bugzilla-daemon at freedesktop.org
2012-Jul-08 12:03 UTC
[Nouveau] [Bug 51862] New: Shader optimizer is extremely slow
https://bugs.freedesktop.org/show_bug.cgi?id=51862
Bug #: 51862
Summary: Shader optimizer is extremely slow
Classification: Unclassified
Product: Mesa
Version: git
Platform: x86-64 (AMD64)
OS/Version: Linux (All)
Status: NEW
Severity: critical
Priority: medium
Component: Drivers/DRI/nouveau
AssignedTo: nouveau at lists.freedesktop.org
ReportedBy: tiziano at tizbac.dyndns.org
Created attachment 63969
--> https://bugs.freedesktop.org/attachment.cgi?id=63969
Comparison of shader compile times between mesa and nvidia driver
Shader optimizer is extremely slow and causes games like Crysis that loads
content while playing to go very slow or freeze temporarily
On nvidia proprietary driver sometimes compiling is even 200x faster
--
Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
bugzilla-daemon at freedesktop.org
2012-Jul-08 12:05 UTC
[Nouveau] [Bug 51862] Shader optimizer is extremely slow
https://bugs.freedesktop.org/show_bug.cgi?id=51862 --- Comment #1 from Tiziano Bacocco <tiziano at tizbac.dyndns.org> 2012-07-08 05:05:09 PDT --- Created attachment 63970 --> https://bugs.freedesktop.org/attachment.cgi?id=63970 Profiler output while running a programm that continously compiles shaders -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug.
bugzilla-daemon at freedesktop.org
2013-Aug-21 06:39 UTC
[Nouveau] [Bug 51862] Shader optimizer is extremely slow
https://bugs.freedesktop.org/show_bug.cgi?id=51862
Ilia Mirkin <imirkin at alum.mit.edu> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution|--- |NOTOURBUG
--- Comment #2 from Ilia Mirkin <imirkin at alum.mit.edu> ---
Looks like most of the time is being spent in ir_expression/etc -- these are
generic mesa glsl facilities. Please retest this with the latest mesa, I
suspect that improvements have been made. If you still think it's slow, feel
free to file a bug with the core mesa team. I don't see anything
nouveau-specific about this issue. Also, please use perf for the measurements,
it can see inside the kernel, which may relevant (that 16% chunk inside the
kernel seems high).
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20130821/6f08f55f/attachment.html>