Add some furnace tests for some of the MaterialX closures #1542
Conversation
fpsunflower
changed the title
|
This is failing across the board, but only the render-mx-layer test. Which is odd because that test was not modified directly by this PR. And it's not just a speckle or occasional LSB errors. The whole thing has a visible intensity shift. Do you think that's because of the change to get_albedo? I'm about to test whether it's sufficient to just update that one ref file without even having to adjust the comparison thresholds (i.e., if it's just different now, but not experiencing significant per-platform variation). |
|
The mx-layer test was added in a previous PR, but was not actually being run. I added it to the list of tests to run but didn't re-generate the image yet. |
|
It looks like you just need to update that one ref image and all will be fine. |
fpsunflower
force-pushed
the
mx-furnace
branch
4 times, most recently
from
255a827 to
2a9a011
Compare
|
The one remaining that fails is the Debug build, and at least some of the test were failing simply by timing out -- some of the render tests take a godawful long time when not optimized. Like that touch-up patch I added earlier today, we may just want to bump up the ctest timeout value for the individual tests higher in debug mode. |
|
I'm tempted to merge this and just bump the timeouts for those failing tests individually, as a separate PR. Shall I do that, or do you really want to do that yourself, @fpsunflower ? |
|
Yeah I'm fine with ignoring the timeouts, but there is one legit failure in debug mode -- an assert is being triggered. I was just able to repro locally, so let me try to sort that out first ... |
Implement get_albedo for the sheen closure, based on fit derived by the MaterialX team Add optional mode to testrender to visualize the result of the get_albedo method. This can be useful for validation of the get_albedo method. Signed-off-by: Chris Kulla <[email protected]>
) Implement get_albedo for the sheen closure, based on fit derived by the MaterialX team Add optional mode to testrender to visualize the result of the get_albedo method. This can be useful for validation of the get_albedo method. Signed-off-by: Chris Kulla <[email protected]>
) Implement get_albedo for the sheen closure, based on fit derived by the MaterialX team Add optional mode to testrender to visualize the result of the get_albedo method. This can be useful for validation of the get_albedo method. Signed-off-by: Chris Kulla <[email protected]>
Relevant changes: * GPU string fix -- needed missing cast (AcademySoftwareFoundation#1553) * Fix typo error that prevented correct typecheck of ternary (AcademySoftwareFoundation#1552) * testrender now supports the standard MaterialX closure: * testrender improvements: modernize sampler (AcademySoftwareFoundation#1534), switch to cone tracing (AcademySoftwareFoundation#1543), support MaterialX closures (AcademySoftwareFoundation#1533, AcademySoftwareFoundation#1536, AcademySoftwareFoundation#1537, AcademySoftwareFoundation#1538, AcademySoftwareFoundation#1541, AcademySoftwareFoundation#1542, AcademySoftwareFoundation#1547) * Bump LLVM to 14.0 * Help reduce PTX nondeterminism (AcademySoftwareFoundation#1570) See merge request spi/dev/3rd-party/osl-feedstock!30
This is to quantify the energy loss/gain of the individual closures. I've also added a way for testrender to show the results of the
get_albedo()method directly, for comparison with furnace tests.I've also added a rational fit for the albedo of sheen, which is taken from the MaterialX project. It doesn't look like an exact match, but it should be good enough for now.