Perform bucket picking in web workers and other performance optimizations #3902
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
philippotto
commented
Mar 18, 2019
| @@ -67,7 +67,7 @@ export default function determineBucketsForFlight( | |||
| const cameraDirection = V3.sub(centerPosition, cameraPosition); | |||
| V3.scale(cameraDirection, 1 / Math.abs(V3.length(cameraDirection)), cameraDirection); | |||
|
|
|||
| const iterStep = 10; | |||
| const iterStep = 8; | |||
There was a problem hiding this comment.
I also tweaked the values for the arbitrary mode pickers a bit, since they are not that perf criticial anymore.
daniel-wer
approved these changes
Mar 19, 2019
There was a problem hiding this comment.
Very nice! I'm excited to measure performance on a deployed version :)
I like the clean design and tests for the latest_task_executor.js.
I do agree that the _refreshLookUpBuffer method is fairly complex and we should think about ways to improve that, but one has to note, that zooming in and out is super-smooth now, no flickering! 👍
frontend/javascripts/oxalis/model/bucket_data_handling/layer_rendering_manager.js
Outdated
Show resolved
Hide resolved
frontend/javascripts/oxalis/model/bucket_data_handling/layer_rendering_manager.js
Outdated
Show resolved
Hide resolved
…workerbucketpicking
hotzenklotz
added a commit
that referenced
this pull request
Mar 25, 2019
…ture-highlight * 'master' of github.com:scalableminds/webknossos: Hide unreported datasets (#3883) Update puppeteer and refresh screenshots (#3914) only show team names of own organization (#3928) Enable merger mode in skeleton and hybrid tracings (#3619) allow uploading nml for public dataset of different orga (#3929) Always make wheel listeners not passive to allow preventDefault (#3939) Enhance tree search functionallity (#3878) add webknossos-connect to setup (#3913) Update README.md (#3923) Add shortcut to maximize golden layout panes (#3927) Perform bucket picking in web workers and other performance optimizations (#3902) remove alt text for abstract brain loading image (#3930) updated documentation front page (#3917)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
TLDR:
A word abut the look up buffer for the TextureBucketManager:
The creation of that buffer is quite complex now and frankly a bit of a shit show. Since the bucket picking is done asynchronously, there are frames in which the TextureBucketManager doesn't know which buckets are supposed to be on the GPU. In that case, all available buckets should be put into the look up buffer, so that the shader can access them. I wrote a comment into the texture bucket which hopefully explains this. If it's unclear, hit me up. In the long-term I think we should do the addressing with cuckoo hashing or something similar.
URL of deployed dev instance (used for testing):
Steps to test:
Issues: