64> @llvm.vp.load.nxv8i64.p0(ptr undef, <vscale x 8 x i1> undef, i32 undef)
884884
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %t30 = load <vscale x 8 x i64>, ptr undef, align 64
885-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: %t31 = call <vscale x 16 x i64> @llvm.vp.load.nxv16i64.p0(ptr undef, <vscale x 16 x i1> undef, i32 undef)
885+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %t31 = call <vscale x 16 x i64> @llvm.vp.load.nxv16i64.p0(ptr undef, <vscale x 16 x i1> undef, i32 undef)
886886
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %t32 = load <vscale x 16 x i64>, ptr undef, align 128
887887
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
888888
;
@@ -958,37 +958,37 @@ define void @store() {
958958
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
959959
;
960960
; TYPEBASED-LABEL: 'store'
961-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 12 for instruction: call void @llvm.vp.store.v2i8.p0(<2 x i8> undef, ptr undef, <2 x i1> undef, i32 undef)
961+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: call void @llvm.vp.store.v2i8.p0(<2 x i8> undef, ptr undef, <2 x i1> undef, i32 undef)
962962
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <2 x i8> undef, ptr undef, align 2
963-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 26 for instruction: call void @llvm.vp.store.v4i8.p0(<4 x i8> undef, ptr undef, <4 x i1> undef, i32 undef)
963+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: call void @llvm.vp.store.v4i8.p0(<4 x i8> undef, ptr undef, <4 x i1> undef, i32 undef)
964964
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <4 x i8> undef, ptr undef, align 4
965-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 54 for instruction: call void @llvm.vp.store.v8i8.p0(<8 x i8> undef, ptr undef, <8 x i1> undef, i32 undef)
965+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: call void @llvm.vp.store.v8i8.p0(<8 x i8> undef, ptr undef, <8 x i1> undef, i32 undef)
966966
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <8 x i8> undef, ptr undef, align 8
967-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 110 for instruction: call void @llvm.vp.store.v16i8.p0(<16 x i8> undef, ptr undef, <16 x i1> undef, i32 undef)
967+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: call void @llvm.vp.store.v16i8.p0(<16 x i8> undef, ptr undef, <16 x i1> undef, i32 undef)
968968
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <16 x i8> undef, ptr undef, align 16
969-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 12 for instruction: call void @llvm.vp.store.v2i64.p0(<2 x i64> undef, ptr undef, <2 x i1> undef, i32 undef)
969+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: call void @llvm.vp.store.v2i64.p0(<2 x i64> undef, ptr undef, <2 x i1> undef, i32 undef)
970970
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <2 x i64> undef, ptr undef, align 16
971-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 26 for instruction: call void @llvm.vp.store.v4i64.p0(<4 x i64> undef, ptr undef, <4 x i1> undef, i32 undef)
971+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.vp.store.v4i64.p0(<4 x i64> undef, ptr undef, <4 x i1> undef, i32 undef)
972972
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: store <4 x i64> undef, ptr undef, align 32
973-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 54 for instruction: call void @llvm.vp.store.v8i64.p0(<8 x i64> undef, ptr undef, <8 x i1> undef, i32 undef)
973+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.vp.store.v8i64.p0(<8 x i64> undef, ptr undef, <8 x i1> undef, i32 undef)
974974
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <8 x i64> undef, ptr undef, align 64
975-
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 110 for instruction: call void @llvm.vp.store.v16i64.p0(<16 x i64> undef, ptr undef, <16 x i1> undef, i32 undef)
975+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.vp.store.v16i64.p0(<16 x i64> undef, ptr undef, <16 x i1> undef, i32 undef)
976976
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: store <16 x i64> undef, ptr undef, align 128
977-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.vp.store.nxv2i8.p0(<vscale x 2 x i8> undef, ptr undef, <vscale x 2 x i1> undef, i32 undef)
977+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: call void @llvm.vp.store.nxv2i8.p0(<vscale x 2 x i8> undef, ptr undef, <vscale x 2 x i1> undef, i32 undef)
978978
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <vscale x 2 x i8> undef, ptr undef, align 2
979-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.vp.store.nxv4i8.p0(<vscale x 4 x i8> undef, ptr undef, <vscale x 4 x i1> undef, i32 undef)
979+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: call void @llvm.vp.store.nxv4i8.p0(<vscale x 4 x i8> undef, ptr undef, <vscale x 4 x i1> undef, i32 undef)
980980
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <vscale x 4 x i8> undef, ptr undef, align 4
981-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.vp.store.nxv8i8.p0(<vscale x 8 x i8> undef, ptr undef, <vscale x 8 x i1> undef, i32 undef)
981+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: call void @llvm.vp.store.nxv8i8.p0(<vscale x 8 x i8> undef, ptr undef, <vscale x 8 x i1> undef, i32 undef)
982982
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <vscale x 8 x i8> undef, ptr undef, align 8
983-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> undef, ptr undef, <vscale x 16 x i1> undef, i32 undef)
983+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.vp.store.nxv16i8.p0(<vscale x 16 x i8> undef, ptr undef, <vscale x 16 x i1> undef, i32 undef)
984984
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: store <vscale x 16 x i8> undef, ptr undef, align 16
985-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.vp.store.nxv2i64.p0(<vscale x 2 x i64> undef, ptr undef, <vscale x 2 x i1> undef, i32 undef)
985+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.vp.store.nxv2i64.p0(<vscale x 2 x i64> undef, ptr undef, <vscale x 2 x i1> undef, i32 undef)
986986
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: store <vscale x 2 x i64> undef, ptr undef, align 16
987-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.vp.store.nxv4i64.p0(<vscale x 4 x i64> undef, ptr undef, <vscale x 4 x i1> undef, i32 undef)
987+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.vp.store.nxv4i64.p0(<vscale x 4 x i64> undef, ptr undef, <vscale x 4 x i1> undef, i32 undef)
988988
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <vscale x 4 x i64> undef, ptr undef, align 32
989-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.vp.store.nxv8i64.p0(<vscale x 8 x i64> undef, ptr undef, <vscale x 8 x i1> undef, i32 undef)
989+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.vp.store.nxv8i64.p0(<vscale x 8 x i64> undef, ptr undef, <vscale x 8 x i1> undef, i32 undef)
990990
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: store <vscale x 8 x i64> undef, ptr undef, align 64
991-
; TYPEBASED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.vp.store.nxv16i64.p0(<vscale x 16 x i64> undef, ptr undef, <vscale x 16 x i1> undef, i32 undef)
991+
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.vp.store.nxv16i64.p0(<vscale x 16 x i64> undef, ptr undef, <vscale x 16 x i1> undef, i32 undef)
992992
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: store <vscale x 16 x i64> undef, ptr undef, align 128
993993
; TYPEBASED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
994994
;

0 commit comments

Comments
 (0)