Intent to ship: WebGPU: Dual source blending
François Beaufort
Contact emails
Explainer
The “dual-source-blending” GPU feature adds additional blend factors and the WGSL @blend_src attribute to allow a fragment shader to blend two color outputs into a single output buffer. https://github.com/gpuweb/gpuweb/pull/4621
Specification
https://gpuweb.github.io/gpuweb/#dom-gpufeaturename-dual-source-blending
Summary
Functionality added to the WebGPU spec after its first shipment in a browser.
Adds the optional GPU feature “dual-source-blending” that enables combining two fragment shader outputs into a single framebuffer. This technique is particularly useful for applications that require complex blending operations, such as those based on Porter-Duff blend modes. By reducing the need for frequent pipeline state object changes, dual source blending can enhance performance and flexibility.
Blink component
TAG review
None
TAG review status
Not applicable
Risks
Interoperability and Compatibility
This feature has not yet been implemented in any browser. It has been approved by the GPU for the Web Community Group, with representatives from Chrome, Firefox, and Safari. See minutes at https://github.com/gpuweb/gpuweb/wiki/GPU-Web-2024-05-29#add-optional-feature-dual_source_blending-4621
Gecko: No signal (Mozilla members have approved https://github.com/gpuweb/gpuweb/pull/4621 and requested during weekly standardization meetings that we postpone filing standard positions until we reach Candidate Recommendation (CR) status in Q4.)
WebKit: Positive (https://github.com/WebKit/standards-positions/issues/294#issuecomment-1877411933)
Web developers: Positive (https://github.com/gpuweb/gpuweb/issues/4283)
Other signals:
WebView application risks
Does this intent deprecate or change behavior of existing APIs, such that it has potentially high risk for Android WebView-based applications?
None
Debuggability
None
Will this feature be supported on all six Blink platforms (Windows, Mac, Linux, ChromeOS, Android, and Android WebView)?
No
All platforms will eventually have support. Will immediately be available on Android, Android WebView, ChromeOS, Mac, and Windows, since those platforms already support WebGPU. Linux is planned to have WebGPU support in the future, so this feature will become available when WebGPU does.
Is this feature fully tested by web-platform-tests?
Yes. See https://github.com/gpuweb/cts/issues/3810
WebGPU/WGSL have a conformance test suite (https://github.com/gpuweb/cts) that is regularly pulled into Chromium and part of the testing of Dawn/Tint in Chromium.
Flag name on chrome://flags
None
Finch feature name
WebGPU.Enabled:UnsafeFeatures
Requires code in //chrome?
False
Tracking bug
https://issues.chromium.org/issues/341973423
Estimated milestones
Anticipated spec changes
Open questions about a feature may be a source of future web compat or interop issues. Please list open issues (e.g. links to known github issues in the project for the feature specification) whose resolution may introduce web compat/interop risk (e.g., changing to naming or structure of the API in a non-backward-compatible way).
None
Link to entry on the Chrome Platform Status
https://chromestatus.com/feature/5167711051841536
This intent message was generated by Chrome Platform Status.
Domenic Denicola
Contact emails
Explainer
The “dual-source-blending” GPU feature adds additional blend factors and the WGSL @blend_src attribute to allow a fragment shader to blend two color outputs into a single output buffer. https://github.com/gpuweb/gpuweb/pull/4621
Specification
https://gpuweb.github.io/gpuweb/#dom-gpufeaturename-dual-source-blending
Summary
Functionality added to the WebGPU spec after its first shipment in a browser.
Adds the optional GPU feature “dual-source-blending” that enables combining two fragment shader outputs into a single framebuffer. This technique is particularly useful for applications that require complex blending operations, such as those based on Porter-Duff blend modes. By reducing the need for frequent pipeline state object changes, dual source blending can enhance performance and flexibility.Blink component
TAG review
None
TAG review status
Not applicable
Risks
Interoperability and Compatibility
This feature has not yet been implemented in any browser. It has been approved by the GPU for the Web Community Group, with representatives from Chrome, Firefox, and Safari. See minutes at https://github.com/gpuweb/gpuweb/wiki/GPU-Web-2024-05-29#add-optional-feature-dual_source_blending-4621
Gecko: No signal (Mozilla members have approved https://github.com/gpuweb/gpuweb/pull/4621 and requested during weekly standardization meetings that we postpone filing standard positions until we reach Candidate Recommendation (CR) status in Q4.)
WebKit: Positive (https://github.com/WebKit/standards-positions/issues/294#issuecomment-1877411933)
Web developers: Positive (https://github.com/gpuweb/gpuweb/issues/4283)
Other signals:
WebView application risks
Does this intent deprecate or change behavior of existing APIs, such that it has potentially high risk for Android WebView-based applications?
None
Debuggability
None
Will this feature be supported on all six Blink platforms (Windows, Mac, Linux, ChromeOS, Android, and Android WebView)?
No
All platforms will eventually have support. Will immediately be available on Android, Android WebView, ChromeOS, Mac, and Windows, since those platforms already support WebGPU. Linux is planned to have WebGPU support in the future, so this feature will become available when WebGPU does.
Is this feature fully tested by web-platform-tests?
Yes. See https://github.com/gpuweb/cts/issues/3810
WebGPU/WGSL have a conformance test suite (https://github.com/gpuweb/cts) that is regularly pulled into Chromium and part of the testing of Dawn/Tint in Chromium.
Flag name on chrome://flags
None
Finch feature name
WebGPU.Enabled:UnsafeFeatures
Requires code in //chrome?
False
Tracking bug
https://issues.chromium.org/issues/341973423
Estimated milestones
Anticipated spec changes
Open questions about a feature may be a source of future web compat or interop issues. Please list open issues (e.g. links to known github issues in the project for the feature specification) whose resolution may introduce web compat/interop risk (e.g., changing to naming or structure of the API in a non-backward-compatible way).
None
Link to entry on the Chrome Platform Status
https://chromestatus.com/feature/5167711051841536
This intent message was generated by Chrome Platform Status.
--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAPpwU5LYcjs9iaFmzLZ%2BtorNvAQ0no4SSxGAm6ZJ62NONiASNA%40mail.gmail.com.
François Beaufort
On Thu, Aug 22, 2024 at 9:13 PM 'François Beaufort' via blink-dev <blin...@chromium.org> wrote:Contact emails
Explainer
The “dual-source-blending” GPU feature adds additional blend factors and the WGSL @blend_src attribute to allow a fragment shader to blend two color outputs into a single output buffer. https://github.com/gpuweb/gpuweb/pull/4621
Specification
https://gpuweb.github.io/gpuweb/#dom-gpufeaturename-dual-source-blending
Summary
Functionality added to the WebGPU spec after its first shipment in a browser.
Adds the optional GPU feature “dual-source-blending” that enables combining two fragment shader outputs into a single framebuffer. This technique is particularly useful for applications that require complex blending operations, such as those based on Porter-Duff blend modes. By reducing the need for frequent pipeline state object changes, dual source blending can enhance performance and flexibility.Blink component
TAG review
None
TAG review status
Not applicable
I can't figure out which exception this would fall under. Can you help explain why TAG review is not applicable here?
Risks
Interoperability and Compatibility
This feature has not yet been implemented in any browser. It has been approved by the GPU for the Web Community Group, with representatives from Chrome, Firefox, and Safari. See minutes at https://github.com/gpuweb/gpuweb/wiki/GPU-Web-2024-05-29#add-optional-feature-dual_source_blending-4621
Gecko: No signal (Mozilla members have approved https://github.com/gpuweb/gpuweb/pull/4621 and requested during weekly standardization meetings that we postpone filing standard positions until we reach Candidate Recommendation (CR) status in Q4.)
WebKit: Positive (https://github.com/WebKit/standards-positions/issues/294#issuecomment-1877411933)
Web developers: Positive (https://github.com/gpuweb/gpuweb/issues/4283)
Other signals:
WebView application risks
Does this intent deprecate or change behavior of existing APIs, such that it has potentially high risk for Android WebView-based applications?
None
Debuggability
None
Will this feature be supported on all six Blink platforms (Windows, Mac, Linux, ChromeOS, Android, and Android WebView)?
No
All platforms will eventually have support. Will immediately be available on Android, Android WebView, ChromeOS, Mac, and Windows, since those platforms already support WebGPU. Linux is planned to have WebGPU support in the future, so this feature will become available when WebGPU does.
Is this feature fully tested by web-platform-tests?
Yes. See https://github.com/gpuweb/cts/issues/3810
WebGPU/WGSL have a conformance test suite (https://github.com/gpuweb/cts) that is regularly pulled into Chromium and part of the testing of Dawn/Tint in Chromium.
But, it is not tested by web platform tests themselves, right? https://github.com/web-platform-tests/wpt/tree/master/webgpu seems to not contain any tests for this feature.(This does not block shipping, but rather is just a request to improve the accuracy of your Chrome Platform Status entry.)
As you can see in https://groups.google.com/a/chromium.org/g/blink-dev/search?q=cts%20web-platform-tests%20subject%3Awebgpu, we’ve always replied “Yes” to this question. The WebGPU/WGSL conformance test suite (https://github.com/gpuweb/cts) is not used only by Chromium. Other engines, like WebKit and Gecko, also rely on this test suite to ensure their implementation aligns with the WebGPU specification.
This test suite can be exported to WPT and we use that export for a subset of tests in Chromium for reftests: https://source.chromium.org/chromium/chromium/src/+/main:third_party/blink/web_tests/wpt_internal/webgpu/
The rest of the tests are executed with a different harness in Chromium for better speed and test result granularity.
Flag name on chrome://flags
None
Finch feature name
WebGPU.Enabled:UnsafeFeatures
This Finch feature name is a bit scary. Does it actually allow unshipping only this new feature? It sounds rather general.
If it really matters, we could rename the Finch feature name to WebGPU.Enabled:BlocklistFeatures to better reflect the meaning.
François Beaufort
Yoav Weiss (@Shopify)
Adding +Corentin Wallez to the threadOn Fri, Aug 23, 2024 at 2:29 PM François Beaufort <fbea...@google.com> wrote:On Fri, Aug 23, 2024 at 7:01 AM Domenic Denicola <dom...@chromium.org> wrote:On Thu, Aug 22, 2024 at 9:13 PM 'François Beaufort' via blink-dev <blin...@chromium.org> wrote:Contact emails
Explainer
The “dual-source-blending” GPU feature adds additional blend factors and the WGSL @blend_src attribute to allow a fragment shader to blend two color outputs into a single output buffer. https://github.com/gpuweb/gpuweb/pull/4621
Specification
https://gpuweb.github.io/gpuweb/#dom-gpufeaturename-dual-source-blending
Summary
Functionality added to the WebGPU spec after its first shipment in a browser.
Adds the optional GPU feature “dual-source-blending” that enables combining two fragment shader outputs into a single framebuffer. This technique is particularly useful for applications that require complex blending operations, such as those based on Porter-Duff blend modes. By reducing the need for frequent pipeline state object changes, dual source blending can enhance performance and flexibility.Blink component
TAG review
None
TAG review status
Not applicable
I can't figure out which exception this would fall under. Can you help explain why TAG review is not applicable here?Back in October 2023, Corentin Wallez, the WebGPU Tech Lead, met with the Blink API owners to address the process for releasing WebGPU features following their inclusion in Chromium. It was mutually agreed that requesting TAG reviews for minor features would be unnecessary. This policy applied to all features added at the time. See https://groups.google.com/a/chromium.org/g/blink-dev/search?q=%22tag%20review%22%20subject%3Awebgpu
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAPpwU5L5XPV4Z9amhPMP8819GR_WtDOS2pwN2iph_bLOKoCZoQ%40mail.gmail.com.
Corentin Wallez
Alex Russell
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAPpwU5LYcjs9iaFmzLZ%2BtorNvAQ0no4SSxGAm6ZJ62NONiASNA%40mail.gmail.com.
--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.
Daniel Bratell
LGTM3 (after two LGTM1s from Alex and Yoav).
The finch flag is just a followup on Domenic's question. It seems the issue is that the documented flag doesn't actually correspond to the on-off switch for this feature so please update the finch flag documentation to reflect the correct switch. Just in case this will need emergency surgery.
There is a discussion to be had about TAG reviews for minor WebGPU features. By the book this should probably have had one, but I'm not sure anyone would have benefited from such a request so it's probably the book that needs updating. Also see Alex' comment below.
/Daniel
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/ad26c92d-a28a-4ed2-9f95-17b02439f276n%40chromium.org.
Corentin Wallez
Finch flag
I think there is a misunderstanding about this Finch flag: the flag is meant to contain a list of comma-separated GPUFeatureNames that are disabled on the GPU process side (which is then reflected in Blink). So it is a general mechanism to disable any optional WebGPU feature, including, but not limited to dual source blending.
For the flag's name, we can easily rename it as no Finch config uses it at the moment. The WebGPU team called it "Unsafe" as it is meant to remotely disable features that we discover are unsafe for a reason or another after shipping. (think of a terrible in the wild exploit that we want to prevent immediately, or random crashes on end user machines).
TAG
We already had a discussion about reviews for minor WebGPU features, though it doesn't seem to mention the TAG (I misremembered earlier this thread) so I think we should discuss it now!
What do we expect the benefits of adding the TAG to the process of launching minor WebGPU features? The TAG review for WebGPU was excessively long and didn't give much actionable feedback. This is understandable given how large and how domain-specific WebGPU is, but likewise for 95% of WebGPU features we expect no actionable feedback, so we'd just be adding noise to the TAG's queue and wasting their time.
The WebGPU group (composed of Apple, Google, and Mozilla folks) is aiming to reach CR by the end of the year, and to pursue a living spec with regular (1-2 per year) CR snapshots afterwards. The process for CR triggers horizontal reviews which I assume would include the TAG? This way they would get a general FYI each CR snapshot. As noted earlier in the thread, we would proactively look for a TAG review on features that aren't trivial or completely domain-specific, where we believe there is a chance of actionable feedback. The vast majority of features are minor in this way, but as mentioned earlier we have several future features that we believe might be of interest to TAG.
Judging by the Github templates, the TAG has no mechanism for sending FYIs, only early (which this is not) and full reviews. A full TAG is a disproportionate amount of effort for minor features even if we made helpful templates for contributors for all the documents required. Any additional process involving the TAG would be gratuitous effort and friction for WebGPU contributors, but full reviews extremely so.
For the WebGPU team and contributors, adding the TAG to the launch process for every feature would add significant friction, when many contributors are already shying away from doing the I2S process due to the perceived amount of bureaucracy required (for example Francois filed this I2S on behalf of Jiawei).
In summary, WebGPU rolling CR snapshots, and proactive reviews with the TAG on the few topics that matter should be sufficient while keeping the process relatively lightweight.
Cheers,
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAPpwU5LYcjs9iaFmzLZ%2BtorNvAQ0no4SSxGAm6ZJ62NONiASNA%40mail.gmail.com.
--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAPpwU5L5XPV4Z9amhPMP8819GR_WtDOS2pwN2iph_bLOKoCZoQ%40mail.gmail.com.
Alex Russell
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAPpwU5LYcjs9iaFmzLZ%2BtorNvAQ0no4SSxGAm6ZJ62NONiASNA%40mail.gmail.com.
--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAPpwU5L5XPV4Z9amhPMP8819GR_WtDOS2pwN2iph_bLOKoCZoQ%40mail.gmail.com.
--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.
Corentin Wallez
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAPpwU5LYcjs9iaFmzLZ%2BtorNvAQ0no4SSxGAm6ZJ62NONiASNA%40mail.gmail.com.
--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAPpwU5L5XPV4Z9amhPMP8819GR_WtDOS2pwN2iph_bLOKoCZoQ%40mail.gmail.com.
--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.