[clang] [llvm] [NVPTX] Support i256 load/store with 256-bit vector load (PR #155198)
Alex MacLean via llvm-commits
llvm-commits at lists.llvm.org
Mon Aug 25 11:02:59 PDT 2025
https://github.com/AlexMaclean updated https://github.com/llvm/llvm-project/pull/155198
>From f7e19af684716853b9b0ea6f0db101db35c9ec77 Mon Sep 17 00:00:00 2001
From: Alex Maclean <amaclean at nvidia.com>
Date: Mon, 25 Aug 2025 02:45:10 +0000
Subject: [PATCH 1/2] [NVPTX] Support i256 load/store with 256-bit vector load
---
clang/lib/Basic/Targets/NVPTX.cpp | 13 +-
clang/test/CodeGen/target-data.c | 4 +-
llvm/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp | 27 +-
llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp | 271 ++++++++++--------
llvm/lib/Target/NVPTX/NVPTXISelLowering.h | 1 -
llvm/lib/Target/NVPTX/NVPTXTargetMachine.cpp | 6 +-
.../CodeGen/NVPTX/load-store-vectors-256.ll | 66 +++++
7 files changed, 234 insertions(+), 154 deletions(-)
diff --git a/clang/lib/Basic/Targets/NVPTX.cpp b/clang/lib/Basic/Targets/NVPTX.cpp
index 5cf2dc187b836..f7abc05903cdf 100644
--- a/clang/lib/Basic/Targets/NVPTX.cpp
+++ b/clang/lib/Basic/Targets/NVPTX.cpp
@@ -69,14 +69,15 @@ NVPTXTargetInfo::NVPTXTargetInfo(const llvm::Triple &Triple,
HasFloat16 = true;
if (TargetPointerWidth == 32)
- resetDataLayout(
- "e-p:32:32-p6:32:32-p7:32:32-i64:64-i128:128-v16:16-v32:32-n16:32:64");
+ resetDataLayout("e-p:32:32-p6:32:32-p7:32:32-i64:64-i128:128-i256:256-v16:"
+ "16-v32:32-n16:32:64");
else if (Opts.NVPTXUseShortPointers)
- resetDataLayout(
- "e-p3:32:32-p4:32:32-p5:32:32-p6:32:32-p7:32:32-i64:64-i128:128-v16:"
- "16-v32:32-n16:32:64");
+ resetDataLayout("e-p3:32:32-p4:32:32-p5:32:32-p6:32:32-p7:32:32-i64:64-"
+ "i128:128-i256:256-v16:"
+ "16-v32:32-n16:32:64");
else
- resetDataLayout("e-p6:32:32-i64:64-i128:128-v16:16-v32:32-n16:32:64");
+ resetDataLayout(
+ "e-p6:32:32-i64:64-i128:128-i256:256-v16:16-v32:32-n16:32:64");
// If possible, get a TargetInfo for our host triple, so we can match its
// types.
diff --git a/clang/test/CodeGen/target-data.c b/clang/test/CodeGen/target-data.c
index 92fe3eb6f171c..eecee69e14122 100644
--- a/clang/test/CodeGen/target-data.c
+++ b/clang/test/CodeGen/target-data.c
@@ -144,11 +144,11 @@
// RUN: %clang_cc1 -triple nvptx-unknown -o - -emit-llvm %s | \
// RUN: FileCheck %s -check-prefix=NVPTX
-// NVPTX: target datalayout = "e-p:32:32-p6:32:32-p7:32:32-i64:64-i128:128-v16:16-v32:32-n16:32:64"
+// NVPTX: target datalayout = "e-p:32:32-p6:32:32-p7:32:32-i64:64-i128:128-i256:256-v16:16-v32:32-n16:32:64"
// RUN: %clang_cc1 -triple nvptx64-unknown -o - -emit-llvm %s | \
// RUN: FileCheck %s -check-prefix=NVPTX64
-// NVPTX64: target datalayout = "e-p6:32:32-i64:64-i128:128-v16:16-v32:32-n16:32:64"
+// NVPTX64: target datalayout = "e-p6:32:32-i64:64-i128:128-i256:256-v16:16-v32:32-n16:32:64"
// RUN: %clang_cc1 -triple r600-unknown -o - -emit-llvm %s | \
// RUN: FileCheck %s -check-prefix=R600
diff --git a/llvm/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp b/llvm/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp
index 3300ed9a5a81c..964b93ed2527c 100644
--- a/llvm/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp
+++ b/llvm/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp
@@ -1097,11 +1097,6 @@ bool NVPTXDAGToDAGISel::tryLoad(SDNode *N) {
if (PlainLoad && PlainLoad->isIndexed())
return false;
- const EVT LoadedEVT = LD->getMemoryVT();
- if (!LoadedEVT.isSimple())
- return false;
- const MVT LoadedVT = LoadedEVT.getSimpleVT();
