[ubuntu-x] Test kernel for perf issues on -intel
michael.larabel at phoronix.com
Thu Apr 16 03:10:25 BST 2009
Michael Larabel wrote:
> Bryce Harrington wrote:
>> Another option which has been suggested is to turn off migration
>> Option "EXAOptimizeMigration" "off"
>> This may have the same "benefit" as the greedy migration, but without
>> the irritating side effects. It would be hugely helpful if people could
>> test this option (independently from the kernel change) and see if they
>> spot any regressions with it set. Meanwhile I'll try to get some debs
>> put together for it.
>> On Wed, Apr 15, 2009 at 11:54:43AM -0700, Bryce Harrington wrote:
>>> Tim pointed me to this:
>>> bryce: 2.6.30-rc2 has a bunch of i915 patches from Dave Arlie. He
>>> big performance improvements which likely address out issues in
>>> Jaunty. Is it possible for you to run that kernel somewhere and test
>>> If anyone has hardware that they're seeing the -intel performance
>>> regression and would be willing to test the above, please do so and
>>> report results here.
>>> (It would be especially helpful if you could run a 3d game that prints
>>> fps, to measure both before and after.)
>>> Ubuntu-x mailing list
>>> Ubuntu-x at lists.ubuntu.com
>>> Modify settings or unsubscribe at:
> Attached are results using the EXAOptimizeMigration off thrown into
> the mix. This option isn't definitively faster, but there seems to be
> a few areas where it's ahead.
Attached are the results with the Linux 2.6.30-rc2 kernel being used
with EXA and no options being set.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 712505 bytes
Desc: not available
Url : https://lists.ubuntu.com/archives/ubuntu-x/attachments/20090415/ddf7132f/attachment-0001.pdf
More information about the Ubuntu-x