Add computeStructuralHash for shader graphs#2976
Open
ashwinbhat wants to merge 7 commits into
Open
Conversation
Introduce computeStructuralHash in MaterialXGenShader, a free function that produces a size_t hash of a finalized ShaderGraph capturing its topology and port types while ignoring instance-specific data (node/port names, parameter values, and texture paths). Two graphs that wire up the same node implementations in the same pattern hash identically even when their constants differ, enabling shader deduplication/caching and distinct-structure reporting across large material libraries. The function is an external visitor over the public API of ShaderGraph, ShaderNode, and ShaderPort: it never mutates the graph and folds node implementation hashes, classifications, port structure (type, semantic, colorspace, unit, geomprop, uniform/bind-input flags), and connection topology into a Boost/Hydra-style hash combiner. The hash is deterministic within a build but session-local: it relies on std::hash and implementation hashes, so it must not be persisted or compared across builds/versions, and it is a fast pre-filter rather than a collision-proof equivalence proof. Also adds: - Python binding (computeStructuralHash) in PyMaterialXGenShader. - A --dumpHash flag to generateshader.py. - A GLSL structural-hash test exercising generation and determinism.
There was a problem hiding this comment.
Pull request overview
This PR introduces a new structural hashing utility for MaterialXGenShader shader graphs to support shader deduplication/caching by hashing graph topology and port structure while ignoring instance-specific details (names, values, paths).
Changes:
- Added
computeStructuralHash(const ShaderGraph&)inMaterialXGenShaderto hash finalized shader graph structure/topology. - Exposed the hash via Python (
PyMaterialXGenShader.computeStructuralHash(shader)), and added a--dumpHashCLI flag togenerateshader.py. - Added a GLSL-side Catch2 test to exercise structural hash determinism.
Reviewed changes
Copilot reviewed 5 out of 5 changed files in this pull request and generated 4 comments.
Show a summary per file
| File | Description |
|---|---|
| source/PyMaterialX/PyMaterialXGenShader/PyUtil.cpp | Adds Python binding for structural hashing via a Shader wrapper function. |
| source/MaterialXTest/MaterialXGenGlsl/GenGlsl.cpp | Adds a Catch2 test that computes/prints structural hashes and checks determinism. |
| source/MaterialXGenShader/ShaderGraphHash.h | Declares the new public computeStructuralHash API. |
| source/MaterialXGenShader/ShaderGraphHash.cpp | Implements the structural hash over graph sockets, nodes, ports, and connections. |
| python/Scripts/generateshader.py | Adds --dumpHash to print the structural hash for generated shaders. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+281
to
+286
| std::string message; | ||
| bool docValid = doc->validate(&message); | ||
| if (!docValid) | ||
| { | ||
| continue; | ||
| } |
Comment on lines
+293
to
+301
| mx::ShaderPtr shader; | ||
| try | ||
| { | ||
| shader = context.getShaderGenerator().generate(element->getName(), element, context); | ||
| } | ||
| catch (const std::exception&) | ||
| { | ||
| continue; | ||
| } |
Comment on lines
+309
to
+317
| mx::ShaderPtr shader2; | ||
| try | ||
| { | ||
| shader2 = context.getShaderGenerator().generate(element->getName(), element, context); | ||
| } | ||
| catch (const std::exception&) | ||
| { | ||
| continue; | ||
| } |
Comment on lines
+331
to
+334
| // Output to Catch2 INFO so it appears with -s flag | ||
| INFO(hashLog.str()); | ||
| SUCCEED(); | ||
| } |
kwokcb
reviewed
Jul 9, 2026
kwokcb
left a comment
kwokcb
left a comment
Contributor
There was a problem hiding this comment.
@jstone-lucasfilm suggested adding enough API exposure for my PR for building graph diagrams from a Shader as seen in #2790.
Basically expose enough to allow you to build your own utilities.
I think this could be an approach to consider here as well vs an opaque non-customizable hash function. I think this could work:
- Add a "hash" class and provide your current implementation as one case.
- Expose it in the API so you can create custom ones in C++, Python or JS.
- e.g. in
generateshader.pyyou could even add a "hasher" option like the validator option for custom hashes.
This follows the approach for customizable entry points such as generators and loaders.
Thoughts ?
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
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.
Introduce computeStructuralHash in MaterialXGenShader, a free function that produces a size_t hash of a finalized ShaderGraph capturing its topology and port types while ignoring instance-specific data (node/port names, parameter values, and texture paths). Two graphs that wire up the same node implementations in the same pattern hash identically even when their constants differ, enabling shader deduplication/caching and distinct-structure reporting across large material libraries.
The function is an external visitor over the public API of ShaderGraph, ShaderNode, and ShaderPort: it never mutates the graph and folds node implementation hashes, classifications, port structure (type, semantic, colorspace, unit, geomprop, uniform/bind-input flags), and connection topology into a Boost/Hydra-style hash combiner.
The hash is deterministic within a build but session-local: it relies on std::hash and implementation hashes, so it must not be persisted or compared across builds/versions, and it is a fast pre-filter rather than a collision-proof equivalence proof.
Also adds: