Bug 100089 - [regression] [bisected] Space Run rendering issues
Summary: [regression] [bisected] Space Run rendering issues
Status: RESOLVED MOVED
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/Gallium/radeonsi (show other bugs)
Version: 17.0
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Default DRI bug account
QA Contact: Default DRI bug account
URL:
Whiteboard:
Keywords:
Depends on:
Blocks: 77449
  Show dependency treegraph
 
Reported: 2017-03-06 21:44 UTC by LunarG
Modified: 2019-09-25 17:57 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
Screen shots of frames in Mesa 13.0.4 and Mesa 17 for Space Run (9.22 MB, application/zip)
2017-03-06 21:44 UTC, LunarG
Details

Description LunarG 2017-03-06 21:44:48 UTC
Created attachment 130093 [details]
Screen shots of frames in Mesa 13.0.4 and Mesa 17 for Space Run

Overview:
Space Run doesn't render smoke from destroyed asteroids, exhaust flames are incorrect, also it looks like "red arrows" aren't rendering. See included screen shots for reference.

OS:
Ubuntu 16.04

Steps to Reproduce:
Start the game and you should immediately see the exhaust flames out the ship are incorrect. Shoot an asteroid to see the problem with smoke that gets created when an asteroid is destroyed.

Expected Results:
Images are included from Mesa 13.0.4 that looks correct.
Comment 1 Kenneth Graunke 2017-03-08 20:16:45 UTC
What graphics card(s) are you seeing this issue on?
Comment 2 Kenneth Graunke 2017-03-08 20:18:24 UTC
Also, can you supply an apitrace?
Comment 3 LunarG 2017-03-13 16:12:58 UTC
This was observed on AMD R9 Fury and AMD RX480 graphic cards using Mesa 17.0.0 (git-cfdda19).

I do have a trace but it was not captured using this Mesa version and is a bit old but the problem can be observed using it. The problem was also observed when running the game in steam with this mesa 17.0.0 version. 

I've uploaded the trace; please let me know if you need one captured with this mesa version and I can make time to get that for you.
Comment 4 LunarG 2017-03-13 17:54:31 UTC
I'm having trouble adding the trace. I can't trim my current trace to the allowable upload size (including compressing it). I also can't create a new trace; apitrace gets an abort signal when trying to run the game with apitrace.

Let me see what I can do...
Comment 5 Samuel Pitoiset 2017-03-13 18:47:26 UTC
You might want to upload the trace with gdrive or dropbox.
Comment 6 LunarG 2017-03-15 21:45:16 UTC
Kenneth Graunke and Samuel Pitoiset should have the game now. I sent you guys an invite to our file share site. Let me know if you have any problems.
Comment 7 LunarG 2017-03-15 21:46:13 UTC
The invite should invite you to our file share site to get the trace.
Comment 8 Samuel Pitoiset 2017-03-15 22:12:51 UTC
Got the trace. The issue can be reproduced. I will investigate. Thanks
Comment 9 Samuel Pitoiset 2017-03-16 15:25:11 UTC
After bisecting, the guilty is:

commit 3081264dbeb5c322456b1b6b9011745ab98982a1
Author: Stanislav Mekhanoshin <Stanislav.Mekhanoshin@amd.com>
Date:   Sat Mar 11 00:29:27 2017 +0000

    [AMDGPU] Remove getBidirectionalReasonRank
    
    This method inverts the Reason field of a scheduling candidate.
    It does right comparison between RegCritical and RegExcess, but
    everything else is broken. In fact it can prefer less strong reason
    such as Weak over RegCritical because Weak > -RegCritical.
    
    The CandReason enum is properly sorted, so just remove artificial
    ranking.
    
    Differential Revision: https://reviews.llvm.org/D30557
    
    git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297536 91177308-0d34-0410-b5e6-96231b3b80d8

Presumably, it's a scheduling issue. The commit looks correct though.
Comment 10 GitLab Migration User 2019-09-25 17:57:56 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/mesa/mesa/issues/1259.


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.