[mlir][tosa] Add row_gather_block_scaled op#192272
Merged
Merged
Conversation
Member
|
@llvm/pr-subscribers-mlir-tosa Author: Peng Sun (psunn) ChangesAdd This includes:
NotesThe op supports both spec-defined forms:
This also tightens existing block-scaled-only ops to reject Patch is 36.87 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/192272.diff 14 Files Affected:
diff --git a/mlir/include/mlir/Dialect/Tosa/IR/TosaComplianceData.h.inc b/mlir/include/mlir/Dialect/Tosa/IR/TosaComplianceData.h.inc
index d3e2cd129028e..50bb9f69c6242 100644
--- a/mlir/include/mlir/Dialect/Tosa/IR/TosaComplianceData.h.inc
+++ b/mlir/include/mlir/Dialect/Tosa/IR/TosaComplianceData.h.inc
@@ -396,6 +396,25 @@ profileComplianceMap = {
{{i16T, i64T, i16T}, SpecificationVersion::V_1_1_DRAFT},
{{i32T, i64T, i32T}, SpecificationVersion::V_1_1_DRAFT}},
anyOf}}},
+ {"tosa.row_gather_block_scaled",
+ {{{Profile::pro_int},
+ {{{i8T, i32T, i32T, i8T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i16T, i32T, i32T, i16T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i32T, i32T, i32T, i32T}, SpecificationVersion::V_1_1_DRAFT}}},
+ {{Profile::pro_fp},
+ {{{i8T, i32T, i32T, i8T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i16T, i32T, i32T, i16T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i32T, i32T, i32T, i32T}, SpecificationVersion::V_1_1_DRAFT},
+ {{fp16T, i32T, i32T, fp16T}, SpecificationVersion::V_1_1_DRAFT},
+ {{fp32T, i32T, i32T, fp32T}, SpecificationVersion::V_1_1_DRAFT}}},
+ {{Profile::pro_fp, Profile::pro_int},
+ {{{boolT, i32T, i32T, boolT}, SpecificationVersion::V_1_1_DRAFT},
+ {{i8T, i64T, i32T, i8T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i16T, i64T, i32T, i16T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i32T, i64T, i32T, i32T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i64T, i64T, i32T, i64T}, SpecificationVersion::V_1_1_DRAFT},
+ {{boolT, i64T, i32T, boolT}, SpecificationVersion::V_1_1_DRAFT}},
+ anyOf}}},
{"tosa.scatter",
{{{Profile::pro_int},
{{{i8T, i32T, i8T, i8T}, SpecificationVersion::V_1_0},
@@ -890,6 +909,57 @@ extensionComplianceMap = {
{{Extension::bf16, Extension::int64},
{{{bf16T, i64T, bf16T}, SpecificationVersion::V_1_1_DRAFT}},
allOf}}},
+ {"tosa.row_gather_block_scaled",
+ {{{Extension::fp8e4m3},
+ {{{fp8e4m3T, i32T, i32T, fp8e4m3T}, SpecificationVersion::V_1_1_DRAFT}}},
+ {{Extension::fp8e5m2},
+ {{{fp8e5m2T, i32T, i32T, fp8e5m2T}, SpecificationVersion::V_1_1_DRAFT}}},
+ {{Extension::bf16},
+ {{{bf16T, i32T, i32T, bf16T}, SpecificationVersion::V_1_1_DRAFT}}},
+ {{Extension::int64},
+ {{{i8T, i64T, i32T, i8T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i16T, i64T, i32T, i16T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i32T, i64T, i32T, i32T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i64T, i64T, i32T, i64T}, SpecificationVersion::V_1_1_DRAFT},
+ {{fp16T, i64T, i32T, fp16T}, SpecificationVersion::V_1_1_DRAFT},
+ {{fp32T, i64T, i32T, fp32T}, SpecificationVersion::V_1_1_DRAFT},
+ {{boolT, i64T, i32T, boolT}, SpecificationVersion::V_1_1_DRAFT}}},
+ {{Extension::fp8e4m3, Extension::int64},
+ {{{fp8e4m3T, i64T, i32T, fp8e4m3T}, SpecificationVersion::V_1_1_DRAFT}},
+ allOf},
+ {{Extension::fp8e5m2, Extension::int64},
+ {{{fp8e5m2T, i64T, i32T, fp8e5m2T}, SpecificationVersion::V_1_1_DRAFT}},
+ allOf},
+ {{Extension::bf16, Extension::int64},
+ {{{bf16T, i64T, i32T, bf16T}, SpecificationVersion::V_1_1_DRAFT}},
+ allOf},
+ {{Extension::mxfp},
+ {{{fp4e2m1T, fp8ue8m0T, i32T, i32T, fp4e2m1T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp6e2m3T, fp8ue8m0T, i32T, i32T, fp6e2m3T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp6e3m2T, fp8ue8m0T, i32T, i32T, fp6e3m2T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp8e4m3T, fp8ue8m0T, i32T, i32T, fp8e4m3T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp8e5m2T, fp8ue8m0T, i32T, i32T, fp8e5m2T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{mxint8T, fp8ue8m0T, i32T, i32T, mxint8T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT}}},
+ {{Extension::mxfp, Extension::int64},
+ {{{fp4e2m1T, fp8ue8m0T, i64T, i32T, fp4e2m1T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp6e2m3T, fp8ue8m0T, i64T, i32T, fp6e2m3T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp6e3m2T, fp8ue8m0T, i64T, i32T, fp6e3m2T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp8e4m3T, fp8ue8m0T, i64T, i32T, fp8e4m3T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp8e5m2T, fp8ue8m0T, i64T, i32T, fp8e5m2T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{mxint8T, fp8ue8m0T, i64T, i32T, mxint8T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT}},
+ allOf}}},
{"tosa.scatter",
{{{Extension::fp8e4m3},
{{{fp8e4m3T, i32T, fp8e4m3T, fp8e4m3T}, SpecificationVersion::V_1_0}}},
diff --git a/mlir/include/mlir/Dialect/Tosa/IR/TosaOpBase.td b/mlir/include/mlir/Dialect/Tosa/IR/TosaOpBase.td
index 1f05aee3e5eec..591073e9985ae 100644
--- a/mlir/include/mlir/Dialect/Tosa/IR/TosaOpBase.td
+++ b/mlir/include/mlir/Dialect/Tosa/IR/TosaOpBase.td
@@ -484,11 +484,12 @@ def Tosa_RoundingModeAttr
: Tosa_I32EnumAttr<"RoundingMode", "Supported rounding modes", "rounding_mode",
[Tosa_ROUNDING_SINGLE_ROUND, Tosa_ROUNDING_INEXACT_ROUND, Tosa_ROUNDING_DOUBLE_ROUND]>;
+def Tosa_BLOCK_SIZE_1 : I32EnumAttrCase<"BLOCK_SIZE_1", 1>;
def Tosa_BLOCK_SIZE_32 : I32EnumAttrCase<"BLOCK_SIZE_32", 32>;
def Tosa_BlockSizeAttr
- : Tosa_I32EnumAttr<"BlockSize", "Block size for the block_scaled formats", "block_size",
- [Tosa_BLOCK_SIZE_32]> {
+ : Tosa_I32EnumAttr<"BlockSize", "Block size for the block_scaled formats",
+ "block_size", [Tosa_BLOCK_SIZE_1, Tosa_BLOCK_SIZE_32]> {
let extraClassDeclaration = [{
static uint32_t getBlockSizeValue(BlockSize blockSize) {
return static_cast<uint32_t>(blockSize);
@@ -496,7 +497,6 @@ def Tosa_BlockSizeAttr
}];
}
-
//===----------------------------------------------------------------------===//
// TOSA Interfaces.
