arno-lunarg
arno-lunarg
@corporateshark I think you just need to disable the shader caching option under shader instrumentation. Cached shaders must be the instrumented ones, and since you disabled validation after that, the...
Regarding perf, how does it behave on your setup when toggling "out of bounds buffer device addresses" on/off?
For 2), Disabling caching should result in no false positive. Can you confirm please? For now, every time you change how shaders are instrumented (i.e, toggle "Descriptors indexing", "out of...
On your side, not much you can do... If you now go back to the previous SDK, do you still observe the same performance hit?
As of those 20 FPS, are you still able to get them when now switching to SDK 1.3.280 ?
Hello @f32by Thank you for reporting this issue, in a way it happens to be the right time for it given what I am working on. I did not managed...
I updated the issue. Not sure about how to fix this. It looks like modifying `enabled_features` in `GpuShaderInstrumentor::PreCallRecordCreateDevice` is a hack, I do not want to go down that route....
You are right, thank you for spotting this!
Still crashing in CTS. 2 tests can be used to try to reproduce, `dEQP-VK.ray_tracing_pipeline.opacity_micromap.disable_opacity_micromap_instance.map_value.2.level_9` and `dEQP-VK.ray_query.builtin.flow.vert.triangles` When only `dEQP-VK.ray_tracing_pipeline.opacity_micromap.disable_opacity_micromap_instance.map_value.2.level_9` is ran, the test fails. if only `dEQP-VK.ray_query.builtin.flow.vert.triangles` is ran, the...
Finally took the time to run those tests, no crash, closing