Adaptive Physics Time Step #989
Merged
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
HunterBarclay
added
the
physics
HunterBarclay
requested review from
PepperLola and
LucaHaverty
and removed request for
a team
LucaHaverty
approved these changes
Jun 26, 2024
|
@LucaHaverty @PepperLola Fixed prettier and lint validation and fixing. Updated workflow to check them separately. |
LucaHaverty
reviewed
Jun 28, 2024
|
@LucaHaverty @PepperLola It looks like the number of substeps is decreasing as the frame rate drops which isn't right. I'll take a look. |
|
K, I was actually shocked at the fidelity with this last change. I couldn't see any significant joint drift |
PepperLola
approved these changes
Jun 29, 2024
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Outdated suggestions cannot be applied.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.
Description
Modulate the physics time step to no longer be fixed, and instead slightly adjust it towards the actual DeltaT. During the original Spike of JoltPhysics, we discovered the significant impact that adjusting the physics delta can have. Ideally we'd have a fixed, consistent time step; however, this is proving difficult as we would like to accommodate the varying performance levels.
Warning
This is a bandage. We may want to look into better ways of smoothing out the physics step.
We should also investigate allowing advance options to manual set/configure physics time steps.
Adjustment Method
We initially set a physics step rate of 120FPS. We then calculate the difference between the last used deltaT and the current frames recorded deltaT. This change is capped out to a constant and the targeted deltaT is then adjusted by this capped difference. We then calculate number of sub steps via the inverse of this deltaT. This way we can try to make up for lack of frames per second with more physical calculation.
Important
All of these values are limited with constants that can be found in
PhysicsSystem.tsDownsides
Sub steps can't seem to be calculated to a good value that doesn't flip between two adjacent numbers, at higher frame rates. In addition, this inconsistent deltaT generally decays the fidelity of the simulation.
JIRA Issue