//===----------------------------------------------------------------------===//
diff --git a/mlir/include/mlir/Dialect/Tosa/IR/TosaOps.td b/mlir/include/mlir/Dialect/Tosa/IR/TosaOps.td
index 45d1388a28749..ba750ef4438db 100644
--- a/mlir/include/mlir/Dialect/Tosa/IR/TosaOps.td
+++ b/mlir/include/mlir/Dialect/Tosa/IR/TosaOps.td
@@ -2487,6 +2487,44 @@ def Tosa_GatherOp : Tosa_InferShapedTypeOp<"gather", [NoMemoryEffect]> {
"operands attr-dict `:` functional-type(operands, results)";
}
+//===----------------------------------------------------------------------===//
+// Operator: row_gather_block_scaled
+//===----------------------------------------------------------------------===//
+def Tosa_RowGatherBlockScaledOp
+ : Tosa_InferShapedTypeOp<"row_gather_block_scaled", [NoMemoryEffect]> {
+ let summary =
+ "Row gather operation for block-scaled and non-block-scaled data.";
+
+ let description = [{
+ Generate a tensor-list which contains a data tensor and an optional scale
+ tensor based on the input indices and row_count. The number of consecutive
+ rows gathered for each index is specified in row_count.
+
+ This operation follows the TOSA 1.1 draft specification and may evolve as
+ the specification is updated.
+
+ This operation is not pure. Undefined behaviour may occur if the specified
+ indices are out of range.
+ }];
+
+ let arguments = (ins Variadic<Tosa_Tensor3D>:$values,
+ Tosa_IndexTensor2D:$indices, Tosa_ScalarInt32Tensor:$row_count,
+ Tosa_BlockSizeAttr:$block_size);
+
+ let results = (outs Variadic<Tosa_Tensor3D>:$output);
+
+ list<Availability> availability = [Profile<[Tosa_PRO_INT, Tosa_PRO_FP]>,
+ Extension<[Tosa_EXT_FP8E4M3,
+ Tosa_EXT_FP8E5M2, Tosa_EXT_BF16,
+ Tosa_EXT_MXFP, Tosa_EXT_INT64]>,
+ ];
+
+ let hasVerifier = 1;
+
+ let assemblyFormat =
+ "operands attr-dict `:` functional-type(operands, results)";
+}
+
//===----------------------------------------------------------------------===//
// Operator: scatter
//===----------------------------------------------------------------------===//
diff --git a/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp b/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
index 2754d3b21d4a6..360ab25875bd7 100644
--- a/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
+++ b/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
@@ -2201,6 +2201,8 @@ LogicalResult MatmulTBlockScaledOp::verify() {
// Verify C is a multiple of block size
const uint32_t blockSize = BlockSizeAttr::getBlockSizeValue(getBlockSize());
+ if (blockSize == 1)
+ return emitOpError("requires block_size to not be BLOCK_SIZE_1");
if (ShapedType::isStatic(C) && C % blockSize != 0)
return emitOpError("expect C to be a multiple of block size, got C=")
<< C << ", block_size=" << blockSize;
@@ -2848,6 +2850,18 @@ static FailureOr<int64_t> getZeroPoint(Value val, bool signExtend) {
return -1;
}
+static FailureOr<int64_t> getConstantScalarIntValue(Value val) {
+ ElementsAttr attr;
+ if (!matchPattern(val, m_Constant(&attr)))
+ return failure();
+
+ if (!llvm::isa<IntegerType>(attr.getElementType()) ||
+ attr.getNumElements() != 1)
+ return failure();
+
+ return attr.getValues<APInt>()[0].getSExtValue();
+}
+
template <typename T>
static LogicalResult verifyZeroPoint(T op, Value val, const int64_t &zp,
const std::string &operand) {
@@ -3096,6 +3110,48 @@ LogicalResult tosa::GatherOp::inferReturnTypeComponents(
return success();
}
+LogicalResult tosa::RowGatherBlockScaledOp::inferReturnTypeComponents(
+ MLIRContext *context, ::std::optional<Location> location,
+ RowGatherBlockScaledOp::Adaptor adaptor,
+ SmallVectorImpl<ShapedTypeComponents> &inferredReturnShapes) {
+ const auto values = adaptor.getValues();
+ if (values.empty() || values.size() > 2)
+ return failure();
+
+ SmallVector<int64_t> dataShape(3, ShapedType::kDynamic);
+ const ShapeAdaptor valuesShape(values.front().getType());
+ if (valuesShape.hasRank()) {
+ dataShape[0] = valuesShape.getDimSize(0);
+ dataShape[2] = valuesShape.getDimSize(2);
+ }
+
+ const ShapeAdaptor indicesShape(adaptor.getIndices().getType());
+ if (indicesShape.hasRank()) {
+ if (dataShape[0] == ShapedType::kDynamic)
+ dataShape[0] = indicesShape.getDimSize(0);
+
+ if (auto rowCount = getConstantScalarIntValue(adaptor.getRowCount());
