With llvm/llvm-project@112499f landed, `MLIR_TABLEGEN_EXE` is given as a
cache variable in the MLIR core project. Other external projects, such
as TORCH-MLIR, should not set the variable as this breaks
cross-compilation.
Summary of changes:
- Switch to C++17 (similar to https://reviews.llvm.org/D131348)
- Update MHLO to build with LLVM commit hash 061e0189
- Replace deprecated `hasValue()` and `getValue()` with `has_value()`
and `value()` respectively (https://reviews.llvm.org/D131349)
- Use `TypedAttr` (https://reviews.llvm.org/D130092)
- Use updated assembly format of `mhlo.compare` op (commit
d03ef01e70fbf9afd0fa1976fbb7ed31838929b3 in MHLO repo)
* Update native function definitions
* Add ops to support bert lowering
- Add empty_strided and as_strided
- Restore zeros_like to op blacklist (Without this, tensors will be unintentionally created with a CPU device rather than lazy)
- Check for composite implicit ops and add device data IR
- Also fix codegen for functionalization
* Add autogen to CMakeList
* Remove PyTorch submodule
* Reduced BERT model size
* Print Mark Step status in Torch MLIR LTC debug string
* Apply fixes to work with latest upstream/main
- Pass importOptions into getMlirTypeFromTorchType during NodeImporter::importNode
Without this, the tensor type created may have a mismatched type as ImportOptions may cause vtensor to be used instead of tensor
* Update shape inference functions
- Fixed compute_shape_native_batch_norm when mean and var are uninitialized
Previously, the number of shapes returned would be <3 if either mean or val was didn't exist. Instead, we now initialize them with a vector matching the number of channels.
- Implemented compute_shape_mul
- Fixed bug in reshape shape inference error message
* Get MLIR backend more consistent with TS backend
- Remove LazyNativeFunctions::_unsafe_view from autogen
- Blacklist ops to make JIT graph more like output of TS backend
- Print graph when SSA value has mismatch of types and results
- Remove normalize_index from LazyShapeInference
- Fix seeds for LTC example models
* Update and clean up shape inference functions
- Prune shape inference functions
- Add shape inference function for GenerateSlice
- Add shape inference function for GenerateCopy
Co-authored-by: Henry Tu <henry.tu@cerebras.net>
This patch makes some rudimentary changes to torch-mlir's use of MLIR
Python bindings to work with the most recent LLVM code. We can perhaps
do better by being more selective in what we link against, instead of
using `MLIRPythonExtension.RegisterEverything`.
The biggest change here is to upgrade RefineTypes to the new sparse
dataflow framework.
Smaller changes:
- minor changes to type parsing
- suppress warnings in e2e tests
When compiling without assertions (i.e. in `NDEBUG` mode), a handful of
statements turn to NOPs, which results in warnings such as missing
return statement or unused variables and function. This patch replaces
such statements with `llvm_unreachable()`, which informs the compiler
about program termination regardless of the `NDEBUG` mode. This also
enables torch-mlir to be compiled using the flags `-Wall`, `-Wextra`,
`-Wpedantic`, and `-Werror`.
In addition to updating the llvm-project submodule, this patch also:
1. updates shape functions and tests so that `func` and `call`
operations refer to the `func` dialect
2. avoid duplicate registration of dialects
The updated LLVM code includes a patch to create bfloat16 array
attributes, thus enabling a different patch to torch-mlir to flesh out
support for the bfloat16 type.
Follows up on #623 for out-of-tree builds of torch-mlir, which
added building `torch-mir-dialects` as a subdirectory.
Our goal is to support both in-tree and out-of-tree builds of
`torch-mlir` with minimum hassle, for instance by using the same
variable names in both setups.
Specific changes to `externals/llvm-external-projects/torch-mlir-dialects/CMakeLists.txt`:
- We use `MLIR_FOUND` to detect that it is being build as a subdirectory
and the llvm+mlir cmake infrastructure is already set up (via
find_package in the parent build) as opposed to an in-tree build.
- For in-tree, the setting of variables and loading of llvm+mlir cmake
infrastructure is now conditionally performed.
- For in-tree, the names of cmake variables being defined for are
adjusted to match those `llvm-project` makes available through
`find_package(MLIR REQUIRED CONFIG)`, under the assumption that those
are the more "standardized" names.
Co-authored-by: Clément Fournier <clement.fournier@amd.com>
Co-authored-by: Liam Fitzpatrick <liam.fitzpatrick@xilinx.com>
Ashay Rane
Marius Brehler
Jae Hoon (Antonio) Kim
Ziheng Jiang
Sean Silva
gpetters94
Maksim Levental
Clément Fournier
Ahmed S. Taei