| Summary: | DIRT: Showdown broken graphics with Mesa built with -Ofast | ||
|---|---|---|---|
| Product: | Mesa | Reporter: | Gregor Münch <gr.muench> |
| Component: | Drivers/Gallium/radeonsi | Assignee: | Default DRI bug account <dri-devel> |
| Status: | RESOLVED NOTABUG | QA Contact: | Default DRI bug account <dri-devel> |
| Severity: | normal | ||
| Priority: | medium | ||
| Version: | git | ||
| Hardware: | x86-64 (AMD64) | ||
| OS: | Linux (All) | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
| Bug Depends on: | |||
| Bug Blocks: | 77449 | ||
| Attachments: | garbled graphics #1 | ||
|
Description
Gregor Münch
2016-11-25 17:17:07 UTC
Could you provide an apitrace that reproduces the issue? I tried getting this to work but I dont get any output file. DEBUGGER="apitrace trace" steam steam://rungameid/201700 What am Im doing wrong here? I installed: apitrace lib32-apitrace If you want to do a regression testing, this somewhat broke arround here: https://cgit.freedesktop.org/mesa/mesa/commit/?id=74e39de9324d2d2333cda6adca50ae2a3fc36de2 Does reverting that commit fix the issue? I managed to get a trace. Hope it works for you. https://drive.google.com/open?id=0B5KUmMDAwP-iUFNDdEpkd2xUMXc (In reply to Samuel Pitoiset from comment #3) > Does reverting that commit fix the issue? I can't test (I would do if someone could show me how). I do weekly Mesa builds and Im quite sure it broke in the week when this commit appeared. Problem is still there as of mesa git from Friday. Also, the trying replaying my trace hangs the GPU. I finally was able to start bisecting this, I was going back to Mesa from early October but couldnt find the culprit. I used PKGBUILD provided from here: https://aur.archlinux.org/packages/lib32-mesa-git/ Then I followed these instructions: https://wiki.archlinux.org/index.php/Arch_User_Repository#Installing_packages https://wiki.archlinux.org/index.php/Bisecting_bugs I guess since the executable name of the game is dirt.i386 so bisecting the 32bit package is the right one? Im a bit lost here, I will try older Kernel and libdrm but after that Im running out of ideas what could caused this, any other ideas would be appreciated. BTW, judging from the screenshots in https://bugs.freedesktop.org/show_bug.cgi?id=99136 it looks this is the same problem. Its interesting to see that enabling a minor heavy graphic option causes such crazy graphical corruption. He enables blood effects and in Dirt you have to set Crowd from OFF to LOW or higher. While ULTRA option makes the problem even worse. Fun Fact: I tested Tomb Raider Benchmark, and the average FPS on went from 48fps with Mesa from October to 58fps with current Mesa git. So really good job on this! Created attachment 130076 [details]
garbled graphics #1
Found the problem. Seems like using -Ofast wasnt a good idea. Using -03 everything works as it should.
Still attaching screenshot showing the problem in case it is useful to others.
Still the same with -Ofast. Ofast disregards strict standards compliance. -Ofast enables all -O3 optimizations. It also enables optimizations that are not valid for all standard-compliant programs. In other words don't use it with Mesa :) |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.