-
// Address Space Setting
const auto CodeAddrSpace = getAddrSpace(LD);
if (canLowerToLDG(*LD, *Subtarget, CodeAddrSpace))
@@ -1111,7 +1106,7 @@ bool NVPTXDAGToDAGISel::tryLoad(SDNode *N) {
SDValue Chain = N->getOperand(0);
const auto [Ordering, Scope] = insertMemoryInstructionFence(DL, Chain, LD);
- const unsigned FromTypeWidth = LoadedVT.getSizeInBits();
+ const unsigned FromTypeWidth = LD->getMemoryVT().getSizeInBits();
// Vector Setting
const unsigned FromType =
@@ -1165,9 +1160,6 @@ static unsigned getStoreVectorNumElts(SDNode *N) {
bool NVPTXDAGToDAGISel::tryLoadVector(SDNode *N) {
MemSDNode *LD = cast<MemSDNode>(N);
- const EVT MemEVT = LD->getMemoryVT();
- if (!MemEVT.isSimple())
- return false;
// Address Space Setting
const auto CodeAddrSpace = getAddrSpace(LD);
@@ -1237,10 +1229,6 @@ bool NVPTXDAGToDAGISel::tryLoadVector(SDNode *N) {
}
bool NVPTXDAGToDAGISel::tryLDG(MemSDNode *LD) {
- const EVT LoadedEVT = LD->getMemoryVT();
- if (!LoadedEVT.isSimple())
- return false;
-
SDLoc DL(LD);
unsigned ExtensionType;
@@ -1357,10 +1345,6 @@ bool NVPTXDAGToDAGISel::tryStore(SDNode *N) {
if (PlainStore && PlainStore->isIndexed())
return false;
- const EVT StoreVT = ST->getMemoryVT();
- if (!StoreVT.isSimple())
- return false;
-
// Address Space Setting
const auto CodeAddrSpace = getAddrSpace(ST);
@@ -1369,7 +1353,7 @@ bool NVPTXDAGToDAGISel::tryStore(SDNode *N) {
const auto [Ordering, Scope] = insertMemoryInstructionFence(DL, Chain, ST);
// Vector Setting
- const unsigned ToTypeWidth = StoreVT.getSimpleVT().getSizeInBits();
+ const unsigned ToTypeWidth = ST->getMemoryVT().getSizeInBits();
// Create the machine instruction DAG
SDValue Value = PlainStore ? PlainStore->getValue() : AtomicStore->getVal();
@@ -1406,8 +1390,7 @@ bool NVPTXDAGToDAGISel::tryStore(SDNode *N) {
bool NVPTXDAGToDAGISel::tryStoreVector(SDNode *N) {
MemSDNode *ST = cast<MemSDNode>(N);
- const EVT StoreVT = ST->getMemoryVT();
- assert(StoreVT.isSimple() && "Store value is not simple");
+ const unsigned TotalWidth = ST->getMemoryVT().getSizeInBits();
// Address Space Setting
const auto CodeAddrSpace = getAddrSpace(ST);
@@ -1420,10 +1403,6 @@ bool NVPTXDAGToDAGISel::tryStoreVector(SDNode *N) {
SDValue Chain = ST->getChain();
const auto [Ordering, Scope] = insertMemoryInstructionFence(DL, Chain, ST);
- // Type Setting: toType + toTypeWidth
- // - for integer type, always use 'u'
- const unsigned TotalWidth = StoreVT.getSimpleVT().getSizeInBits();
-
const unsigned NumElts = getStoreVectorNumElts(ST);
SmallVector<SDValue, 16> Ops;
diff --git a/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp b/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
index 997c33f1f6a76..8dabd9ba490db 100644
--- a/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
+++ b/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
@@ -198,6 +198,12 @@ static bool IsPTXVectorType(MVT VT) {
static std::optional<std::pair<unsigned int, MVT>>
getVectorLoweringShape(EVT VectorEVT, const NVPTXSubtarget &STI,
unsigned AddressSpace) {
+ const bool CanLowerTo256Bit = STI.has256BitVectorLoadStore(AddressSpace);
+
+ if (CanLowerTo256Bit && VectorEVT.isScalarInteger() &&
+ VectorEVT.getSizeInBits() == 256)
+ return {{4, MVT::i64}};
+
if (!VectorEVT.isSimple())
return std::nullopt;
const MVT VectorVT = VectorEVT.getSimpleVT();
@@ -214,8 +220,6 @@ getVectorLoweringShape(EVT VectorEVT, const NVPTXSubtarget &STI,
// The size of the PTX virtual register that holds a packed type.
