You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Parse and convert shader once during app lifetime or bundle the cache with the app. This implies a fixed number of shaders and variations.
Performance gain are noticeable on BN validation tests or with users app.
This works at Native level. I don't see a need to expose it to JS.
1 cache per engine. Construct it once per App, then add a reference in engine
Use (optional?) compressor like LZ4 for extra compression
Store API in cache header so it's not possible to mismatch
It makes sense to use UnitTests to generate the cache and inject it back in the build. How to do it efficiently on the CI? Is it possible to generate GLSL on win32 platform. Or Metal from a MacOS CI agent? -> leave the work to users
Logging/exception when a shader is not found in cache and GLSLANG/SPIRV is disabled
Add build options to disable SPIRV/GLSlang build and link.
Parse and convert shader once during app lifetime or bundle the cache with the app. This implies a fixed number of shaders and variations.
Performance gain are noticeable on BN validation tests or with users app.
This works at Native level. I don't see a need to expose it to JS.
PR : #1402
The text was updated successfully, but these errors were encountered: