OK thanks - it's just at the moment it's not clear if extensions aren't supported due to missing implementation (so it won't work anywhere), or an unsupported system (in which case I can potentially get it working and test my code if I find the right system setup). Extra visibility on that would be good.
- Ashley On Thu, 12 Jan 2023 at 12:21, Corentin Wallez <[email protected]> wrote: > Hey Ashley, > > Happy to talk in more detail outside of the I2S thread. In general WebGPU > optional features are optional as they can depend on hardware capabilities. > The only guarantee is that around texture compression where the following > must be true: BC || (ETC2 && ASTC). For shader-float16 for example, we are > planning to have it implemented around the V1 timeframe, but it can't be > universal as it depends on hardware capabilities and platform compiler > support (which is the blocking part of your setup). For F16 one solution to > bridge availability is to use type aliases: you can do `type maybe_f16 = > f16;` or the same with f32 and use `maybe_f16` everywhere. > > Hopefully this helps, but in general after V1 we should have a clear > roadmap to give visibility to the ecosystem. > > Cheers, > > Corentin > > On Thu, Jan 12, 2023 at 11:55 AM Ashley Gullen <[email protected]> wrote: > >> Can I ask what the release schedule is for the various WebGPU extensions >> as well? It's not currently clear to me which extensions are planned to be >> shipped with the first release of WebGPU and which are to be implemented >> later down the line. >> >> As of today's Canary 111.0.5534.0 on Windows 11 with a nVidia GeForce GTX >> 1070, I can see the following adapter features: indirect-first-instance, >> texture-compression-bc, depth32float-stencil8, rg11b10ufloat-renderable >> However the current spec's feature index lists 5 other extensions, >> notably including shader-f16 and timestamp-query. I'm particularly >> interested in the status of shader-f16, as our product allows for >> third-party shaders, and I want to avoid any backwards compatibility >> complications if it's introduced later down the line. But I'm also not >> clear if the OS/hardware combo is meant to support this too. >> >> As a WebGPU developer it would be useful to see official documentation >> somewhere of the individual development/shipping status for each WebGPU >> extension. >> >> Thanks for all the great work on WebGPU! >> >> >> On Wed, 11 Jan 2023 at 17:57, Corentin Wallez <[email protected]> >> wrote: >> >>> Thank you everyone for the LGTMs! >>> >>> We're super excited to get a high-quality WebGPU implementation in >>> developer's hands! The list of issues we are tracking on the implementation >>> side are >>> <https://bugs.chromium.org/p/dawn/issues/list?q=Milestone%3DV1&can=2> >>> these >>> <https://bugs.chromium.org/p/tint/issues/list?q=Milestone%3DV1&can=2> >>> three >>> <https://bugs.chromium.org/p/chromium/issues/list?q=Proj%3DWebGPU-Milestone-V1&can=2> >>> (for >>> the three parts that make WebGPU) and we expect to fix most issues before >>> shipping, with clear rational of why the ones that aren't can be delayed. >>> Of course we expect to find new issues as we keep expanding the test suite >>> but will add them to the bug trackers I just linked for V1 tracking as well. >>> >>> On Wednesday, January 11, 2023 at 6:18:17 PM UTC+1 Yoav Weiss wrote: >>> >>>> LGTM3 >>>> >>>> On Wed, Jan 11, 2023 at 6:06 PM Rick Byers <[email protected]> wrote: >>>> >>>>> We discussed in the API owner meeting and agree that we are impressed >>>>> with all the work that's gone into interoperability here and the broad >>>>> general support across implementers. We have high confidence that the team >>>>> is setting an appropriately high quality bar both for shipping and for >>>>> ongoing updates and interoperability improvements post-ship. So LGTM2 to >>>>> ship whenever the team feels it's ready. >>>>> >>>>> This feature has been a long time in the making and I'm excited to see >>>>> it launch ASAP, great work! >>>>> >>>>> Rick >>>>> >>>>> On Wed, Jan 11, 2023 at 11:52 AM slightlyoff via Chromestatus < >>>>> [email protected]> wrote: >>>>> >>>>>> An enthusiastic LGTM1 from me. Congrats on getting to this point; a >>>>>> big moment. >>>>>> >>>>>> -- >>>>>> 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 [email protected]. >>>>>> To view this discussion on the web visit >>>>>> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/0000000000004c1db705f1ffd237%40google.com >>>>>> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/0000000000004c1db705f1ffd237%40google.com?utm_medium=email&utm_source=footer> >>>>>> . >>>>>> >>>>> -- >>>>> 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 [email protected]. >>>>> >>>> To view this discussion on the web visit >>>>> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAFUtAY98ckLv_13k9sBxQfDC%3DtDyqPA9_hLOMiVG4Fn4%2B9njwQ%40mail.gmail.com >>>>> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAFUtAY98ckLv_13k9sBxQfDC%3DtDyqPA9_hLOMiVG4Fn4%2B9njwQ%40mail.gmail.com?utm_medium=email&utm_source=footer> >>>>> . >>>>> >>>> -- >>> 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 [email protected]. >>> To view this discussion on the web visit >>> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/7e739536-1bef-4bee-a44a-d2350ab79f3fn%40chromium.org >>> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/7e739536-1bef-4bee-a44a-d2350ab79f3fn%40chromium.org?utm_medium=email&utm_source=footer> >>> . >>> >> -- 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 [email protected]. To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAABs73h_PdGHfAMNVVO9Z0Bj71kjBKd2SkC77v8h63sNTFHrCQ%40mail.gmail.com.