+ succeeded(rowCount) && rowCount.value() > 0) {
+ const int64_t indicesW = indicesShape.getDimSize(1);
+ if (ShapedType::isStatic(indicesW))
+ dataShape[1] = indicesW * rowCount.value();
+ }
+ }
+
+ inferredReturnShapes.push_back(ShapedTypeComponents(dataShape));
+ if (values.size() == 1)
+ return success();
+
+ SmallVector<int64_t> scaleShape = dataShape;
+ const uint32_t blockSize =
+ BlockSizeAttr::getBlockSizeValue(adaptor.getBlockSize());
+ if (ShapedType::isStatic(dataShape[2]))
+ scaleShape[2] = dataShape[2] / blockSize;
+
+ inferredReturnShapes.push_back(ShapedTypeComponents(scaleShape));
+ return success();
+}
+
LogicalResult tosa::GatherOp::verify() {
if (verifySameElementTypes(*this, /* inType = */ getValues().getType(),
/* outType = */ getOutput().getType())
@@ -3145,6 +3201,137 @@ LogicalResult tosa::GatherOp::verify() {
return success();
}
+LogicalResult tosa::RowGatherBlockScaledOp::verify() {
+ const OperandRange values = getValues();
+ const ResultRange output = getOutput();
+ if (values.empty() || values.size() > 2)
+ return emitOpError()
+ << "expects values tensor list length to be 1 or 2, got "
+ << values.size();
+ if (output.size() != values.size())
+ return emitOpError()
+ << "expects output tensor list length to match values tensor list "
+ "length, got "
+ << output.size() << " results for " << values.size()
+ << " input tensors";
+
+ const uint32_t blockSize = BlockSizeAttr::getBlockSizeValue(getBlockSize());
+ if (values.size() == 1 && blockSize != 1)
+ return emitOpError()
+ << "requires block_size to be BLOCK_SIZE_1 when values tensor list "
+ "length is 1";
+ if (values.size() == 2 && blockSize == 1)
+ return emitOpError()
+ << "requires block_size to not be BLOCK_SIZE_1 when values tensor "
+ "list length is 2";
+
+ if (failed(verifySameElementTypes(*this, values[0].getType(),
+ output[0].getType(), "values[0]",
+ "output[0]")))
+ return failure();
+ if (values.size() == 2 && failed(verifySameElementTypes(
+ *this, values[1].getType(), output[1].getType(),
+ "values[1]", "output[1]")))
+ return failure();
+
+ if (auto rowCount = getConstantScalarIntValue(getRowCount());
+ succeeded(rowCount) && rowCount.value() <= 0)
+ return emitOpError() << "requires row_count to be > 0, got "
+ << rowCount.value();
+
+ int64_t n = ShapedType::kDynamic;
+ int64_t k = ShapedType::kDynamic;
+ int64_t c = ShapedType::kDynamic;
+ int64_t w = ShapedType::kDynamic;
+ int64_t multiplesOfC = ShapedType::kDynamic;
+
+ const ShapeAdaptor valuesDataShape(values[0].getType());
+ if (valuesDataShape.hasRank()) {
+ n = valuesDataShape.getDimSize(0);
+ k = valuesDataShape.getDimSize(1);
+ c = valuesDataShape.getDimSize(2);
+ }
+
+ if (ShapedType::isStatic(c) && c % blockSize != 0)
+ return emitOpError() << "expects channels of values[0] (" << c
+ << ") to be divisible by block_size (" << blockSize
+ << ")";
+
+ const ShapeAdaptor indicesShape(getIndices().getType());
+ if (indicesShape.hasRank()) {
+ if (failed(tryUpdateDimOrFailure(*this, n, indicesShape.getDimSize(0),
+ "indices", "batch")))
+ return failure();
+ w = indicesShape.getDimSize(1);
+ }
+
+ const ShapeAdaptor outputDataShape(output[0].getType());
+ if (outputDataShape.hasRank()) {
+ if (failed(tryUpdateDimOrFailure(*this, n, outputDataShape.getDimSize(0),
+ "output[0]", "batch")) ||
+ failed(tryUpdateDimOrFailure(*this, c, outputDataShape.getDimSize(2),
+ "output[0]", "channels")))
+ return failure();
+
+ if (auto rowCount = getConstantScalarIntValue(getRowCount());
+ succeeded(rowCount) && rowCount.value() > 0 &&
+ ShapedType::isStatic(w)) {
+ const int64_t expectedOutputRows = w * rowCount.value();
+ if (ShapedType::isStatic(outputDataShape.getDimSize(1)) &&
+ outputDataShape.getDimSize(1) != expectedOutputRows)
+ return emitOpError() << "requires output[0] dimension 1 to have size "
+ << expectedOutputRows << ", got "
+ << outputDataShape.getDimSize(1);
+ }
+ }
+
+ if (values.size() == 2) {
+ const ShapeAdaptor valuesScaleShape(values[1].getType());
+ if (valuesScaleShape.hasRank()) {
+ if (failed(tryUpdateDimOrFailure(*this, n, valuesScaleShape.getDimSize(0),
+ "values[1]", "batch")) ||
+ failed(tryUpdateDimOrFailure(*this, k, valuesScaleShape.getDimSize(1),
+ "values[1]", "rows")))
+ return failure();
+ multiplesOfC = valuesScaleShape.getDimSize(2);
+ }
+
+ const ShapeAdaptor outputScaleShape(output[1].getType());
+ if (outputScaleShape.hasRank()) {
+ if (failed(tryUpdateDimOrFailure(*this, n, outputScaleShape.getDimSize(0),
+ "output[1]", "batch")))
+ return failure();
+
+ if (auto rowCount = getConstantScalarIntValue(getRowCount());
+ succeeded(rowCount) && rowCount.value() > 0 &&
+ ShapedType::isStatic(w)) {
+ const int64_t expectedOutputRows = w * rowCount.value();
+ if (ShapedType::isStatic(outputScaleShape.getDimSize(1)) &&
+ outputScaleShape.getDimSize(1) != expectedOutputRows)
+ return emitOpError() << "requires output[1] dimension 1 to have size "
+ << expectedOutputRows << ", got "
+ << outputScaleShape.getDimSize(1);
+ }
+
+ if (ShapedType::isDynamic(multiplesOfC))
+ multiplesOfC = outputScaleShape.getDimSize(2);
+ else if (ShapedType::isStatic(outputScaleShape.getDimSize(2)) &&
+ multiplesOfC != outputScaleShape.getDimSize(2))
+ return emitOpError()
+ << "expected channels of output[1] to match size "
+ << multiplesOfC << ", got " << outputScaleShape.getDimSize(2);
+ }
+
+ if (ShapedType::isStatic(c) && ShapedType::isStatic(multiplesOfC) &&
+ multiplesOfC != c / blockSize)
+ return emitOpError()
+ << "expects channels of scale tensors to equal C/block_size (" << c
+ << "/" << blockSize << "), got " << multiplesOfC;
+ }
+
+ return success();
+}
+
LogicalResult tosa::ResizeOp::inferReturnTypeComponents(
MLIRContext *context, ::std::optional<Location> location,
ResizeOp::Adaptor adaptor,
@@ -3987,6 +4174,8 @@ LogicalResult Conv2DBlockScaledOp::verify() {
// Verify IC is a multiple of block size
const uint32_t blockSize = BlockSizeAttr::getBlockSizeValue(getBlockSize());
+ if (blockSize == 1)
+ return emitOpError("requires block_size to not be BLOCK_SIZE_1");
if (ShapedType::isStatic(IC) && IC % blockSize != 0)
return emitOpError("expect IC to be a multiple of block size, got IC=")
<< IC << ", block_size=" << blockSize;
@@ -4577,6 +4766,8 @@ LogicalResult CastFromBlockScaledOp::verify() {
if (inputDataShape.hasRank()) {
const unsigned int blockSize =
BlockSizeAttr::getBlockSizeValue(getBlockSize());
+ if (blockSize == 1)
+ return emitOpError("requires block_size to not be BLOCK_SIZE_1");
const int64_t inputDataLastDim =
inputDataShape.getDimSize(inputDataShape.getRank() - 1);
if (inputDataLastDim % blockSize != 0)
@@ -4650,6 +4841,8 @@ LogicalResult CastToBlockScaledOp::verify() {
const unsigned int blockSize =
BlockSizeAttr::getBlockSizeValue(getBlockSize());
+ if (blockSize == 1)
+ return emitOpError("requires block_size to not be BLOCK_SIZE_1");
const ShapeAdaptor inputDataShape = ShapeAdaptor(inputDataType);
if (inputDataShape.hasRank()) {
const int64_t inputDataLastDim =
diff --git a/mlir/lib/Dialect/Tosa/Transforms/TosaProfileCompliance.cpp b/mlir/lib/Dialect/Tosa/Transforms/TosaProfileCompliance.cpp
index 01c85be4f704f..4ea225b860f6c 100644
--- a/mlir/lib/Dialect/Tosa/Transforms/TosaProfileCompliance.cpp
+++ b/mlir/lib/Dialect/Tosa/Transforms/TosaProfileCompliance.cpp
@@ -185,6 +185,18 @@ LogicalResult ProfileInfoDepot::populateProfileInfo(tosa::GatherOp op) {
return success();
}
+template <>
+LogicalResult
+ProfileInfoDepot::populateProfileInfo(tosa::RowGatherBlockScaledOp op) {
+ for (Value value : op.getValues())
+ addValue(value);
+ addValue(op.getIndices());
+ addValue(op.getRowCount());
+ for (Value result : op.getOutput())
+ addValue(result);
+ return success();
+}
+
template <>
LogicalResult ProfileInfoDepot::populateProfileInfo(tosa::ScatterOp op) {
addValue(op.getValuesIn());
@@ -288,6 +300,7 @@ LogicalResult ProfileInfoDepot::populatationDispatch(Operation *op) {
POPULATE_PROFILE_INFO_CUSTOM(Tile)
POPULATE_PROFILE_INFO_CUSTOM(Transpose)
POPULATE_PROFILE_INFO_CUSTOM(Gather)
+ POPULATE_PROFILE_INFO_CUSTOM(RowGatherBlockScaled)
POPULATE_PROFILE_INFO_CUSTOM(Scatter)
POPULATE_PROFILE_INFO_CUSTOM(Resize)
POPULATE_PROFILE_INFO_CUSTOM(Select)
@@ -598,10 +611,11 @@ SmallVector<OpComplianceInfo<T>> TosaProfileCompliance::findMatchedEntries(
SmallVector<VersionedTypeInfo> sets = compInfo[i].operandTypeInfoSet;
for (const auto &set : sets) {
SmallVector<TypeInfo> expected = set.first;
- assert(present.size() == expected.size() &&
- "the entries for profile-based compliance do not match between...
[truncated]
|
Member
|
@llvm/pr-subscribers-mlir Author: Peng Sun (psunn) ChangesAdd This includes:
NotesThe op supports both spec-defined forms:
This also tightens existing block-scaled-only ops to reject Patch is 36.87 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/192272.diff 14 Files Affected:
diff --git a/mlir/include/mlir/Dialect/Tosa/IR/TosaComplianceData.h.inc b/mlir/include/mlir/Dialect/Tosa/IR/TosaComplianceData.h.inc
index d3e2cd129028e..50bb9f69c6242 100644
--- a/mlir/include/mlir/Dialect/Tosa/IR/TosaComplianceData.h.inc
+++ b/mlir/include/mlir/Dialect/Tosa/IR/TosaComplianceData.h.inc
@@ -396,6 +396,25 @@ profileComplianceMap = {
{{i16T, i64T, i16T}, SpecificationVersion::V_1_1_DRAFT},
{{i32T, i64T, i32T}, SpecificationVersion::V_1_1_DRAFT}},
anyOf}}},
+ {"tosa.row_gather_block_scaled",
+ {{{Profile::pro_int},
+ {{{i8T, i32T, i32T, i8T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i16T, i32T, i32T, i16T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i32T, i32T, i32T, i32T}, SpecificationVersion::V_1_1_DRAFT}}},
+ {{Profile::pro_fp},
+ {{{i8T, i32T, i32T, i8T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i16T, i32T, i32T, i16T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i32T, i32T, i32T, i32T}, SpecificationVersion::V_1_1_DRAFT},
+ {{fp16T, i32T, i32T, fp16T}, SpecificationVersion::V_1_1_DRAFT},
+ {{fp32T, i32T, i32T, fp32T}, SpecificationVersion::V_1_1_DRAFT}}},
+ {{Profile::pro_fp, Profile::pro_int},
+ {{{boolT, i32T, i32T, boolT}, SpecificationVersion::V_1_1_DRAFT},
+ {{i8T, i64T, i32T, i8T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i16T, i64T, i32T, i16T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i32T, i64T, i32T, i32T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i64T, i64T, i32T, i64T}, SpecificationVersion::V_1_1_DRAFT},
+ {{boolT, i64T, i32T, boolT}, SpecificationVersion::V_1_1_DRAFT}},
+ anyOf}}},
{"tosa.scatter",
{{{Profile::pro_int},
{{{i8T, i32T, i8T, i8T}, SpecificationVersion::V_1_0},
@@ -890,6 +909,57 @@ extensionComplianceMap = {
{{Extension::bf16, Extension::int64},
{{{bf16T, i64T, bf16T}, SpecificationVersion::V_1_1_DRAFT}},
allOf}}},
+ {"tosa.row_gather_block_scaled",
+ {{{Extension::fp8e4m3},
+ {{{fp8e4m3T, i32T, i32T, fp8e4m3T}, SpecificationVersion::V_1_1_DRAFT}}},
+ {{Extension::fp8e5m2},
+ {{{fp8e5m2T, i32T, i32T, fp8e5m2T}, SpecificationVersion::V_1_1_DRAFT}}},
+ {{Extension::bf16},
+ {{{bf16T, i32T, i32T, bf16T}, SpecificationVersion::V_1_1_DRAFT}}},
+ {{Extension::int64},
+ {{{i8T, i64T, i32T, i8T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i16T, i64T, i32T, i16T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i32T, i64T, i32T, i32T}, SpecificationVersion::V_1_1_DRAFT},
+ {{i64T, i64T, i32T, i64T}, SpecificationVersion::V_1_1_DRAFT},
+ {{fp16T, i64T, i32T, fp16T}, SpecificationVersion::V_1_1_DRAFT},
+ {{fp32T, i64T, i32T, fp32T}, SpecificationVersion::V_1_1_DRAFT},
+ {{boolT, i64T, i32T, boolT}, SpecificationVersion::V_1_1_DRAFT}}},
+ {{Extension::fp8e4m3, Extension::int64},
+ {{{fp8e4m3T, i64T, i32T, fp8e4m3T}, SpecificationVersion::V_1_1_DRAFT}},
+ allOf},
+ {{Extension::fp8e5m2, Extension::int64},
+ {{{fp8e5m2T, i64T, i32T, fp8e5m2T}, SpecificationVersion::V_1_1_DRAFT}},
+ allOf},
+ {{Extension::bf16, Extension::int64},
+ {{{bf16T, i64T, i32T, bf16T}, SpecificationVersion::V_1_1_DRAFT}},
+ allOf},
+ {{Extension::mxfp},
+ {{{fp4e2m1T, fp8ue8m0T, i32T, i32T, fp4e2m1T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp6e2m3T, fp8ue8m0T, i32T, i32T, fp6e2m3T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp6e3m2T, fp8ue8m0T, i32T, i32T, fp6e3m2T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp8e4m3T, fp8ue8m0T, i32T, i32T, fp8e4m3T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp8e5m2T, fp8ue8m0T, i32T, i32T, fp8e5m2T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{mxint8T, fp8ue8m0T, i32T, i32T, mxint8T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT}}},
+ {{Extension::mxfp, Extension::int64},
+ {{{fp4e2m1T, fp8ue8m0T, i64T, i32T, fp4e2m1T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp6e2m3T, fp8ue8m0T, i64T, i32T, fp6e2m3T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp6e3m2T, fp8ue8m0T, i64T, i32T, fp6e3m2T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp8e4m3T, fp8ue8m0T, i64T, i32T, fp8e4m3T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{fp8e5m2T, fp8ue8m0T, i64T, i32T, fp8e5m2T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT},
+ {{mxint8T, fp8ue8m0T, i64T, i32T, mxint8T, fp8ue8m0T},
+ SpecificationVersion::V_1_1_DRAFT}},
+ allOf}}},
{"tosa.scatter",
{{{Extension::fp8e4m3},
{{{fp8e4m3T, i32T, fp8e4m3T, fp8e4m3T}, SpecificationVersion::V_1_0}}},
diff --git a/mlir/include/mlir/Dialect/Tosa/IR/TosaOpBase.td b/mlir/include/mlir/Dialect/Tosa/IR/TosaOpBase.td
index 1f05aee3e5eec..591073e9985ae 100644
--- a/mlir/include/mlir/Dialect/Tosa/IR/TosaOpBase.td
+++ b/mlir/include/mlir/Dialect/Tosa/IR/TosaOpBase.td
@@ -484,11 +484,12 @@ def Tosa_RoundingModeAttr
: Tosa_I32EnumAttr<"RoundingMode", "Supported rounding modes", "rounding_mode",
[Tosa_ROUNDING_SINGLE_ROUND, Tosa_ROUNDING_INEXACT_ROUND, Tosa_ROUNDING_DOUBLE_ROUND]>;
+def Tosa_BLOCK_SIZE_1 : I32EnumAttrCase<"BLOCK_SIZE_1", 1>;
def Tosa_BLOCK_SIZE_32 : I32EnumAttrCase<"BLOCK_SIZE_32", 32>;
def Tosa_BlockSizeAttr
- : Tosa_I32EnumAttr<"BlockSize", "Block size for the block_scaled formats", "block_size",
- [Tosa_BLOCK_SIZE_32]> {
+ : Tosa_I32EnumAttr<"BlockSize", "Block size for the block_scaled formats",
+ "block_size", [Tosa_BLOCK_SIZE_1, Tosa_BLOCK_SIZE_32]> {
let extraClassDeclaration = [{
static uint32_t getBlockSizeValue(BlockSize blockSize) {
return static_cast<uint32_t>(blockSize);
@@ -496,7 +497,6 @@ def Tosa_BlockSizeAttr
}];
}
-
//===----------------------------------------------------------------------===//
// TOSA Interfaces.
