NVIDIAGameWorks / NVIDIAGameWorks/rtx-remix

[Toolkit bug][HdRemix]: Unloading a stage does not free vram + issue when loading captures

Open
#484 5 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

bug hdremix jira runtime toolkit
Dominant language
No language data
Stars
1.7k
Forks
99
PR merge metrics
No merged PRs in 30d

Description

Contact Details (Optional)

No response

Describe the bug

From the latest changelog:
Added the option to unload the stage to reclaim resources without closing the app. This function is useful if you need to free up memory to do certain actions (for example, boot the game while running RTX Remix, or run RTX Remix’s AI Texture Tools). This option can be found in the top left menu.

Unloading a stage does not free my gpu's vram so there is no chance to load up the game or think about any texture upscaling without closing the toolkit first.

Loading a capture with the stage unloaded results in the project failing to load / throwing an error message.

https://github.com/NVIDIAGameWorks/rtx-remix/assets/45299104/56308722-ff32-4095-ae24-5518eaae7f42

How do you reproduce the bug?
  1. Unload a stage and watch vram usage
  2. Load a capture after stage unloading
What is the expected behavior?

The toolkit should free gpu vram.
Loading a capture after unloading a stage should reload the stage.

Maybe throw the user into the homescreen when unloading a stage so that he has to re-load the project from there

Version

2024.3.0

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Reproduce the two reported flows: unloading a stage while watching VRAM usage, then loading a capture with the stage unloaded. Trace the stage-unload and capture-loading entry points to determine why resources remain allocated and why loading fails. Done means unloading reclaims GPU VRAM and loading a capture reloads the stage or otherwise completes successfully.

Written by the indexing model from the issue text.

Assessment

Domain
computer-graphics, tooling
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
45/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.