-
-
Notifications
You must be signed in to change notification settings - Fork 879
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
Performance is dramatically decreased After #1610 #1614
Comments
@nathanaelneveux must be the |
That does seem to be the case - I was able to replicate this by loading my test map x30 and then turning off and on the I'm not sure the At worse this shouldn't be too difficult to remake with a round trip through |
But |
It produces an |
Just to put some numbers behind it - using test map x30:
In all cases once the spriteBatch has been loaded images are not recreated - _generateAtlas is not called again after the initial load. |
Ok so I have a proof of concept I'd like to add caching to this final atlas image since I could see situations where a tileset would be reused and so shouldn't be re-generated as a flipped image but I think I'll need some kind of I did rely heavily on this project for converting the raw image bytes back into a |
Coincidentally I think I fixed what you need in #1624? |
Nice! That should work perfect |
Any progress on this issue @nathanaelneveux? |
Here is the branch I've been working in. I should have posted that sooner so you could track where I was at. My original proof of concept had a discoloring issue that took me an embarrassingly long time to realize that I was flipping the RGBA to ABGR. My most recent commit to the above branch passes all test and maintains performance but I'm not totally happy with it yet.
|
Can you test the updated PR @Hwan-seok? |
This sounds disturbing... What if you replace |
@st-pasha |
Fix for the performance decrease mentioned in #1614
Current bug behaviour
I was very looked forward to #1605 and I upgraded flame to this.
But I found the performance Is Very slow after upgrade. (see videos at the end).
Android can check the performance at least but IOS cannot even see the app launches because it shutting down immediately after launched due to memory issue.
So I bisected what commit went wrong.
After some reverts, I found #1610 is the problem.
It is very normal at #1596 (commit hash 60df3b9)
But after #1610 (b4ad498), It has dramatically slown performance.
I carefully assume that it has some memory leak and GCs constantly.
Expected behaviour
Fix the performance issue.
Flutter doctor output
More environment information
More information
Before #1610
Screen_Recording_20220510-143933_tagroom.mp4
After #1610 - I didn't manipulated video speed
Screen_Recording_20220510-144535_tagroom.mp4
The text was updated successfully, but these errors were encountered: