Search Issue Tracker
Won't Fix
Votes
1
Found in [Package]
7.1.1
Issue ID
1189015
Regression
Yes
URP 7.1+ build times have regressed, might be as high as 10x longer
It seems build times have regressed between SRP 7.0.2 and 7.1.1
Note: checking can take a long time due to this slow build issue
Steps to repro:
- download and open project
- delete Library/ShaderCache
- select only first 30 scenes in build settings
- build project
- observe time taken is >30 mins
- change SRP 7.0.2 (not available in package manager? Get this from github)
- delete Library/ShaderCache
- build project
- observe time taken is <5 mins
Expected: similar or same time to build
Actual: much longer when using SRP 7.1.1
Reproduces: 2019.3.0b4 (no other versions checked because it is a package regression)
-
nasos_333
Oct 12, 2020 13:39
Hi,
I have an issue with URP project, i have multiple shaders and scripts and after a while the project started to go very slow when I do any changes to any shader, it takes around 1.5 minutes for each little change, making it impossible to work with fast iterations after my changes, is this something similar to this issue ?
Thanks
Add comment
All about bugs
View bugs we have successfully reproduced, and vote for the bugs you want to see fixed most urgently.
Latest issues
- Crash on ShaderLab::ShaderState::ApplyShaderState when entering Play Mode and calling ShaderWarmup.WarmupShaderFromCollection with a shader that has dynamic_branch
- Crash on operator delete[] when a serializable structure in the System namespace is serialized by a MonoBehaviour
- [macOS] forward and backward mouse buttons are recognized as middle mouse button
- NavMeshModifier.OnEnable/OnDisable gets progressively slower when called repeatedly
- "System.Net.CookieContainer" throws an "ArgumentException" error when using a non-UTF-8 machine locale and hostname
Resolution Note:
This is caused by FXC compiler taking longer based on the size of the uniform buffers. Since we increased light limits, the compilation time has increased. There's an workaround to trick FXC compiler that is to declare each uniform buffer as an array of 2 inside a cbuffer. However there's too much risk to take this approach given how much uncertainty there is and the amount of platforms. i.e, behavior could change in different version of the FXC compiler, or we can see different behavior on different platform. Based on this and the fact that we already have work in progress to optimize shader compilation time we decided we will not fix the cause of this regression.