* For draws where some fragments discard, D3D12 & Vulkan currently can't
distinguish which specific fragments did and didn't discard because they're
processed batched in separate callbacks rather than incrementally like D3D11 &
GL which can track a discarded offset by doing one stencil count at a time.
This means we can't get primitive information or ordering, we only get a list
of the passing fragments and a set of failing fragments.
* Note that on all APIs we also can't distinguish when instancing is present,
because there could be multiple "identical" primitives from different
primitives, some of which discard and some don't.
* This includes:
- Incorrect comparisons
- Missing type annotations where it can't be inferred
- Variable shadowing and bad use
- Unused imports
- asserts for non-None on types
* Passes clean on strict pyright checking except for
`reportMissingParameterType`, `reportUnusedVariable`, and
`reportOptionalMemberAccess` which are all too spammy and low value to be
worth addressing.
* These come from long ago when using PyCharm with not as good type checking and
with renderdoc module stubs that were incomplete so needed help identifying
types.
* assert is handled specially by the type checker so we don't want to have
duplicate self.check / assert
* The original reason for TestCase.check was to print the assertion message but
this can be handled by checking for the AssertionError exception
Added D3D11_Workgroup_Zoo, D3D12 Workgroup Zoo, Vk Workgroup Zoo performance tests
Performance tests primarily for performance testing workgroup debugging speed
Changed workgroup and subgroup result variable from "data" -> "testResult" to help to reduce conflicts
More checking of GSM local/global cache behaviour when debugging
One test is not GPU stable and its results are verified against hard coded expectation (this is to test the expected behaviour of the local GSM cache on the active thread)
Subgroup_Zoo : unit tests, non-trivial convergence tests moved to Workgroup_Zoo
Workgroup_Zoo : convergence tests, small number of unit tests (not full coverage)
Added checks for workgroup convergence in Workgroup_Zoo tests
* Vulkan uses barrier()
* D3D12 uses AllMemoryBarrierWithGroupSync()
* dispatches workgroup of 2x1x1
* test debug results for workgroup 1,0,0
Check the renderdoc log for lines matching "Assertion" or "Error"
Using new helper function in testcase.py
def check_renderdoc_log(self, asserts: bool = True, errors: bool = True):
Tests specifically aimed at workgroup debugging i.e GSM and non-aligned subgroups
Not focused on unit tests of subgroup/quad instructions that is handled by *_Subgroup_Zoo
DebugShaderInputs struct contains existing the parameters:
sample
primitive
and a new parameter
view
Default constructor for DebugShaderInputs sets the parameters to NoPreference (~0U).
New 'Exploded' visualisation mode in BufferViewer with new exploder controls
hidden when not in 'Exploded' mode.
Change 'solidShading' and 'solidShadeMode' to 'visualisation' and
'visualisationMode'.
Hide the 'highlightVerts' widget when using 'Exploded' vis for both
real-estate and practical implementation reasons.
baldurk