unsigned PackRegSize;
- bool CanLowerTo256Bit = STI.has256BitVectorLoadStore(AddressSpace);
-
// We only handle "native" vector sizes for now, e.g. <4 x double> is not
// legal. We can (and should) split that into 2 stores of <2 x double> here
// but I'm leaving that as a TODO for now.
@@ -3085,9 +3089,114 @@ SDValue NVPTXTargetLowering::LowerVASTART(SDValue Op, SelectionDAG &DAG) const {
MachinePointerInfo(SV));
}
+/// ReplaceVectorLoad - Convert vector loads into multi-output scalar loads.
+static std::optional<std::pair<SDValue, SDValue>>
+replaceLoadVector(SDNode *N, SelectionDAG &DAG, const NVPTXSubtarget &STI) {
+ LoadSDNode *LD = cast<LoadSDNode>(N);
+ const EVT ResVT = LD->getValueType(0);
+ const EVT MemVT = LD->getMemoryVT();
+
+ // If we're doing sign/zero extension as part of the load, avoid lowering to
+ // a LoadV node. TODO: consider relaxing this restriction.
+ if (ResVT != MemVT)
+ return std::nullopt;
+
+ const auto NumEltsAndEltVT =
+ getVectorLoweringShape(ResVT, STI, LD->getAddressSpace());
+ if (!NumEltsAndEltVT)
+ return std::nullopt;
+ const auto [NumElts, EltVT] = NumEltsAndEltVT.value();
+
+ Align Alignment = LD->getAlign();
+ const auto &TD = DAG.getDataLayout();
+ Align PrefAlign = TD.getPrefTypeAlign(MemVT.getTypeForEVT(*DAG.getContext()));
+ if (Alignment < PrefAlign) {
+ // This load is not sufficiently aligned, so bail out and let this vector
+ // load be scalarized. Note that we may still be able to emit smaller
+ // vector loads. For example, if we are loading a <4 x float> with an
+ // alignment of 8, this check will fail but the legalizer will try again
+ // with 2 x <2 x float>, which will succeed with an alignment of 8.
+ return std::nullopt;
+ }
+
+ // Since LoadV2 is a target node, we cannot rely on DAG type legalization.
+ // Therefore, we must ensure the type is legal. For i1 and i8, we set the
+ // loaded type to i16 and propagate the "real" type as the memory type.
+ const MVT LoadEltVT = (EltVT.getSizeInBits() < 16) ? MVT::i16 : EltVT;
+
+ unsigned Opcode;
+ switch (NumElts) {
+ default:
+ return std::nullopt;
+ case 2:
+ Opcode = NVPTXISD::LoadV2;
+ break;
+ case 4:
+ Opcode = NVPTXISD::LoadV4;
+ break;
+ case 8:
+ Opcode = NVPTXISD::LoadV8;
+ break;
+ }
+ auto ListVTs = SmallVector<EVT, 9>(NumElts, LoadEltVT);
+ ListVTs.push_back(MVT::Other);
+ SDVTList LdResVTs = DAG.getVTList(ListVTs);
+
+ SDLoc DL(LD);
+
+ // Copy regular operands
+ SmallVector<SDValue, 8> OtherOps(LD->ops());
+
+ // The select routine does not have access to the LoadSDNode instance, so
+ // pass along the extension information
+ OtherOps.push_back(DAG.getIntPtrConstant(LD->getExtensionType(), DL));
+
+ SDValue NewLD = DAG.getMemIntrinsicNode(Opcode, DL, LdResVTs, OtherOps, MemVT,
+ LD->getMemOperand());
+
+ SmallVector<SDValue> ScalarRes;
+ if (EltVT.isVector()) {
+ assert(EVT(EltVT.getVectorElementType()) == ResVT.getVectorElementType());
+ assert(NumElts * EltVT.getVectorNumElements() ==
+ ResVT.getVectorNumElements());
+ // Generate EXTRACT_VECTOR_ELTs to split v2[i,f,bf]16/v4i8 subvectors back
+ // into individual elements.
