* Ignore changes on non-main branches
* Increase the timeout for git cloning the repo
* Improve the error message on git error
Change-Id: I22af2b8042c14c3c413f52d65dc5b89b7f5c3f11
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/132820
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Parses the CL descriptions and adds missing hashtags to Gerrit changes.
Also add ./tools/push-to-gerrit which runs this after pushing the local branch's changes to 'main'.
Use this for the VSCode 'push' task.
Change-Id: I4c3f5982f6fdc7c1c6ebe770fc7811b1b38795d1
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/133061
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Ben Clayton <bclayton@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
A command that can time how long a command takes to execute for each file in a filepath glob.
Useful for finding the top-N shaders that take the longest to compile in a large corpus directory.
Change-Id: I416f300f7344480a939b9304bd3b49c378af9fef
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/132641
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
This adds a DAWN_FETCH_DEPENDENCIES CMake option (default: OFF) that if
turned ON fetches dependencies using a basic Python script that reads
DEPS files instead of requiring the installation of depot_tools by all
projects that depend on Dawn.
The script is not equivalent to gclient as it is not aware of advanced
configuration options (it always gets the list of dependencies that is
hardcoded at the beginning) but at least for a regular use of Dawn as a
dependency that provides a WebGPU backend (the intended use of this
DAWN_FETCH_DEPENDENCIES option) it is enough.
This is an attempt at merging upstream [this
repository](https://github.com/eliemichel/WebGPU-distribution/tree/dawn)
that I currently recommend in [Learn WebGPU for native
C++](https://eliemichel.github.io/LearnWebGPU/getting-started/hello-webgpu.html)
as a shallow intermediary to avoid requiring depot_tools (and turn some
options on/off).
By default, the script performs shallow clones, getting only the very
commit that is needed. This is a great improvement over naive clone:
| Regular | Shallow |
----------------|---------|---------|
Downloaded size | 194 MB | 15 MB |
Stored size | 294 MB | 103 MB |
Change-Id: Iaedb4af78036a1696b68787c36f1d3d70e18ba2c
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/131750
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
We don't have CTS results for the associated CI bot.
Change-Id: I3cfc62bd299e2c7bdb30fd0389f20159ab2f154e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/132482
Kokoro: Austin Eng <enga@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
* Bump timeouts
* Include changes description in results
* Use median times instead of means
* Use CPU time instead of real time
* Bump prority of the benchmark process
Change-Id: I40cb5635016fab4d6a1d1ee2434e94c120f7f6f5
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/121700
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Allow perfmon to benchmark shaders that are defined outside of the repo
Change-Id: Iee53b64cf357facdaa9941602d8d571611680097
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/128040
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
It's intended for the Unexpected Pass Finder to clean these up.
This should help prevent known flakes from causing retries.
Bug: chromium:1426512
Change-Id: I58c7baf2585dc4ca2e061e4a53c7b7d0f41e0af9
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/127420
Auto-Submit: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
Use the new `cmdline.ts --coverage` flag in https://github.com/gpuweb/cts/pull/2206 to write out the per-test coverage to a unique file.
Change-Id: Iea273217eede2fa615b78cd6f69f036976dcfd35
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/117883
Kokoro: Ben Clayton <bclayton@google.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
It appears that FXC and DXC do some validation post dead-code-elimination.
These tests have been updated so that the return value is assigned to a storage buffer, ensuring that all validation is performed.
Many DXC tests are affected by https://github.com/microsoft/DirectXShaderCompiler/issues/5082, which have been SKIP'ed.
Fixed: tint:1859
Change-Id: I0001a9a9821846cd0855c3d8ce2bec79ab8e64ef
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/122662
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: James Price <jrprice@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Emits the new OverloadFlag::kMustUse flag on the annotated overloads.
Nothing consumes this, yet.
Bug: tint:1844
Change-Id: I4eb0943a23eaf5de98cd63444a686cffe62fb36e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/120920
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
The default behaviour is a 1 minute timeout, which depending on
network load and other factors, may not be enough time to clone the
CTS repo when running the tool.
Increasing the timeout for this step to 10 minutes, since the tool is
expected to take a significant amount of time to run, so a bit of
extra time on setup isn't noticable.
Change-Id: If51ed154a8b2373bb9b246f29467fb17c7d7e345
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/119300
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Auto-Submit: Ryan Harrison <rharrison@chromium.org>
These will include the builtin language types.
Bug: tint:1810
Change-Id: I695a9ee833e1035eb1d17913d709038ae4c561d8
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/118502
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
The change to use `go run` over `go build` has upset the windows kokoro builds.
Use absolute paths.
Change-Id: Id9700529b3460a5ab579f82f616a95c3498c152c
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/118722
Kokoro: Kokoro <noreply+kokoro@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Commit-Queue: James Price <jrprice@google.com>
Reviewed-by: James Price <jrprice@google.com>
People keep on running the executable from the bin cache, wondering why stuff is out of date.
`go run` has pretty decent builtin caching, so just use that.
Change-Id: I6305a222f3f6b5b0b1ce23d7a89227aea96b9ef4
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/118720
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Ben Clayton <bclayton@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
* Use common_compile_options(turbo-cov) to fix RTTI linker errors on macOS + debug builds.
* Delete the .profraw files after they've been consumed.
