llvm
/
torch-mlir
mirror of https://github.com/llvm/torch-mlir
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Bump revision and fix issues. 

* llvm revision = 4836188ad9b3334b0c1e055d45ccaa54ed797e4b
* iree revision = 091482e8fdf599d6cb5c701d5b3ccb27fc66c014
pull/1/head
Stella Laurenzo 2020-06-19 10:38:51 -07:00
parent 529873d13c
commit fc5f10c5c5
4 changed files with 9 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
README.md
View File

@ -27,7 +27,7 @@ export LDFLAGS=-fuse-ld=$(which ld.lld-$LLVM_VERSION)
export LLVM_SRC_DIR=/path/to/llvm-project
# Check out last known good commit.
(cd $LLVM_SRC_DIR && git checkout 52cae05e087b3d4fd02849fc37c387c720055ffb)
(cd $LLVM_SRC_DIR && git checkout 4836188ad9b3334b0c1e055d45ccaa54ed797e4b)
./tools/install_mlir.sh
./tools/cmake_configure.sh
@ -70,7 +70,7 @@ Notes:
* Python sources are symlinked to the output directory at configure time.
Adding sources will require a reconfigure. Editing should not.
* It is a very common issue to have both python 2.7 (aka. "python") and python
3.x (aka. "python3") on a system at a time (and we can only hope that one
3.x (aka. "python3") on a system at a time (and we can only hope that one
day this ends). Since the native library at development time binds to a
specific version, if you try to run with a different python, you will get
an error about the "native" module not being found.

4
lib/E2E/LowerRankedShapes.cpp
View File

@ -26,7 +26,7 @@ public:
LogicalResult
matchAndRewrite(shape::BroadcastOp op, ArrayRef<Value> operands,
ConversionPatternRewriter &rewriter) const override {
shape::BroadcastOp::OperandAdaptor adaptor(operands);
shape::BroadcastOp::Adaptor adaptor(operands);
auto lhs = adaptor.lhs().getDefiningOp<shape::FromExtentsOp>();
auto rhs = adaptor.rhs().getDefiningOp<shape::FromExtentsOp>();
if (!lhs || !rhs)
@ -88,7 +88,7 @@ public:
LogicalResult
matchAndRewrite(tcp::GetExtentOp op, ArrayRef<Value> operands,
ConversionPatternRewriter &rewriter) const override {
tcp::GetExtentOp::OperandAdaptor adaptor(operands);
tcp::GetExtentOp::Adaptor adaptor(operands);
auto fromExtents = adaptor.shape().getDefiningOp<shape::FromExtentsOp>();
if (!fromExtents)
return rewriter.notifyMatchFailure(op, "not a from_extents op");

4
lib/E2E/LowerToLLVM.cpp
View File

@ -28,7 +28,7 @@ public:
LogicalResult
matchAndRewrite(tcp::AbortIfOp op, ArrayRef<Value> operands,
ConversionPatternRewriter &rewriter) const override {
tcp::AbortIfOp::OperandAdaptor adaptor(operands);
tcp::AbortIfOp::Adaptor adaptor(operands);
rewriter.replaceOpWithNewOp<LLVM::CallOp>(op, abortIfFunc, adaptor.pred());
return success();
}
@ -67,7 +67,7 @@ class LowerToLLVM : public LowerToLLVMBase<LowerToLLVM> {
populateStdToLLVMConversionPatterns(converter, patterns);
patterns.insert<LowerAbortIf>(abortIfFunc);
if (failed(applyFullConversion(module, target, patterns, &converter))) {
if (failed(applyFullConversion(module, target, patterns))) {
return signalPassFailure();
}
}

6
lib/E2E/LowerToMemRefABI.cpp
View File

@ -28,7 +28,7 @@ public:
LogicalResult
matchAndRewrite(TensorStoreOp op, ArrayRef<Value> operands,
ConversionPatternRewriter &rewriter) const override {
TensorStoreOp::OperandAdaptor adaptor(operands);
TensorStoreOp::Adaptor adaptor(operands);
// The tensor has been converted to an unranked memref. We need to cast
// it to the original memref type and copy it to the destination.
//
@ -52,7 +52,7 @@ public:
LogicalResult
matchAndRewrite(TensorLoadOp op, ArrayRef<Value> operands,
ConversionPatternRewriter &rewriter) const override {
TensorLoadOp::OperandAdaptor adaptor(operands);
TensorLoadOp::Adaptor adaptor(operands);
auto type = UnrankedMemRefType::get(op.getType().getElementType(), 0);
// TODO: This won't work. The LLVM unranked memref calling convention
// doesn't allow returning an unranked memref becuase it lowers it to
@ -93,7 +93,7 @@ public:
LogicalResult
matchAndRewrite(shape::ShapeOfOp op, ArrayRef<Value> operands,
ConversionPatternRewriter &rewriter) const override {
shape::ShapeOfOp::OperandAdaptor adaptor(operands);
shape::ShapeOfOp::Adaptor adaptor(operands);
auto tensorType = op.arg().getType().cast<RankedTensorType>();
auto rankedMemRefType =
MemRefType::get(tensorType.getShape(), tensorType.getElementType());