Search Issue Tracker
Fixed in 2019.2.X
Votes
0
Found in
2017.4.1f1
Issue ID
1035800
Regression
No
[iOS] Higher memory consumption in Load/Store Bandwidth frame using commandBuffer.Blit()
To reproduce:
1. Open project
2. Build to XCode and runs in a Metal iOS Device.
3. Capture a frame to see only one RenderCommandEncoder generated. Frame Load&Store Bandwidth is around 8MB
4. Open CameraController.cs in the editor and Uncomment Lines of 71 - 76, which adds the blit operation to the commandbuffer
5. Build project to XCode and runs in a Metal iOS Device.
6. Capture a frame to see there are 3 RenderCommandEncoders generated(expected 1). Frame Load&Store Bandwidth is around 80MB
Reproduced: 2017.1.3p4, 2017.2.2p4, 2017.3.2f1, 2017.4.3f1, 2018.1.0f2, 2018.2.0a6
Not reproducible: 2018.2.0a8, 2018.2.0b4
Note:
- fixed in 2018.2 between a6-a8
- tested with:
- iPhone 6S Plus iOS 10.3.2 (Frame Load&Store Bandwidth in 6th step is around 80MB)
- iPhone 7 iOS 11.0 (Frame Load&Store Bandwidth in 6th step is around 35-40MB)
All about bugs
View bugs we have successfully reproduced, and vote for the bugs you want to see fixed most urgently.
Latest issues
- UI Text color is incorrect when text is rendered on SubMeshUI
- "Invalid SortingGroup index set in Renderer" errors when making many modifications in same frame
- Crash on GizmoBufferSizeHistory::GetBuffer() when opening a specific scene
- RenderTexture is not initialized when it is set as Material's Texture
- Branch node outputs a random value instead of a constant on the false conditon
Add comment