+ for (const unsigned I : llvm::seq(NumElts)) {
+ SDValue SubVector = NewLD.getValue(I);
+ DAG.ExtractVectorElements(SubVector, ScalarRes);
+ }
+ } else {
+ for (const unsigned I : llvm::seq(NumElts)) {
+ SDValue Res = NewLD.getValue(I);
+ if (LoadEltVT != EltVT)
+ Res = DAG.getNode(ISD::TRUNCATE, DL, EltVT, Res);
+ ScalarRes.push_back(Res);
+ }
+ }
+
+ SDValue LoadChain = NewLD.getValue(NumElts);
+
+ const MVT BuildVecVT =
+ MVT::getVectorVT(EltVT.getScalarType(), ScalarRes.size());
+ SDValue BuildVec = DAG.getBuildVector(BuildVecVT, DL, ScalarRes);
+ SDValue LoadValue = DAG.getBitcast(ResVT, BuildVec);
+
+ return {{LoadValue, LoadChain}};
+}
+
static void replaceLoadVector(SDNode *N, SelectionDAG &DAG,
SmallVectorImpl<SDValue> &Results,
- const NVPTXSubtarget &STI);
+ const NVPTXSubtarget &STI) {
+ if (auto Res = replaceLoadVector(N, DAG, STI))
+ Results.append({Res->first, Res->second});
+}
+
+static SDValue lowerLoadVector(SDNode *N, SelectionDAG &DAG,
+ const NVPTXSubtarget &STI) {
+ if (auto Res = replaceLoadVector(N, DAG, STI))
+ return DAG.getMergeValues({Res->first, Res->second}, SDLoc(N));
+ return SDValue();
+}
// v = ld i1* addr
// =>
@@ -3128,20 +3237,8 @@ SDValue NVPTXTargetLowering::LowerLOAD(SDValue Op, SelectionDAG &DAG) const {
llvm_unreachable("Unexpected custom lowering for load");
}
-SDValue NVPTXTargetLowering::LowerSTORE(SDValue Op, SelectionDAG &DAG) const {
- StoreSDNode *Store = cast<StoreSDNode>(Op);
- EVT VT = Store->getMemoryVT();
-
- if (VT == MVT::i1)
- return LowerSTOREi1(Op, DAG);
-
- // Lower store of any other vector type, including v2f32 as we want to break
- // it apart since this is not a widely-supported type.
- return LowerSTOREVector(Op, DAG);
-}
-
-SDValue
-NVPTXTargetLowering::LowerSTOREVector(SDValue Op, SelectionDAG &DAG) const {
+static SDValue lowerSTOREVector(SDValue Op, SelectionDAG &DAG,
+ const NVPTXSubtarget &STI) {
MemSDNode *N = cast<MemSDNode>(Op.getNode());
SDValue Val = N->getOperand(1);
SDLoc DL(N);
@@ -3233,6 +3330,18 @@ NVPTXTargetLowering::LowerSTOREVector(SDValue Op, SelectionDAG &DAG) const {
return NewSt;
}
+SDValue NVPTXTargetLowering::LowerSTORE(SDValue Op, SelectionDAG &DAG) const {
+ StoreSDNode *Store = cast<StoreSDNode>(Op);
+ EVT VT = Store->getMemoryVT();
+
+ if (VT == MVT::i1)
+ return LowerSTOREi1(Op, DAG);
+
+ // Lower store of any other vector type, including v2f32 as we want to break
+ // it apart since this is not a widely-supported type.
+ return LowerSTOREVector(Op, DAG);
+}
+
// st i1 v, addr
// =>
// v1 = zxt v to i16
@@ -5126,11 +5235,34 @@ static SDValue combinePackingMovIntoStore(SDNode *N,
ST->getMemoryVT(), ST->getMemOperand());
}
-static SDValue PerformStoreCombine(SDNode *N,
- TargetLowering::DAGCombinerInfo &DCI) {
+static SDValue combineSTORE(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
+ const NVPTXSubtarget &STI) {
+
+ if (DCI.isBeforeLegalize() && N->getOpcode() == ISD::STORE) {
+ // Here is our chance to custom lower a store with a non-simple type.
+ // Unfortunately, we can't do this in the legalizer because there is no
+ // way to setOperationAction for an non-simple type.
+ StoreSDNode *ST = cast<StoreSDNode>(N);
+ if (!ST->getValue().getValueType().isSimple())
+ return lowerSTOREVector(SDValue(ST, 0), DCI.DAG, STI);
+ }
+
return combinePackingMovIntoStore(N, DCI, 1, 2);
}
+static SDValue combineLOAD(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
+ const NVPTXSubtarget &STI) {
+ if (DCI.isBeforeLegalize() && N->getOpcode() == ISD::LOAD) {
+ // Here is our chance to custom lower a load with a non-simple type.
+ // Unfortunately, we can't do this in the legalizer because there is no
+ // way to setOperationAction for an non-simple type.
+ if (!N->getValueType(0).isSimple())
+ return lowerLoadVector(N, DCI.DAG, STI);
+ }
+
+ return combineUnpackingMovIntoLoad(N, DCI);
+}
+
/// PerformADDCombine - Target-specific dag combine xforms for ISD::ADD.
///
static SDValue PerformADDCombine(SDNode *N,
@@ -5858,7 +5990,7 @@ SDValue NVPTXTargetLowering::PerformDAGCombine(SDNode *N,
case ISD::LOAD:
case NVPTXISD::LoadV2:
case NVPTXISD::LoadV4:
- return combineUnpackingMovIntoLoad(N, DCI);
+ return combineLOAD(N, DCI, STI);
case ISD::MUL:
return PerformMULCombine(N, DCI, OptLevel);
case NVPTXISD::PRMT:
@@ -5875,7 +6007,7 @@ SDValue NVPTXTargetLowering::PerformDAGCombine(SDNode *N,
case ISD::STORE:
case NVPTXISD::StoreV2:
case NVPTXISD::StoreV4:
- return PerformStoreCombine(N, DCI);
+ return combineSTORE(N, DCI, STI);
case ISD::VSELECT:
return PerformVSELECTCombine(N, DCI);
}
@@ -5904,103 +6036,6 @@ static void ReplaceBITCAST(SDNode *Node, SelectionDAG &DAG,
DAG.getNode(ISD::BUILD_VECTOR, DL, MVT::v2i8, {Vec0, Vec1}));
}
-/// ReplaceVectorLoad - Convert vector loads into multi-output scalar loads.
-static void replaceLoadVector(SDNode *N, SelectionDAG &DAG,
- SmallVectorImpl<SDValue> &Results,
- const NVPTXSubtarget &STI) {
- LoadSDNode *LD = cast<LoadSDNode>(N);
- const EVT ResVT = LD->getValueType(0);
- const EVT MemVT = LD->getMemoryVT();
-
- // If we're doing sign/zero extension as part of the load, avoid lowering to
- // a LoadV node. TODO: consider relaxing this restriction.
- if (ResVT != MemVT)
- return;
-
- const auto NumEltsAndEltVT =
- getVectorLoweringShape(ResVT, STI, LD->getAddressSpace());
- if (!NumEltsAndEltVT)
- return;
- const auto [NumElts, EltVT] = NumEltsAndEltVT.value();
-
- Align Alignment = LD->getAlign();
- const auto &TD = DAG.getDataLayout();
- Align PrefAlign = TD.getPrefTypeAlign(MemVT.getTypeForEVT(*DAG.getContext()));
- if (Alignment < PrefAlign) {
- // This load is not sufficiently aligned, so bail out and let this vector
- // load be scalarized. Note that we may still be able to emit smaller
- // vector loads. For example, if we are loading a <4 x float> with an
- // alignment of 8, this check will fail but the legalizer will try again
- // with 2 x <2 x float>, which will succeed with an alignment of 8.
- return;
- }
-
- // Since LoadV2 is a target node, we cannot rely on DAG type legalization.
- // Therefore, we must ensure the type is legal. For i1 and i8, we set the
- // loaded type to i16 and propagate the "real" type as the memory type.
- const MVT LoadEltVT = (EltVT.getSizeInBits() < 16) ? MVT::i16 : EltVT;
-
- unsigned Opcode;
- switch (NumElts) {
- default:
- return;
- case 2:
- Opcode = NVPTXISD::LoadV2;
- break;
- case 4:
- Opcode = NVPTXISD::LoadV4;
- break;
- case 8:
- Opcode = NVPTXISD::LoadV8;
- break;
- }
- auto ListVTs = SmallVector<EVT, 9>(NumElts, LoadEltVT);
- ListVTs.push_back(MVT::Other);
- SDVTList LdResVTs = DAG.getVTList(ListVTs);
-
- SDLoc DL(LD);
-
- // Copy regular operands
- SmallVector<SDValue, 8> OtherOps(LD->ops());
-
- // The select routine does not have access to the LoadSDNode instance, so
- // pass along the extension information
- OtherOps.push_back(DAG.getIntPtrConstant(LD->getExtensionType(), DL));
-
- SDValue NewLD = DAG.getMemIntrinsicNode(Opcode, DL, LdResVTs, OtherOps,
- LD->getMemoryVT(),
- LD->getMemOperand());
-
- SmallVector<SDValue> ScalarRes;
- if (EltVT.isVector()) {
- assert(EVT(EltVT.getVectorElementType()) == ResVT.getVectorElementType());
- assert(NumElts * EltVT.getVectorNumElements() ==
- ResVT.getVectorNumElements());
- // Generate EXTRACT_VECTOR_ELTs to split v2[i,f,bf]16/v4i8 subvectors back
- // into individual elements.
