Improve performance when compiling small shaders. #6396
Merged
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
jkwak-work
previously approved these changes
Feb 20, 2025
Comment on lines
+796
to
+797
| String slice; | ||
| HashCode64 hashCode; |
There was a problem hiding this comment.
Probably we should make the members private.
Not sure if we want to make them const as well since it is Immutable.
Comment on lines
+20312
to
20315
| [require(glsl)] | ||
| __specialized_for_target(glsl) | ||
| [[vk::spirv_instruction(1, "NonSemantic.DebugBreak")]] | ||
| void __glslDebugBreak(); |
There was a problem hiding this comment.
Maybe we should move this to glsl.meta.slang?
source/slang/slang-ir-link.cpp
Outdated
| IRBuilder builderStorage; | ||
|
|
||
| // The "global" specialization environment. | ||
| IRSpecEnv globalEnv; | ||
| }; | ||
|
|
||
| void insertGlobalValueSymbol(IRSharedSpecContext* sharedContext, IRInst* gv); | ||
|
|
||
| struct WitnessTableCloenInfo : RefObject |
There was a problem hiding this comment.
Copilot found typo here.
WitnessTableCloenInfo -> WitnessTableCloneInfo
There was a problem hiding this comment.
good catch, how did you setup copilot to do the review?
|
|
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Outdated suggestions cannot be applied.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.
Addresses #6358.
This PR address three performance issues when compiling small shaders:
These three optimizations combined achieves a 75% reduction in the benchmark provided in #6358.
Copilot summary:
This pull request includes several changes aimed at improving the handling of GLSL modules, enhancing the differentiation capabilities, and optimizing the IR linking process. The most important changes are grouped by their themes below:
GLSL Module Handling:
SemanticsVisitor::importModuleIntoScopemethod to set theglslModuleDeclwhen the module name is "glsl".OperatorOverloadCacheKeyto include anisGLSLModeflag and adjusted its hash code calculation accordingly.ResolveInvoketo set theisGLSLModeflag based on the presence of the GLSL module and cache the resolved operator overloads conditionally. [1] [2]Differentiation Enhancements:
IDifferentiableinterface with theKnownBuiltin("IDifferentiable")attribute.NullDifferentialstruct to remove theexportkeyword.IDifferentiableinterface to useIRKnownBuiltinDecorationinstead ofIRNameHintDecoration.IR Linking Optimization:
ImmutableHashedStringclass for efficient string handling in the IR linking process.IRSharedSpecContext::SymbolDictionaryto useImmutableHashedStringand added a new dictionary for imported symbols. [1] [2]Type Checking Cache:
TypeCheckingCacheto inherit fromRefObjectand updated its usage inLinkageandSessionclasses. [1] [2] [3]Debugging and Specialized Targets:
__specialized_for_target(glsl)attribute and corresponding SPIR-V instruction.These changes collectively improve the handling of GLSL modules, enhance differentiation capabilities, and optimize the IR linking process, making the codebase more efficient and maintainable.