* Include the test-case parameters in the coverage information
Change-Id: I02c6f408ed6e65c6f7e9eb1b4cd8443e4cfa7939
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/117882
Kokoro: Ben Clayton <bclayton@google.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
We weren't actually enabling WGSL validating in Tint in the E2E test
runner.
Mark a few tests as SKIP for cases that do not actually validate.
Change-Id: I62a55ce701f704d744c32f2cd5cb97683e270e96
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/117960
Commit-Queue: James Price <jrprice@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: James Price <jrprice@google.com>
Kokoro: Ben Clayton <bclayton@google.com>
Default to no-validation.
Change-Id: I4b3506e2addf580256175b13efbeb47f8ef96aaf
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/117881
Kokoro: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Common:
* The turbocov build target (somehow) never got hooked up to the root CMakeLists.txt file. This is now fixed.
macOS:
* Emit coverage for 'AppleClang' compiler.
* Have run-cts find the tucked-away `llvm-profdata` executable.
Windows:
* Various fixes for building with clang
* Fix turbocov stdout CRLF corruption
* Fix bad JSON with backslashes
Change-Id: I481cceafe2e72b544e13168172fc1456e5df2005
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/117880
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro: Ben Clayton <bclayton@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Can dramatically reduce latency for small test runs.
Requires: https://github.com/gpuweb/cts/pull/2167
Change-Id: Ib5ff5d0fbcaf151ffe67bf30e97c9ccdf34443e0
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/117200
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Ben Clayton <bclayton@google.com>
Enable coverage collection when using the test server, which is substantially faster than running in separate, isolated processes.
Use clang's `__llvm_profile_*` APIs to reset the counters between each test case run.
Change-Id: I01f8d0c1b3f215f66cfa50ef0fd51f2522c2ea57
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/113880
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
Add the "--adapter" flag to run-cts as a helper for setting this.
Make "--verbose" print the picked adapter.
Rename "dawn-backend" flag to just "backend" - this is already specific to the "gpu provider" (dawn).
Change-Id: Idc8d0eb3ccf5fa23325c06f0f9520aa9b528d9dd
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/116295
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Auto-Submit: Ben Clayton <bclayton@google.com>
Kokoro: Ben Clayton <bclayton@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
With the CTS global timeout extended, and tweaks to how const-eval loops are emitted, we now have a whole lot of passing tests.
Adjust the 'Slow' threshold to account for the new global timeout.
Cq-Include-Trybots: luci.chromium.try:dawn-try-win10-x86-rel,linux-dawn-rel,mac-dawn-rel,win-dawn-rel
Change-Id: I9a4d3278048e72e6560552164522395423dedf99
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/116296
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
These configurations are no longer used and replaced with newer
machines.
Bug: dawn:1387, dawn:1444
Change-Id: I13fcaf8fc5a42329cd74e9fd9a8ec9d10de63f68
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/116285
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Auto-Submit: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
The tooling does not expect, nor can handel a type matcher to be used as a parameter or a direct template parameter to another type. Make this an error.
Change-Id: I28c04c00c7fa1cb5130c53215a90a59d660d6fa3
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/114380
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Displays a per-test coverage viewer in your browser
Change-Id: I0b808bfadf01dab0540143760580cd7ca680e93b
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/113644
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
This is the standard place for tooling.
Update src/dawn/node/README.md with the new paths, and drop inferred
arguments from the examples.
Change-Id: Ib944ca045366b81b8897d9548112a8889e097769
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/113643
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
A library for dealing with coverage information, and compressing CTS test information down into something that can be easily downloaded and visualized in a web page
Change-Id: If025e1555665609f7c963bad14c9c91d8324a390
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/113642
Reviewed-by: Austin Eng <enga@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
A tiny utility that emits just the per segment coverage in a binary stream. This avoids the overhead of encoding to JSON, which provides substantial performance improvements.
Change-Id: I36a588069d69c5c800d31bca8dd5c542bcdbe313
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/113641
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
appendConsumedResultsForSkippedTests() adds fake results for expectations that are 'Skip'ed, so that these sub-trees aren't collapsed as all passing.
However, this code was not handling the fact that there might actually be results for the cases.
This happened because there was an expectation collision in the expectations.txt file, and tests were not being skipped even though there was an expectation with a Skip (a collision expectation was used instead).
Change-Id: I2a1543d231db44dc8aa6683d051f884f4cb96853
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/113520
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Ben Clayton <bclayton@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
10x speed improvement on my 3990x machine
Change-Id: Ie4b2377719e9a25860ec05bb7649aa652d26ccf0
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/113022
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Ben Clayton <bclayton@google.com>
Can be run with:
./tools/run git-stats
Change-Id: Ie46be01ff3318ff50acdb13eb809d5304f867f04
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/111180
Auto-Submit: Ben Clayton <bclayton@google.com>
Commit-Queue: David Neto <dneto@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
A simple tool to benchmark tint based on a template file.
Bug: tint:1122
Change-Id: I34ee0fb98e6d2187c145cc38ec3cb48606242cb4
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/110820
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
And add tests.
This is useful for other tooling.
Change-Id: Ia399071baf6d4bb617f3c73e4ccd4ed72d522c2e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/111020
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Will be used by GN to know the outputs of the generate cache step.
Change-Id: Iae09477fb355eac41ff9b2204605f46ed78e08ec
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/110620
Reviewed-by: Austin Eng <enga@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>