Refactoring Fusion Executor, pulling out compiled kernel #3468
Conversation
This file contains 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
…d in executor::compile.
…need to run compileRTC and runRTC calls in tests with CompiledKernel instances directly. Also fix profiling calls in KernelExecutor.
…owering but is now checked after lowering.
|
!test |
csarofeen
commented
Dec 22, 2024
csrc/fusion.cpp
Outdated
| @@ -231,6 +231,8 @@ void Fusion::removeVal(Val* val) { | |||
| void Fusion::addInput(Val* input) { | |||
| assertInContainer(input, "Cannot register input "); | |||
|
|
|||
| std::cout << "Registering input: " << input->toString() << std::endl; | |||
| buffer << cuda_src.rdbuf(); | ||
| return buffer.str(); | ||
| } | ||
| } // namespace |
There was a problem hiding this comment.
Everything above this is only code motion.
| : compile_params_(compile_params), | ||
| lowered_(std::make_unique<GpuLower>(fusion, compile_params)) { | ||
| FUSER_PERF_SCOPE("CompiledKernel::CompiledKernel"); | ||
| // TODO: No hooks can be sent because this is in the constructor |
csrc/runtime/compiled_kernel.cpp
Outdated
| lowered_->run(); | ||
| } | ||
|
|
||
| // TODO:Rename to "compile" |
csrc/runtime/compiled_kernel.cpp
Outdated
| // between them. | ||
| return val->isFusionInput() && !val->isA<TensorView>(); | ||
| })) { | ||
| // TODO: parameter cache is too big a hammer here. We should consider |
There was a problem hiding this comment.
This is an old TODO I don't plan to address here. I'm not sure what the challenge of caching with scalar inputs is.
| // This could be refactored. | ||
| struct CompiledKernel : public NonCopyable { | ||
| NVF_API ~CompiledKernel(); | ||
| struct CudaExecutable : public NonCopyable { |
There was a problem hiding this comment.
TODO can this be merged into compiled_kernel.h?
| @@ -58,20 +47,11 @@ struct CompiledKernel : public NonCopyable { | |||
| int register_spills = -1; | |||
| }; | |||
|
|
|||
| // Returns executable function and the ptxas log from compilation | |||
There was a problem hiding this comment.
Moved to compiled_kernel.cpp
| @@ -253,12 +233,5 @@ void validateCircularBuffering( | |||
| kir::Kernel* kernel, | |||
| ExpressionEvaluator& expr_eval); | |||
|
|
|||
| //! Query the target GPU version number NVRTC compiles CUDA kernels for | |||
There was a problem hiding this comment.
Moved to compiled_kernel.cpp
There was a problem hiding this comment.
Actually these are likely just removed as they should now be contained in runtime/compiled_kernel.cpp
| @@ -32,117 +30,6 @@ | |||
| #include <cstdlib> | |||
|
|
|||
| namespace nvfuser { | |||
|
|
|||
There was a problem hiding this comment.
Moved to compiled_kernel.cpp
| @@ -194,14 +81,6 @@ bool detectComputeSanitizer() { | |||
|
|
|||
| namespace nvfuser { | |||
|
|
|||
| namespace executor_utils { | |||
There was a problem hiding this comment.
Moved to compiled_kernel.cpp
|
TODO list:
|
|
!test |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Pull out kernel compilation from the KernelExecutor, trying to separate out the two concepts as we will move towards a world where the execution of a kernel is done through HostIr.
Moved code: