Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add missing <algorithm> include #1

Merged
merged 1 commit into from
Feb 8, 2022
Merged

Conversation

sparist
Copy link
Collaborator

@sparist sparist commented Feb 4, 2022

DCP-2603

@sparist sparist requested a review from wesgarland February 4, 2022 04:44
@sparist sparist self-assigned this Feb 4, 2022
@sparist sparist assigned wesgarland and sparist and unassigned wesgarland and sparist Feb 8, 2022
@sparist sparist merged commit 07e7c49 into dcp/develop Feb 8, 2022
@sparist sparist deleted the dcp/tickets/DCP-2603 branch February 8, 2022 21:53
@sparist sparist restored the dcp/tickets/DCP-2603 branch February 9, 2022 00:06
@sparist sparist deleted the dcp/tickets/DCP-2603 branch February 9, 2022 00:11
sparist pushed a commit that referenced this pull request Oct 18, 2023
There are two problems with multiplanar textures in the OpenGL backend.

1) When the the multiplanar external texture transform runs, it splits
a texture_external variable into two texture_2d variables. The
Resolver then creates new texture/sampler pairs for the samplers with
which the plane1 texture is used, and new binding points for those
that bubble up to the global level. Since Dawn collects
sampler/texture pairs from the Inspector before the transforms run, it
doesn't know about these extra pairs, doesn't name them and doesn't
bind anything to them.

2) The multiplanar transform also inserts calls to
textureDimensions(), a texture-only builtin that doesn't require a
sampler. In the CombineSamplers transform, since OpenGLES doesn't have
the concept of a texture-only sampler2d, this necessitates the
creation of new binding points for the texture with a "placeholder"
sampler, which Dawn also doesn't know about, so it doesn't name them
or bind anything to them.

Fix #1 is on the Dawn side: when iterating over external texture
bindings, create a map from plane0 texture -> plane1 texture. When
iterating over the sampler/texture uses that the Tint inspector has
discovered, if the texture is an external texture, add a new pair with
the same sampler and the plane1 texture.

Fix for #2 is on the Tint side: if a function contains both a
texture-only reference and a texture/sampler reference to the same
texture, use the latter and remove the former. Texture-only builtins
won't use the sampler anyway. This is essentially an optimization, but
it's one we know will run on the functions generated by the
multiplanar transform, since it always calls textureDimensions() in a
function also containing the textureSample() call on the same texture.

Change-Id: I531ab09acdc47b435d72033aabd9eefa57f8f6e3
Bug: tint:1774
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/152660
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants