Use constant to generate the correct assertion message. This avoids
confusion when lowering the max rank specialization for debugging.
PiperOrigin-RevId: 344769021
Previously this started at rank 2 after checking for scalars and equal shapes. This resulted in cases such as <1xf32> + <2xf32> being treated as impossible.
PiperOrigin-RevId: 341043965
Additionally:
- Forward listeners through new if/else op builders.
This corrects an error that led to incomplete legalization of broadcasted op
lowering.
- Use OpConversionPattern to ensure up to date operand values are used.
PiperOrigin-RevId: 339838833
* Unified TF->Cubin and TF->Kernel_with_host side lowering in `kernel_creator.h|cc`
* Added a pass that attaches GPU binary blob to GPUModuleOp
* Refactored most of the code.
* Added tf_to_kernel binary that emits obj file
PiperOrigin-RevId: 330494488
The transformation of unranked to ranked operations no longer generates cast
operations for shapes and sizes. Instead, we use the newly introduced support
for extent tensor and index types directly.
PiperOrigin-RevId: 325057440
Shuffle files around, use TableGen to register passes, and introduce
a `mlir-hlo-opt.cpp` file to hold the main entry point of the -opt tool
and stop relying on static registration for dialect/passes.
PiperOrigin-RevId: 323674455
A. Unique TensorFlower
Adrian Kuegel
Alexander Belyaev
Lucy Fox
Tres Popp
Thomas Joerg
Chao Xie
Mehdi Amini
Stephan Herhut