Avoid boxing compiler jobs during kernel resolution - #635
Merged
Merged
Conversation
Specialize compile_or_lookup on the backend target and parameter types. Keep it out of line to avoid duplicating the body in callers specialized per kernel, while preserving the existing GPUCompiler cache path.
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #635 +/- ##
=======================================
Coverage 81.53% 81.53%
=======================================
Files 50 50
Lines 3607 3607
=======================================
Hits 2941 2941
Misses 666 666 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Kernel resolution constructs a
CompilerJobimmediately before passing it through an unspecializedcompile_or_lookupboundary, which can force the job onto the heap. Allow specialization on the job's target and compiler-parameter types, and keep the method@noinlineso its body is not duplicated into callers specialized per kernel. Kernel identity remains a field of the job, so this does not introduce per-kernel specialization or another cache.Applies the approach from KernelAbstractions.jl#769.
Syntax and whitespace checks passed. Runtime behavior and allocation savings have not been measured on Intel GPU hardware.