Skip to content

[MLA-1743] Backport inference GC Optimizations #4916

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

Merged
merged 4 commits into from
Feb 8, 2021

Conversation

chriselion
Copy link
Contributor

@chriselion chriselion commented Feb 6, 2021

Proposed change(s)

Backport most of the changes from:
#4887

Exceptions:

  • Had to make ObservationWriter.AddList internal visibility (since it's a new API)
  • Couldn't add VectorSensor ILIst overload (since it's a new API)
  • Didn't change timer profile strings (not sure if that's too visible of a change)

On a test scene (3DBall with decisions made every frame), the GC Allocations drop from 151.8KB to 135.5KB.

Useful links (Github issues, JIRA tickets, ML-Agents forum threads etc.)

https://jira.unity3d.com/browse/MLA-1743 (backport)
https://jira.unity3d.com/browse/MLA-1724 (original)

Types of change(s)

  • Bug fix

Checklist

@chriselion chriselion changed the base branch from master to release_2_verified February 6, 2021 00:31
@chriselion chriselion requested a review from surfnerd February 6, 2021 00:32
@chriselion chriselion changed the title R2v backport mla 1724 ienumerable gc [MLA-1743] Backport inference GC Optimizations Feb 6, 2021
@chriselion chriselion merged commit 69955f8 into release_2_verified Feb 8, 2021
@delete-merged-branch delete-merged-branch bot deleted the r2v-backport-MLA-1724-ienumerable-gc branch February 8, 2021 18:00
@github-actions github-actions bot locked as resolved and limited conversation to collaborators Feb 8, 2022
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants