Radeon DRM GPU Driver Performance: Linux 3.4 To 3.8

[phoronix]

Phoronix: Radeon DRM GPU Driver Performance: Linux 3.4 To 3.8

Last week benchmarks were delivered on Phoronix that showed the Intel DRM GPU driver performance between the Linux 3.2 and 3.8 kernels. In this article are similar benchmarks of the Radeon DRM driver in recent kernel releases using AMD graphics hardware but going back only to the Linux 3.4 kernel due to show-stopping issues.

http://www.phoronix.com/vr.php?view=18437

[glisse]

Phoronix: Radeon DRM GPU Driver Performance: Linux 3.4 To 3.8

Last week benchmarks were delivered on Phoronix that showed the Intel DRM GPU driver performance between the Linux 3.2 and 3.8 kernels. In this article are similar benchmarks of the Radeon DRM driver in recent kernel releases using AMD graphics hardware but going back only to the Linux 3.4 kernel due to show-stopping issues.

http://www.phoronix.com/vr.php?view=18437

FYI 3.8 also give you hyperz ...

[Vim_User]

How much can we trust this benchmark if recent comparisons are showing very clearly that Unity can have unpredictable influences on 3D performance?

[dungeon]

Nah, i was hope to see supertuxkart 0.8 test on radeon, but nope. Is this true or not .

http://openbenchmarking.org/result/1...PTS-SUPERTUX37

[Hamish Wilson]

So, has there been any word to explain the problems with both Prey and Nexuiz here (assuming it is not something like AA being triggered)? Everything else is indeed looking good.

[fhuberts]

On a HD4650.
This is an oooold and slooow card.
Kind of useless benchmark again...

[GreatEmerald]

On a HD4650.
This is an oooold and slooow card.
Kind of useless benchmark again...

Not for us old and slow card owners

[Calinou]

On a HD4650.
This is an oooold and slooow card.
Kind of useless benchmark again...

I think an HD7970 GHz edition should be tested. :P

[Hamish Wilson]

Lay of Micheal for once there people - the 4650 is probably a very common card to be used by people using the R600g drivers.

[Rexilion]

How much can we trust this benchmark if recent comparisons are showing very clearly that Unity can have unpredictable influences on 3D performance?

To the fullest extend, since Unity was used like it's used to all the way through the entire benchmark? (duh...)

So, has there been any word to explain the problems with both Prey and Nexuiz here (assuming it is not something like AA being triggered)? Everything else is indeed looking good.

Prey and Nexuiz performance can be explained by actually studying the articale.

- kernel v3.4 makes radeon advertise mesa version 2.1
- kernel v3.5 and higher makes radeon advertise mesa version 3.0