- for (const unsigned I : llvm::seq(NumElts)) {
- SDValue SubVector = NewLD.getValue(I);
- DAG.ExtractVectorElements(SubVector, ScalarRes);
- }
- } else {
- for (const unsigned I : llvm::seq(NumElts)) {
- SDValue Res = NewLD.getValue(I);
- if (LoadEltVT != EltVT)
- Res = DAG.getNode(ISD::TRUNCATE, DL, EltVT, Res);
- ScalarRes.push_back(Res);
- }
- }
-
- SDValue LoadChain = NewLD.getValue(NumElts);
-
- const MVT BuildVecVT =
- MVT::getVectorVT(EltVT.getScalarType(), ScalarRes.size());
- SDValue BuildVec = DAG.getBuildVector(BuildVecVT, DL, ScalarRes);
- SDValue LoadValue = DAG.getBitcast(ResVT, BuildVec);
-
- Results.append({LoadValue, LoadChain});
-}
-
// Lower vector return type of tcgen05.ld intrinsics
static void ReplaceTcgen05Ld(SDNode *N, SelectionDAG &DAG,
SmallVectorImpl<SDValue> &Results,
diff --git a/llvm/lib/Target/NVPTX/NVPTXISelLowering.h b/llvm/lib/Target/NVPTX/NVPTXISelLowering.h
index e7f1a4b4c98c4..5d11932c52198 100644
--- a/llvm/lib/Target/NVPTX/NVPTXISelLowering.h
+++ b/llvm/lib/Target/NVPTX/NVPTXISelLowering.h
@@ -311,7 +311,6 @@ class NVPTXTargetLowering : public TargetLowering {
SDValue LowerLOAD(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerSTORE(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerSTOREi1(SDValue Op, SelectionDAG &DAG) const;
- SDValue LowerSTOREVector(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerShiftRightParts(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerShiftLeftParts(SDValue Op, SelectionDAG &DAG) const;
diff --git a/llvm/lib/Target/NVPTX/NVPTXTargetMachine.cpp b/llvm/lib/Target/NVPTX/NVPTXTargetMachine.cpp
index 0603994606d71..833f014a4c870 100644
--- a/llvm/lib/Target/NVPTX/NVPTXTargetMachine.cpp
+++ b/llvm/lib/Target/NVPTX/NVPTXTargetMachine.cpp
@@ -126,12 +126,12 @@ static std::string computeDataLayout(bool is64Bit, bool UseShortPointers) {
// (addrspace:3).
if (!is64Bit)
Ret += "-p:32:32-p6:32:32-p7:32:32";
- else if (UseShortPointers) {
+ else if (UseShortPointers)
Ret += "-p3:32:32-p4:32:32-p5:32:32-p6:32:32-p7:32:32";
- } else
+ else
Ret += "-p6:32:32";
- Ret += "-i64:64-i128:128-v16:16-v32:32-n16:32:64";
+ Ret += "-i64:64-i128:128-i256:256-v16:16-v32:32-n16:32:64";
return Ret;
}
diff --git a/llvm/test/CodeGen/NVPTX/load-store-vectors-256.ll b/llvm/test/CodeGen/NVPTX/load-store-vectors-256.ll
index a846607d816c5..60dd5d9308d2a 100644
--- a/llvm/test/CodeGen/NVPTX/load-store-vectors-256.ll
+++ b/llvm/test/CodeGen/NVPTX/load-store-vectors-256.ll
@@ -1506,3 +1506,69 @@ define void @local_volatile_4xdouble(ptr addrspace(5) %a, ptr addrspace(5) %b) {
store volatile <4 x double> %a.load, ptr addrspace(5) %b
ret void
}
+
+define void @test_i256_global(ptr addrspace(1) %a, ptr addrspace(1) %b) {
+; SM90-LABEL: test_i256_global(
+; SM90: {
+; SM90-NEXT: .reg .b64 %rd<7>;
+; SM90-EMPTY:
+; SM90-NEXT: // %bb.0:
+; SM90-NEXT: ld.param.b64 %rd1, [test_i256_global_param_0];
+; SM90-NEXT: ld.global.v2.b64 {%rd2, %rd3}, [%rd1];
+; SM90-NEXT: ld.global.v2.b64 {%rd4, %rd5}, [%rd1+16];
+; SM90-NEXT: ld.param.b64 %rd6, [test_i256_global_param_1];
+; SM90-NEXT: st.global.v2.b64 [%rd6+16], {%rd4, %rd5};
+; SM90-NEXT: st.global.v2.b64 [%rd6], {%rd2, %rd3};
+; SM90-NEXT: ret;
+;
+; SM100-LABEL: test_i256_global(
+; SM100: {
+; SM100-NEXT: .reg .b64 %rd<7>;
+; SM100-EMPTY:
+; SM100-NEXT: // %bb.0:
+; SM100-NEXT: ld.param.b64 %rd1, [test_i256_global_param_0];
+; SM100-NEXT: ld.global.v4.b64 {%rd2, %rd3, %rd4, %rd5}, [%rd1];
+; SM100-NEXT: ld.param.b64 %rd6, [test_i256_global_param_1];
+; SM100-NEXT: st.global.v4.b64 [%rd6], {%rd2, %rd3, %rd4, %rd5};
+; SM100-NEXT: ret;
+ %a.load = load i256, ptr addrspace(1) %a, align 32
+ store i256 %a.load, ptr addrspace(1) %b, align 32
+ ret void
+}
+
+
+define void @test_i256_global_unaligned(ptr addrspace(1) %a, ptr addrspace(1) %b) {
+; CHECK-LABEL: test_i256_global_unaligned(
+; CHECK: {
+; CHECK-NEXT: .reg .b64 %rd<7>;
+; CHECK-EMPTY:
+; CHECK-NEXT: // %bb.0:
+; CHECK-NEXT: ld.param.b64 %rd1, [test_i256_global_unaligned_param_0];
+; CHECK-NEXT: ld.global.v2.b64 {%rd2, %rd3}, [%rd1];
+; CHECK-NEXT: ld.global.v2.b64 {%rd4, %rd5}, [%rd1+16];
+; CHECK-NEXT: ld.param.b64 %rd6, [test_i256_global_unaligned_param_1];
+; CHECK-NEXT: st.global.v2.b64 [%rd6+16], {%rd4, %rd5};
+; CHECK-NEXT: st.global.v2.b64 [%rd6], {%rd2, %rd3};
+; CHECK-NEXT: ret;
+ %a.load = load i256, ptr addrspace(1) %a, align 16
+ store i256 %a.load, ptr addrspace(1) %b, align 16
+ ret void
+}
+
+define void @test_i256_generic(ptr %a, ptr %b) {
+; CHECK-LABEL: test_i256_generic(
+; CHECK: {
+; CHECK-NEXT: .reg .b64 %rd<7>;
+; CHECK-EMPTY:
+; CHECK-NEXT: // %bb.0:
+; CHECK-NEXT: ld.param.b64 %rd1, [test_i256_generic_param_0];
+; CHECK-NEXT: ld.v2.b64 {%rd2, %rd3}, [%rd1];
+; CHECK-NEXT: ld.v2.b64 {%rd4, %rd5}, [%rd1+16];
+; CHECK-NEXT: ld.param.b64 %rd6, [test_i256_generic_param_1];
+; CHECK-NEXT: st.v2.b64 [%rd6+16], {%rd4, %rd5};
+; CHECK-NEXT: st.v2.b64 [%rd6], {%rd2, %rd3};
+; CHECK-NEXT: ret;
+ %a.load = load i256, ptr %a, align 32
+ store i256 %a.load, ptr %b, align 32
+ ret void
+}
>From bd18e21281a960bdb47bd44da6dbd058fb266c7b Mon Sep 17 00:00:00 2001
From: Alex Maclean <amaclean at nvidia.com>
Date: Mon, 25 Aug 2025 17:36:19 +0000
Subject: [PATCH 2/2] address comments - tests
---
llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp | 2 +-
.../CodeGen/NVPTX/load-store-atomic.err.ll | 20 +++++++++++++
.../CodeGen/NVPTX/load-store-vectors-256.ll | 29 +++++++++++++++++++
3 files changed, 50 insertions(+), 1 deletion(-)
create mode 100644 llvm/test/CodeGen/NVPTX/load-store-atomic.err.ll
diff --git a/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp b/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
index 8dabd9ba490db..4fb3f68f89995 100644
--- a/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
+++ b/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
@@ -3339,7 +3339,7 @@ SDValue NVPTXTargetLowering::LowerSTORE(SDValue Op, SelectionDAG &DAG) const {
// Lower store of any other vector type, including v2f32 as we want to break
// it apart since this is not a widely-supported type.
