encounter
/
dawn-cmake
mirror of https://github.com/encounter/dawn-cmake.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
dawn-cmake/webgpu-cts
History
Austin Eng 00dbf0bf67 Add workarounds to blit a buffer to a depth/stencil texture 
Works around issues on Metal Intel where CopyB2T and WriteTexture
with depth/stencil textures do not work correctly.
Fixes test failures with depth16unorm in the CTS.

Deletes UseTempTextureInStencilTextureToBufferCopy in favor of
the stencil blit. The former supposedly fixes a problem where
the stencil data is not flushed into the real stencil texture
by performing another T2T copy. This only works because the Metal
Intel backend also happens to allocate s8 as d32s8. Copying
the depth aspect as well seems to make the driver remember to
flush the data.
The stencil blit is a better fix for the problem since entirely
avoids getting the driver into a bad state where the stencil data
is not in sync.

Fixed: dawn:1389
Change-Id: If34b1d58996157036c164a5bc329e38b5e53f67a
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/117910
Reviewed-by: Loko Kung <lokokung@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Austin Eng <enga@chromium.org>
 		2023-01-30 17:13:12 +00:00
..
scripts Generate and load CTS cache 2022-11-21 15:16:51 +00:00
webtests Roll third_party/webgpu-cts/ 9fbed0099..3f9b09b36 (8 commits) 2023-01-24 20:32:05 +00:00
BUILD.gn Add test list to deps 2022-08-11 15:16:38 +00:00
PRESUBMIT.py Reland "Add expectations.txt validation to presubmit" 2022-12-08 21:57:55 +00:00
README.md Update CTS docs 2022-09-14 17:13:17 +00:00
expectations.txt Add workarounds to blit a buffer to a depth/stencil texture 2023-01-30 17:13:12 +00:00
test_page.html Change base directory for CTS runner, part 2 2022-06-24 18:50:19 +00:00
test_runner.js Roll third_party/webgpu-cts/ 0673bb68c..6d3a80420 (1 commit) 2023-01-05 21:44:37 +00:00
worker_test_globs.txt Move WebGPU worker globs 2022-07-01 01:20:14 +00:00

README.md

Running the WebGPU CTS Locally with Chrome

Running the WebGPU CTS locally with Chrome requires a Chromium checkout.

Follow these instructions for checking out and building Chrome. You'll also need to build the telemetry_gpu_integration_test target.

At the root of a Chromium checkout, run: ./content/test/gpu/run_gpu_integration_test.py webgpu_cts --browser=exact --browser-executable=path/to/your/chrome-executable

If you don't want to build Chrome, you can still run the CTS, by passing the path to an existing Chrome executable to the --browser-executable argument. However, if you would like to use all harness functionality (symbolizing stack dumps, etc.). You will still need to build the telemetry_gpu_integration_test target.

Useful command-line arguments:

  • --help: See more options and argument documentation.
  • -l: List all tests that would be run.
  • --test-filter: Filter tests.
  • --passthrough --show-stdout: Show browser output. See also --browser-logging-verbosity.
  • --extra-browser-args: Pass extra args to the browser executable.
  • --jobs=N: Run with multiple parallel browser instances.
  • --stable-jobs: Assign tests to each job in a stable order. Used on the bots for consistency and ease of reproduction.
  • --enable-dawn-backend-validation: Enable Dawn's backend validation.
  • --use-webgpu-adapter=[default,swiftshader,compat]: Forwarded to the browser to select a particular WebGPU adapter.

Running a local CTS build on Swarming

Often, it's useful to test changes on Chrome's infrastructure if it's difficult to reproduce a bug locally. To do that, we can package our local build as an "isolate" and upload it to Swarming to run there. This is often much faster than uploading your CL to Gerrit and triggering tryjobs.

Note that since you're doing a local build, you need to be on the same type of machine as the job you'd like to trigger in swarming. To run a job on a Windows bot, you need to build the isolate on Windows.

  1. Build the isolate

    vpython3 tools/mb/mb.py isolate out/Release telemetry_gpu_integration_test

  2. Upload the isolate

    ./tools/luci-go/isolate archive -cas-instance chromium-swarm -i out/Release/telemetry_gpu_integration_test.isolate

    This will output a hash like: 95199eb624d8ddb6ffdfe7a2fc41bc08573aebe3d17363a119cb1e9ca45761ae/734

    Save this hash for use in the next command.

  3. Trigger the swarming job.

    The command structure is as follows:

    ./tools/luci-go/swarming trigger -S https://chromium-swarm.appspot.com <dimensions...> -digest <YOUR_ISOLATE_HASH> -- <command> ...

    Say you want to trigger a job on a Linux Intel bot. It's easiest to check an existing task to see the right args you would use. For example: https://chromium-swarm.appspot.com/task?id=5d552b8def31ab11.

    In the table on the left hand side, you can see the bot's Dimensions. In the Raw Output on the right or below the table, you can see the commands run on this bot. Copying those, you would use:

    ./tools/luci-go/swarming trigger -S https://chromium-swarm.appspot.com -d "pool=chromium.tests.gpu" -d "cpu=x86-64" -d "gpu=8086:9bc5-20.0.8" -d "os=Ubuntu-18.04.6" -digest 95199eb624d8ddb6ffdfe7a2fc41bc08573aebe3d17363a119cb1e9ca45761ae/734 -- vpython3 testing/test_env.py testing/scripts/run_gpu_integration_test_as_googletest.py content/test/gpu/run_gpu_integration_test.py --isolated-script-test-output=${ISOLATED_OUTDIR}/output.json webgpu_cts --browser=release --passthrough -v --show-stdout --extra-browser-args="--enable-logging=stderr --js-flags=--expose-gc --force_high_performance_gpu --enable-features=Vulkan" --total-shards=14 --shard-index=0 --jobs=4 --stable-jobs

    The command will output a link to the Swarming task for you to see the results.