[Bug 110948] New: Recoverable GPU (rcs0) hangs in CPU bound 3D benchmarks

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Wed Jun 19 11:04:44 UTC 2019


https://bugs.freedesktop.org/show_bug.cgi?id=110948

            Bug ID: 110948
           Summary: Recoverable GPU (rcs0) hangs in CPU bound 3D
                    benchmarks
           Product: DRI
           Version: DRI git
          Hardware: Other
                OS: All
            Status: NEW
          Severity: normal
          Priority: medium
         Component: DRM/Intel
          Assignee: intel-gfx-bugs at lists.freedesktop.org
          Reporter: eero.t.tamminen at intel.com
        QA Contact: intel-gfx-bugs at lists.freedesktop.org
                CC: intel-gfx-bugs at lists.freedesktop.org

Created attachment 144591
  --> https://bugs.freedesktop.org/attachment.cgi?id=144591&action=edit
BXT i915 error state

Setup:
* HW: BXT J4205, SKL GT2, KBL GT3e
* OS: Ubuntu 18.04
* kernel: drm-tip compiled from git
* GUI: Unity, using Mesa & Xorg compiled from git

Test-case:
1. Run GfxBench (v5) Driver2 offscreen benchmarks:
   testfw_app --gfx glfw --gl_api desktop_core --width 1920 --height 1080
--fullscreen 1 --test_id gl_driver2_off
2. Run SynMark2 (v7) Batch5-7 tests
   $ for t in 5 6 7; do synmark2 -homeDir $PWD OglBatch$t; done

Expected outcome:
* No GPU hangs, performance of each test-run is about same

Actual outcome:
* Every 5-10th run there's recoverable GPU hang and clearly lower perf

First time I saw this 15th of June.  So the regression may have come between
following drm-tip commits:
* a0195925ba at 2019-06-14 16:19:08: drm-tip: 2019y-06m-14d-16h-18m-09s UTC
integration manifest
* 39e3d39be3 at 2019-06-15 15:21:37: drm-tip: 2019y-06m-15d-15h-20m-54s UTC
integration manifest

Notes:
* These are odd tests to be GPU hanging, as GPU functionality-wise they're
simple and more CPU than GPU bound

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20190619/06c9e70e/attachment.html>


More information about the intel-gfx-bugs mailing list