//===----------------------------------------------------------------------===//
diff --git a/mlir/include/mlir/Dialect/Tosa/IR/TosaOps.td b/mlir/include/mlir/Dialect/Tosa/IR/TosaOps.td
index 45d1388a28749..ba750ef4438db 100644
--- a/mlir/include/mlir/Dialect/Tosa/IR/TosaOps.td
+++ b/mlir/include/mlir/Dialect/Tosa/IR/TosaOps.td
@@ -2487,6 +2487,44 @@ def Tosa_GatherOp : Tosa_InferShapedTypeOp<"gather", [NoMemoryEffect]> {
"operands attr-dict `:` functional-type(operands, results)";
}
+//===----------------------------------------------------------------------===//
+// Operator: row_gather_block_scaled
+//===----------------------------------------------------------------------===//
+def Tosa_RowGatherBlockScaledOp
+ : Tosa_InferShapedTypeOp<"row_gather_block_scaled", [NoMemoryEffect]> {
+ let summary =
+ "Row gather operation for block-scaled and non-block-scaled data.";
+
+ let description = [{
+ Generate a tensor-list which contains a data tensor and an optional scale
+ tensor based on the input indices and row_count. The number of consecutive
+ rows gathered for each index is specified in row_count.
+
+ This operation follows the TOSA 1.1 draft specification and may evolve as
+ the specification is updated.
+
+ This operation is not pure. Undefined behaviour may occur if the specified
+ indices are out of range.
+ }];
+
+ let arguments = (ins Variadic<Tosa_Tensor3D>:$values,
+ Tosa_IndexTensor2D:$indices, Tosa_ScalarInt32Tensor:$row_count,
+ Tosa_BlockSizeAttr:$block_size);
+
+ let results = (outs Variadic<Tosa_Tensor3D>:$output);
+
+ list<Availability> availability = [Profile<[Tosa_PRO_INT, Tosa_PRO_FP]>,
+ Extension<[Tosa_EXT_FP8E4M3,
+ Tosa_EXT_FP8E5M2, Tosa_EXT_BF16,
+ Tosa_EXT_MXFP, Tosa_EXT_INT64]>,
+ ];
+
+ let hasVerifier = 1;
+
+ let assemblyFormat =
+ "operands attr-dict `:` functional-type(operands, results)";
+}
+
//===----------------------------------------------------------------------===//
// Operator: scatter
//===----------------------------------------------------------------------===//
diff --git a/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp b/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
index 2754d3b21d4a6..360ab25875bd7 100644
--- a/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
+++ b/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
@@ -2201,6 +2201,8 @@ LogicalResult MatmulTBlockScaledOp::verify() {
// Verify C is a multiple of block size
const uint32_t blockSize = BlockSizeAttr::getBlockSizeValue(getBlockSize());
+ if (blockSize == 1)
+ return emitOpError("requires block_size to not be BLOCK_SIZE_1");
if (ShapedType::isStatic(C) && C % blockSize != 0)
return emitOpError("expect C to be a multiple of block size, got C=")
<< C << ", block_size=" << blockSize;
@@ -2848,6 +2850,18 @@ static FailureOr<int64_t> getZeroPoint(Value val, bool signExtend) {
return -1;
}
+static FailureOr<int64_t> getConstantScalarIntValue(Value val) {
+ ElementsAttr attr;
+ if (!matchPattern(val, m_Constant(&attr)))
+ return failure();
+
+ if (!llvm::isa<IntegerType>(attr.getElementType()) ||
+ attr.getNumElements() != 1)
+ return failure();
+
+ return attr.getValues<APInt>()[0].getSExtValue();
+}
+
template <typename T>
static LogicalResult verifyZeroPoint(T op, Value val, const int64_t &zp,
const std::string &operand) {
@@ -3096,6 +3110,48 @@ LogicalResult tosa::GatherOp::inferReturnTypeComponents(
return success();
}
+LogicalResult tosa::RowGatherBlockScaledOp::inferReturnTypeComponents(
+ MLIRContext *context, ::std::optional<Location> location,
+ RowGatherBlockScaledOp::Adaptor adaptor,
+ SmallVectorImpl<ShapedTypeComponents> &inferredReturnShapes) {
+ const auto values = adaptor.getValues();
+ if (values.empty() || values.size() > 2)
+ return failure();
+
+ SmallVector<int64_t> dataShape(3, ShapedType::kDynamic);
+ const ShapeAdaptor valuesShape(values.front().getType());
+ if (valuesShape.hasRank()) {
+ dataShape[0] = valuesShape.getDimSize(0);
+ dataShape[2] = valuesShape.getDimSize(2);
+ }
+
+ const ShapeAdaptor indicesShape(adaptor.getIndices().getType());
+ if (indicesShape.hasRank()) {
+ if (dataShape[0] == ShapedType::kDynamic)
+ dataShape[0] = indicesShape.getDimSize(0);
+
+ if (auto rowCount = getConstantScalarIntValue(adaptor.getRowCount());
+ succeeded(rowCount) && rowCount.value() > 0) {
+ const int64_t indicesW = indicesShape.getDimSize(1);
+ if (ShapedType::isStatic(indicesW))
+ dataShape[1] = indicesW * rowCount.value();
+ }
+ }
+
+ inferredReturnShapes.push_back(ShapedTypeComponents(dataShape));
+ if (values.size() == 1)
+ return success();
+
+ SmallVector<int64_t> scaleShape = dataShape;
+ const uint32_t blockSize =
+ BlockSizeAttr::getBlockSizeValue(adaptor.getBlockSize());
+ if (ShapedType::isStatic(dataShape[2]))
+ scaleShape[2] = dataShape[2] / blockSize;
+
+ inferredReturnShapes.push_back(ShapedTypeComponents(scaleShape));
+ return success();
+}
+
LogicalResult tosa::GatherOp::verify() {
if (verifySameElementTypes(*this, /* inType = */ getValues().getType(),
/* outType = */ getOutput().getType())
@@ -3145,6 +3201,137 @@ LogicalResult tosa::GatherOp::verify() {
return success();
}
+LogicalResult tosa::RowGatherBlockScaledOp::verify() {
+ const OperandRange values = getValues();
+ const ResultRange output = getOutput();
+ if (values.empty() || values.size() > 2)
+ return emitOpError()
+ << "expects values tensor list length to be 1 or 2, got "
+ << values.size();
+ if (output.size() != values.size())
+ return emitOpError()
+ << "expects output tensor list length to match values tensor list "
+ "length, got "
+ << output.size() << " results for " << values.size()
+ << " input tensors";
+
+ const uint32_t blockSize = BlockSizeAttr::getBlockSizeValue(getBlockSize());
+ if (values.size() == 1 && blockSize != 1)
+ return emitOpError()
+ << "requires block_size to be BLOCK_SIZE_1 when values tensor list "
+ "length is 1";
+ if (values.size() == 2 && blockSize == 1)
+ return emitOpError()
+ << "requires block_size to not be BLOCK_SIZE_1 when values tensor "
+ "list length is 2";
+
+ if (failed(verifySameElementTypes(*this, values[0].getType(),
+ output[0].getType(), "values[0]",
+ "output[0]")))
+ return failure();
+ if (values.size() == 2 && failed(verifySameElementTypes(
+ *this, values[1].getType(), output[1].getType(),
+ "values[1]", "output[1]")))
+ return failure();
+
+ if (auto rowCount = getConstantScalarIntValue(getRowCount());
+ succeeded(rowCount) && rowCount.value() <= 0)
+ return emitOpError() << "requires row_count to be > 0, got "
+ << rowCount.value();
+
+ int64_t n = ShapedType::kDynamic;
+ int64_t k = ShapedType::kDynamic;
+ int64_t c = ShapedType::kDynamic;
+ int64_t w = ShapedType::kDynamic;
+ int64_t multiplesOfC = ShapedType::kDynamic;
+
+ const ShapeAdaptor valuesDataShape(values[0].getType());
+ if (valuesDataShape.hasRank()) {
+ n = valuesDataShape.getDimSize(0);
+ k = valuesDataShape.getDimSize(1);
+ c = valuesDataShape.getDimSize(2);
+ }
+
+ if (ShapedType::isStatic(c) && c % blockSize != 0)
+ return emitOpError() << "expects channels of values[0] (" << c
+ << ") to be divisible by block_size (" << blockSize
+ << ")";
+
+ const ShapeAdaptor indicesShape(getIndices().getType());
+ if (indicesShape.hasRank()) {
+ if (failed(tryUpdateDimOrFailure(*this, n, indicesShape.getDimSize(0),
+ "indices", "batch")))
+ return failure();
+ w = indicesShape.getDimSize(1);
+ }
+
+ const ShapeAdaptor outputDataShape(output[0].getType());
+ if (outputDataShape.hasRank()) {
+ if (failed(tryUpdateDimOrFailure(*this, n, outputDataShape.getDimSize(0),
+ "output[0]", "batch")) ||
+ failed(tryUpdateDimOrFailure(*this, c, outputDataShape.getDimSize(2),
+ "output[0]", "channels")))
+ return failure();
+
+ if (auto rowCount = getConstantScalarIntValue(getRowCount());
+ succeeded(rowCount) && rowCount.value() > 0 &&
+ ShapedType::isStatic(w)) {
+ const int64_t expectedOutputRows = w * rowCount.value();
+ if (ShapedType::isStatic(outputDataShape.getDimSize(1)) &&
+ outputDataShape.getDimSize(1) != expectedOutputRows)
+ return emitOpError() << "requires output[0] dimension 1 to have size "
+ << expectedOutputRows << ", got "
+ << outputDataShape.getDimSize(1);
+ }
+ }
+
+ if (values.size() == 2) {
+ const ShapeAdaptor valuesScaleShape(values[1].getType());
+ if (valuesScaleShape.hasRank()) {
+ if (failed(tryUpdateDimOrFailure(*this, n, valuesScaleShape.getDimSize(0),
+ "values[1]", "batch")) ||
+ failed(tryUpdateDimOrFailure(*this, k, valuesScaleShape.getDimSize(1),
+ "values[1]", "rows")))
+ return failure();
+ multiplesOfC = valuesScaleShape.getDimSize(2);
+ }
+
+ const ShapeAdaptor outputScaleShape(output[1].getType());
+ if (outputScaleShape.hasRank()) {
+ if (failed(tryUpdateDimOrFailure(*this, n, outputScaleShape.getDimSize(0),
+ "output[1]", "batch")))
+ return failure();
+
+ if (auto rowCount = getConstantScalarIntValue(getRowCount());
+ succeeded(rowCount) && rowCount.value() > 0 &&
+ ShapedType::isStatic(w)) {
+ const int64_t expectedOutputRows = w * rowCount.value();
+ if (ShapedType::isStatic(outputScaleShape.getDimSize(1)) &&
+ outputScaleShape.getDimSize(1) != expectedOutputRows)
+ return emitOpError() << "requires output[1] dimension 1 to have size "
+ << expectedOutputRows << ", got "
+ << outputScaleShape.getDimSize(1);
+ }
+
+ if (ShapedType::isDynamic(multiplesOfC))
+ multiplesOfC = outputScaleShape.getDimSize(2);
+ else if (ShapedType::isStatic(outputScaleShape.getDimSize(2)) &&
+ multiplesOfC != outputScaleShape.getDimSize(2))
+ return emitOpError()
+ << "expected channels of output[1] to match size "
+ << multiplesOfC << ", got " << outputScaleShape.getDimSize(2);
+ }
+
+ if (ShapedType::isStatic(c) && ShapedType::isStatic(multiplesOfC) &&
+ multiplesOfC != c / blockSize)
+ return emitOpError()
+ << "expects channels of scale tensors to equal C/block_size (" << c
+ << "/" << blockSize << "), got " << multiplesOfC;
+ }
+
+ return success();
+}
+
LogicalResult tosa::ResizeOp::inferReturnTypeComponents(
MLIRContext *context, ::std::optional<Location> location,
ResizeOp::Adaptor adaptor,
@@ -3987,6 +4174,8 @@ LogicalResult Conv2DBlockScaledOp::verify() {
// Verify IC is a multiple of block size
const uint32_t blockSize = BlockSizeAttr::getBlockSizeValue(getBlockSize());
+ if (blockSize == 1)
+ return emitOpError("requires block_size to not be BLOCK_SIZE_1");
if (ShapedType::isStatic(IC) && IC % blockSize != 0)
return emitOpError("expect IC to be a multiple of block size, got IC=")
<< IC << ", block_size=" << blockSize;
@@ -4577,6 +4766,8 @@ LogicalResult CastFromBlockScaledOp::verify() {
if (inputDataShape.hasRank()) {
const unsigned int blockSize =
BlockSizeAttr::getBlockSizeValue(getBlockSize());
+ if (blockSize == 1)
+ return emitOpError("requires block_size to not be BLOCK_SIZE_1");
const int64_t inputDataLastDim =
