Intel KMS vs. UMS With Ubuntu 10.04

[phoronix]

Phoronix: Intel KMS vs. UMS With Ubuntu 10.04

Last week we published benchmarks looking at the ATI Radeon KMS vs. UMS performance and found the user-space mode-setting support with the ATI driver (that is also limited to using DRI1 with these older code-paths) to perform significantly faster than the newer kernel mode-setting routes in most instances. To see how the performance difference is on the Intel side between the kernel mode-setting and user-space mode-setting implementations we ran a set of benchmarks on this side as well using Ubuntu 10.04.

http://www.phoronix.com/vr.php?view=14706

[darkbasic]

xf86-video-intel-2.11 should be much more faster thanks to the page-flipping ioctl.

[droidhacker]

So.... garbage in garbage out?

[avilella]

great article. more, please!

[droidhacker]

"ATI Radeon KMS vs. UMS performance and found the user-space mode-setting support with the ATI driver (that is also limited to using DRI1 with these older code-paths) to perform significantly faster than the newer kernel mode-setting routes in most instances."
That is fairly inconsistent with the actual results shown. What was actually shown was that under some circumstances, UMS was slightly faster than KMA, and under other circumstances, KMA was SIGNIFICANTLY FASTER than UMS. KMS also showed a MUCH MORE CONSISTENT performance than UMS.

[FireBurn]

Seeing as Intel doesn't support UMS any more with their new driver and Ubuntu backported the updates from the latest driver back to the last supported UMS Intel driver I'm not sure if these results have any relevance

Any one running Intel's code "the way it's intended" won't have the option for UMS. Only Ubuntu's bastardised version has the feature

[glasen]

The Ubuntu driver is a vanilla 2.9.1 version plus a small patch which allows a flicker free transitition from Plymouth to GDM ("101_copy-fb.patch").

I don't think there will be many bugfixes or backports from driver version 2.10 and above, because with version 2.10 many things have changed (no more UMS, code cleanup, function renaming, etc.) and it will be very hard to backport these changes.

[garytr24]

Any one running Intel's code "the way it's intended" won't have the option for UMS. Only Ubuntu's bastardised version has the feature

I think you meant the X server is bastardized?

[FireBurn]

I think you meant the X server is bastardized?

I think they've played with the kernel drm, libdrm, xorg-server as well as the drivers

[KDesk]

Would be interesting to aditionaly know the KMS performance of the vanilla intel 2.9.1, the 2.10, and the 2.11 rc.