vendor optimize #703
Merged
vendor optimize #703
Conversation
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
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
Comment on lines
+415
to
+418
| # From https://github.com/JuliaGPU/GPUCompiler.jl/blob/7b9322faa34685026c4601a5084eecf5a5d7f3fe/src/ptx.jl#L149 | ||
| function vendored_optimize_module!(@nospecialize(job), | ||
| mod::LLVM.Module, | ||
| instcombine::Bool=false |
There was a problem hiding this comment.
[JuliaFormatter] reported by reviewdog 🐶
Suggested change
| # From https://github.com/JuliaGPU/GPUCompiler.jl/blob/7b9322faa34685026c4601a5084eecf5a5d7f3fe/src/ptx.jl#L149 | |
| function vendored_optimize_module!(@nospecialize(job), | |
| mod::LLVM.Module, | |
| instcombine::Bool=false | |
| function vendored_optimize_module!( | |
| @nospecialize(job), mod::LLVM.Module, instcombine::Bool=false | |
| ) |
| instcombine::Bool=false | ||
| ) | ||
| tm = GPUCompiler.llvm_machine(job.config.target) | ||
| # TODO: Use the registered target passes (JuliaGPU/GPUCompiler.jl#450) |
There was a problem hiding this comment.
[JuliaFormatter] reported by reviewdog 🐶
Suggested change
| # TODO: Use the registered target passes (JuliaGPU/GPUCompiler.jl#450) | |
| LLVM.@dispose pb = LLVM.NewPMPassBuilder() begin |
giordano
reviewed
Feb 7, 2025
giordano
reviewed
Feb 7, 2025
ext/ReactantCUDAExt.jl
Outdated
Comment on lines
423
to
427
| LLVM.register!(pb, LLVM.NVVMReflectPass()) | ||
|
|
||
| LLVM.add!(pb, LLVM.NewPMFunctionPassManager()) do fpm | ||
| # TODO: need to run this earlier; optimize_module! is called after addOptimizationPasses! | ||
| LLVM.add!(fpm, LLVM.NVVMReflectPass()) |
There was a problem hiding this comment.
Suggested change
| LLVM.register!(pb, LLVM.NVVMReflectPass()) | |
| LLVM.add!(pb, LLVM.NewPMFunctionPassManager()) do fpm | |
| # TODO: need to run this earlier; optimize_module! is called after addOptimizationPasses! | |
| LLVM.add!(fpm, LLVM.NVVMReflectPass()) | |
| LLVM.register!(pb, GPUCompiler.NVVMReflectPass()) | |
| LLVM.add!(pb, LLVM.NewPMFunctionPassManager()) do fpm | |
| # TODO: need to run this earlier; optimize_module! is called after addOptimizationPasses! | |
| LLVM.add!(fpm, GPUCompiler.NVVMReflectPass()) |
giordano
reviewed
Feb 7, 2025
giordano
reviewed
Feb 7, 2025
| LLVM.add!(fpm, LLVM.InstSimplifyPass()) # clean-up redundancy | ||
| end | ||
| LLVM.add!(fpm, LLVM.NewPMLoopPassManager(; use_memory_ssa=true)) do lpm | ||
| LLVM.add!(lpm, LLVM.LICMPass()) # the inner runtime check might be |
There was a problem hiding this comment.
[JuliaFormatter] reported by reviewdog 🐶
Suggested change
| LLVM.add!(lpm, LLVM.LICMPass()) # the inner runtime check might be | |
| # outer loop invariant |
wsmoses
added a commit
that referenced
this pull request
Feb 8, 2025
* vendor optimize * Update ReactantCUDAExt.jl * Update ext/ReactantCUDAExt.jl * Update ext/ReactantCUDAExt.jl * try forcing random seed for basic test --------- Co-authored-by: Mosè Giordano <765740+giordano@users.noreply.github.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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.
@gbaraldi vendoring this in interim to unblock kernels