- return LowerSTOREVector(Op, DAG);
+ return lowerSTOREVector(Op, DAG, STI);
}
// st i1 v, addr
diff --git a/llvm/test/CodeGen/NVPTX/load-store-atomic.err.ll b/llvm/test/CodeGen/NVPTX/load-store-atomic.err.ll
new file mode 100644
index 0000000000000..a295356d44fab
--- /dev/null
+++ b/llvm/test/CodeGen/NVPTX/load-store-atomic.err.ll
@@ -0,0 +1,20 @@
+; RUN: not llc < %s -march=nvptx64 -mcpu=sm_100 -mattr=+ptx88 2>&1 | FileCheck %s
+
+; CHECK: error: unsupported atomic store
+; CHECK: error: unsupported atomic load
+; CHECK: error: unsupported atomic store
+; CHECK: error: unsupported atomic load
+
+;; TODO: we could actually support this but we don't currently support b128
+;; load lowering.
+define void @test_i128_generic_atomic(ptr %a, ptr %b) {
+ %a.load = load atomic i128, ptr %a seq_cst, align 16
+ store atomic i128 %a.load, ptr %b seq_cst, align 16
+ ret void
+}
+
+define void @test_i256_global_atomic(ptr addrspace(1) %a, ptr addrspace(1) %b) {
+ %a.load = load atomic i256, ptr addrspace(1) %a seq_cst, align 32
+ store atomic i256 %a.load, ptr addrspace(1) %b seq_cst, align 32
+ ret void
+}
diff --git a/llvm/test/CodeGen/NVPTX/load-store-vectors-256.ll b/llvm/test/CodeGen/NVPTX/load-store-vectors-256.ll
index 60dd5d9308d2a..9f61ded03cdfa 100644
--- a/llvm/test/CodeGen/NVPTX/load-store-vectors-256.ll
+++ b/llvm/test/CodeGen/NVPTX/load-store-vectors-256.ll
@@ -1572,3 +1572,32 @@ define void @test_i256_generic(ptr %a, ptr %b) {
store i256 %a.load, ptr %b, align 32
ret void
}
+
+define void @test_i256_global_volatile(ptr addrspace(1) %a, ptr addrspace(1) %b) {
+; SM90-LABEL: test_i256_global_volatile(
+; SM90: {
+; SM90-NEXT: .reg .b64 %rd<7>;
+; SM90-EMPTY:
+; SM90-NEXT: // %bb.0:
+; SM90-NEXT: ld.param.b64 %rd1, [test_i256_global_volatile_param_0];
+; SM90-NEXT: ld.volatile.global.v2.b64 {%rd2, %rd3}, [%rd1];
+; SM90-NEXT: ld.volatile.global.v2.b64 {%rd4, %rd5}, [%rd1+16];
+; SM90-NEXT: ld.param.b64 %rd6, [test_i256_global_volatile_param_1];
+; SM90-NEXT: st.volatile.global.v2.b64 [%rd6+16], {%rd4, %rd5};
+; SM90-NEXT: st.volatile.global.v2.b64 [%rd6], {%rd2, %rd3};
+; SM90-NEXT: ret;
+;
+; SM100-LABEL: test_i256_global_volatile(
+; SM100: {
+; SM100-NEXT: .reg .b64 %rd<7>;
+; SM100-EMPTY:
+; SM100-NEXT: // %bb.0:
+; SM100-NEXT: ld.param.b64 %rd1, [test_i256_global_volatile_param_0];
+; SM100-NEXT: ld.volatile.global.v4.b64 {%rd2, %rd3, %rd4, %rd5}, [%rd1];
+; SM100-NEXT: ld.param.b64 %rd6, [test_i256_global_volatile_param_1];
+; SM100-NEXT: st.volatile.global.v4.b64 [%rd6], {%rd2, %rd3, %rd4, %rd5};
+; SM100-NEXT: ret;
+ %a.load = load volatile i256, ptr addrspace(1) %a, align 32
+ store volatile i256 %a.load, ptr addrspace(1) %b, align 32
+ ret void
+}
More information about the llvm-commits
mailing list