[Bug 102334] New: [BXT]15% performance drop in SynMark v7.0 Multithread use-case
bugzilla-daemon at freedesktop.org
bugzilla-daemon at freedesktop.org
Mon Aug 21 12:46:13 UTC 2017
https://bugs.freedesktop.org/show_bug.cgi?id=102334
Bug ID: 102334
Summary: [BXT]15% performance drop in SynMark v7.0 Multithread
use-case
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
i915 platform: BXT
drm-tip kernel has caused ~15% drop in SynMark v7.0 Multithread 3D use-case
between these commits:
kernel git://anongit.freedesktop.org/drm-tip at
560582ef03d1f6bf9010232fcc6613188a00d198 2017-07-17_16-27-42 drm-tip:
2017y-07m-17d-16h-27m-16s UTC integration manifest
kernel git://anongit.freedesktop.org/drm-tip at
10de1e17faaab452782e5a1baffd1b30a639a261 2017-07-18_10-09-14 drm-tip:
2017y-07m-18d-10h-08m-42s UTC integration manifest
(This test has quite a large variance, but 15% drop in this period is very
clear from the longer term trend.)
This drop is BXT specific and visible on 2 different BXT (18 EU) J4205 devices.
Multithread test has 4 work threads doing low and hi resolution reflections /
environment map rendering, with main thread rendering the final HDR image with
tone mapping and bloom. This requires enough draw calls & syching that this
test is (mostly) CPU bound.
Note: The less CPU bound version (HdrBloom) of this test, which has only 1 work
thread instead of 4, and which uses higher resolution & details for that,
doesn't show any (clear) performance regression.
--
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20170821/defe6dd2/attachment.html>
More information about the intel-gfx-bugs
mailing list