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.

Sign up for GitHub

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

Jump to bottom

Allow user to adapt GPU memory consumption to allow for better quality #4015

Merged
merged 10 commits into from
Apr 23, 2019
Merged

Allow user to adapt GPU memory consumption to allow for better quality #4015

merged 10 commits into from
Apr 23, 2019

Conversation

philippotto
Copy link
Member

@philippotto philippotto commented Apr 12, 2019 •
edited
Loading

  • the user can select how much gpu memory should be used (from low to very high)
  • the actual setting is a "gpu factor" which is multiplied with 512 internally to get the amount of buckets which should fit on to the gpu (per layer). this means, we can relabel the actual settings (e.g., "medium" becomes "low") without having to migrate the user settings.

todo:

  • the setting is a user setting right now, however from the UI it belongs better to the dataset rendering settings 🤔 not sure, how we should handle this. the "tracing" tab has a weird name, anyway, since this tab is also shown in view mode.
  • ensure that setting can be changed even when webgl crashed
  • ensure that setting cannot be overridden by recommended task settings
  • wait for lab feedback

URL of deployed dev instance (used for testing):

Steps to test:

  • render datasets in different window sizes/layout configs/qualities

Issues:

@philippotto philippotto self-assigned this Apr 12, 2019
@philippotto philippotto requested a review from daniel-wer April 15, 2019 07:19
daniel-wer
daniel-wer approved these changes Apr 17, 2019
View reviewed changes
Copy link
Member

@daniel-wer daniel-wer left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nice, code looks good! Can you add a ToDo or an issue to remove all the "quality" related code once we're sure it's no longer needed.

I'll do a final test once the dev server is back up, initial testing looked very good :)

frontend/javascripts/oxalis/model_initialization.js Outdated
constants.GPU_FACTOR_MULTIPLIER *
(gpuFactor != null ? gpuFactor : constants.DEFAULT_GPU_MEMORY_FACTOR);

console.log("demanding requiredBucketCapacity:", requiredBucketCapacity);
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not sure whether you want to keep those two console.logs

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I removed one and kept the other so that we potentially have some hardware information available in airbrake.

@philippotto philippotto merged commit 3e423c1 into master Apr 23, 2019
hotzenklotz added a commit that referenced this pull request Apr 25, 2019
@hotzenklotz
Merge branch 'master' into mobile
f9bd0f4 
* master: (43 commits)
  update screenshots (#4038)
  docker: don't set sbt/ivy cache (#4034)
  Slow down brush size change for small sizes (#4012)
  voulme tool bar now depends only on the active tool (#4029)
  Allow user to adapt GPU memory consumption to allow for better quality (#4015)
  Updates the Docs (#4020)
  hiding pricing and features in iframe (#4019)
  Fix sitemap for publication detail view (#4024)
  Add sitemap.xml (#4006)
  Allow empty trees  (#4010)
  Prefer annotation zoom over dataset zoom (#3992)
  Add details view for publications (#3994)
  remove enzyme test (#3997)
  Tensorflow segmentation (#3461)
  ensure that max zoom step is not exceeded when changing viewport area (#3996)
  allow isosurfaces for hybrid tracings when setting window.allowIsosurfaces (#3998)
  re-fetch buckets if fetching them failed (#3999)
  Only re-compute bounding sphere of skeletons if something changed (#3995)
  Fix dropdown login padding (#3988)
  Persist tree visibility (#3942)
  ...
@philippotto philippotto deleted the custom-gpu-mem branch May 20, 2019 09:11
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@daniel-wer daniel-wer daniel-wer approved these changes

Assignees

@philippotto philippotto

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

Maximum address space is not properly respected during calculation of max zoom step Render higher quality
2 participants
@philippotto @daniel-wer