A Mesa Fix Lands To Take Care Of The R9 290 Issue, Intel/Radeon Performance Problems

[phoronix]

Phoronix: A Mesa Fix Lands To Take Care Of The R9 290 Issue, Intel/Radeon Performance Problems

A fix landed in Mesa Git today that should address various performance issues people have been seeing in different rare setups. The fix mostly seems to be for Radeon/Intel users seeing low performance recently with glxgears but also appears to help those affected by the much talked about R9 290 regression...

http://www.phoronix.com/scan.php?pag...-Fix-Gears-290

[nuetzel]

Would someone who is hit by the R9 290 regression please verify that it is FIXED, finally?
Mesa git master dc3bb5d and above.

[Veerappan]

Would someone who is hit by the R9 290 regression please verify that it is FIXED, finally?
Mesa git master dc3bb5d and above.

The fd.o bug tracking it has several people claiming that this fixes their 290 performance.

I'd test myself, but the best card I have is an HD7850

[Saladien]

Sorry for my uninformed question. But how long until this hits for example Arch or Ubuntu? Ive just bought a 290X yesterday and it will arrive tomorrow. Would be great to know if i have to wait long . Thx

[Michael]

Sorry for my uninformed question. But how long until this hits for example Arch or Ubuntu? Ive just bought a 290X yesterday and it will arrive tomorrow. Would be great to know if i have to wait long . Thx

Ubuntu 16.10, unless you use Oibaf/Padoka PPAs.

[Mystro256]

Sorry for my uninformed question. But how long until this hits for example Arch or Ubuntu? Ive just bought a 290X yesterday and it will arrive tomorrow. Would be great to know if i have to wait long . Thx

As for Arch, probably within the week of the mesa 12.1 release (end of September I think).

[Saladien]

Thanks you two

[Veerappan]

Like Michael said, it should hit Ubuntu 16.10, and it's probable that the fixes will also be backported to Mesa 12.0.3. The original fix was queued for 12.0.2 until a regression was spotted by an automated testing system at Intel. Now that the Intel regression is probably fixed, and the 290 performance is back to higher levels, it's probable that the fix will be backported to the next 12.0 stable release.

So it's definitely likely to be in Ubuntu 16.10, and it'll probably make it to arch and friends pretty soon as well... Depends on when Mesa 12.0.3 is released, and if your distro of choice has Mesa 12.0 already.

If you're using a PPA (padoka/oibaf) in Ubuntu to get updated graphics drivers, you could probably have the fix today/tomorrow (or at least very soon). Arch has a similar package for pulling in git snapshots of mesa if you want bleeding edge drivers.

Note that once you start going down the route of building your drivers from the upstream development repository, it's hard to go back. Not from a technical standpoint, but mentally.

[elldekaa]

Note that once you start going down the route of building your drivers from the upstream development repository, it's hard to go back. Not from a technical standpoint, but mentally.

hahaha I like this comment

[duby229]

I applied that patch to my tree. Before:

$ vblank_mode=0 glxgears
ATTENTION: default value of option vblank_mode overridden by environment.
16922 frames in 5.0 seconds = 3384.292 FPS
17131 frames in 5.0 seconds = 3426.033 FPS
17176 frames in 5.0 seconds = 3435.125 FPS

After:

$ vblank_mode=0 glxgears
ATTENTION: default value of option vblank_mode overridden by environment.
26543 frames in 5.0 seconds = 5308.593 FPS
26631 frames in 5.0 seconds = 5326.078 FPS
26310 frames in 5.0 seconds = 5261.816 FPS
26708 frames in 5.0 seconds = 5341.584 FPS

I'm on an R9 390.

That really doesn't mean anything. The difference there could easily be explained by the window spawning in a different spot on screen. If you don't believe me, then run glxgears and move the window around, see for yourself. Just the simple matter of the window being in a different spot changes the framerate. glxgears is not a benchmark. It's not capable of showing performance differences. Please benchmark a game, that would actually tell any real difference.