inputDataShape.getDimSize(inputDataShape.getRank() - 1);
if (inputDataLastDim % blockSize != 0)
@@ -4650,6 +4841,8 @@ LogicalResult CastToBlockScaledOp::verify() {
const unsigned int blockSize =
BlockSizeAttr::getBlockSizeValue(getBlockSize());
+ if (blockSize == 1)
+ return emitOpError("requires block_size to not be BLOCK_SIZE_1");
const ShapeAdaptor inputDataShape = ShapeAdaptor(inputDataType);
if (inputDataShape.hasRank()) {
const int64_t inputDataLastDim =
diff --git a/mlir/lib/Dialect/Tosa/Transforms/TosaProfileCompliance.cpp b/mlir/lib/Dialect/Tosa/Transforms/TosaProfileCompliance.cpp
index 01c85be4f704f..4ea225b860f6c 100644
--- a/mlir/lib/Dialect/Tosa/Transforms/TosaProfileCompliance.cpp
+++ b/mlir/lib/Dialect/Tosa/Transforms/TosaProfileCompliance.cpp
@@ -185,6 +185,18 @@ LogicalResult ProfileInfoDepot::populateProfileInfo(tosa::GatherOp op) {
return success();
}
+template <>
+LogicalResult
+ProfileInfoDepot::populateProfileInfo(tosa::RowGatherBlockScaledOp op) {
+ for (Value value : op.getValues())
+ addValue(value);
+ addValue(op.getIndices());
+ addValue(op.getRowCount());
+ for (Value result : op.getOutput())
+ addValue(result);
+ return success();
+}
+
template <>
LogicalResult ProfileInfoDepot::populateProfileInfo(tosa::ScatterOp op) {
addValue(op.getValuesIn());
@@ -288,6 +300,7 @@ LogicalResult ProfileInfoDepot::populatationDispatch(Operation *op) {
POPULATE_PROFILE_INFO_CUSTOM(Tile)
POPULATE_PROFILE_INFO_CUSTOM(Transpose)
POPULATE_PROFILE_INFO_CUSTOM(Gather)
+ POPULATE_PROFILE_INFO_CUSTOM(RowGatherBlockScaled)
POPULATE_PROFILE_INFO_CUSTOM(Scatter)
POPULATE_PROFILE_INFO_CUSTOM(Resize)
POPULATE_PROFILE_INFO_CUSTOM(Select)
@@ -598,10 +611,11 @@ SmallVector<OpComplianceInfo<T>> TosaProfileCompliance::findMatchedEntries(
SmallVector<VersionedTypeInfo> sets = compInfo[i].operandTypeInfoSet;
for (const auto &set : sets) {
SmallVector<TypeInfo> expected = set.first;
- assert(present.size() == expected.size() &&
- "the entries for profile-based compliance do not match between...
[truncated]
|
Contributor
Author
Contributor
|
LGTM except for one nit |
psunn
force-pushed
the
row_gather_block_scaled
branch
from
b5cc923 to
1c6765f
Compare
Contributor
Author
|
Updated to resolve merge conflicts. |
lhutton1
reviewed
Apr 16, 2026
psunn
force-pushed
the
row_gather_block_scaled
branch
3 times, most recently
from
bad45d3 to
4e88d6b
Compare
Note: defer op-specific level checks for ROW_GATHER_BLOCK_SCALED while the TOSA 1.1 draft is still evolving. Signed-off-by: Peng Sun <peng.sun@arm.com>
psunn
force-pushed
the
row_gather_block_scaled
branch
from
8c57610 to
a9ca5fb
Compare
Contributor
Author
|
updated to rebase on head of current main |
lhutton1
added a commit
to lhutton1/llvm-project
that referenced
this pull request
Apr 17, 2026
Fixes a validation test after a merge race condition with llvm#192122 and llvm#192272. Change-Id: I3b87e0ef9c04432b3af1a5aae7630dbae662e802
lhutton1
added a commit
that referenced
this pull request
Apr 17, 2026
llvm-sync Bot
pushed a commit
to arm/arm-toolchain
that referenced
this pull request
Apr 17, 2026
Fixes a validation test after a merge race condition with llvm/llvm-project#192122 and llvm/llvm-project#192272.
cpullvm-upstream-sync Bot
pushed a commit
to navaneethshan/cpullvm-toolchain-1
that referenced
this pull request
Apr 17, 2026
Fixes a validation test after a merge race condition with llvm/llvm-project#192122 and llvm/llvm-project#192272.
alexfh
pushed a commit
to alexfh/llvm-project
that referenced
this pull request
Apr 18, 2026
Add `tosa.row_gather_block_scaled` to the MLIR TOSA dialect, aligned with the current TOSA 1.1 draft spec and the implementation in `tosa-tools`. This includes: - op definition - verifier and shape inference support - validation / profile compliance wiring - availability and extension handling - lit tests for parsing, verification, shape inference, and version / extension gating The op supports both spec-defined forms: - non-block-scaled: 1 input value tensor, `BLOCK_SIZE_1`, 1 output - block-scaled: data + scale tensor list, non-`BLOCK_SIZE_1`, 2 outputs Op-specific level checks for ROW_GATHER_BLOCK_SCALED have been deferred while the TOSA 1.1 draft is still evolving. Signed-off-by: Peng Sun <peng.sun@arm.com>
alexfh
pushed a commit
to alexfh/llvm-project
that referenced
this pull request
Apr 18, 2026
Fixes a validation test after a merge race condition with llvm#192122 and llvm#192272.
llvm-upstreamsync Bot
pushed a commit
to qualcomm/cpullvm-toolchain
that referenced
this pull request
Apr 24, 2026
Fixes a validation test after a merge race condition with llvm/llvm-project#192122 and llvm/llvm-project#192272.
KHicketts
pushed a commit
to KHicketts/llvm-project
that referenced
this pull request
Apr 30, 2026
Fixes a validation test after a merge race condition with llvm#192122 and llvm#192272.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Add
tosa.row_gather_block_scaledto the MLIR TOSA dialect, aligned with the current TOSA 1.1 draft spec and the implementation intosa-tools.This includes:
Notes
The op supports both spec-defined forms:
BLOCK_SIZE_1, 1 outputBLOCK_SIZE_1, 2 outputsThis also tightens existing block-scaled-only ops to reject
BLOCK_SIZE_1now that it is part of the shared enum.Op-specific level checks for ROW_GATHER_BLOCK_SCALED have been deferred while the TOSA 1.1 draft is still evolving.