Linux 3.1 To Linux 3.8 Radeon DRM Benchmarks

[phoronix]

Phoronix: Linux 3.1 To Linux 3.8 Radeon DRM Benchmarks

After recently carrying out legacy Radeon benchmarks comparing Mesa/Gallium3D versions from an ATI Radeon X1800XT (R520) graphics card, up today is a vintage Linux kernel DRM comparison. For seeing if modern Linux kernels are still influencing the performance of this vintage ATI Radeon graphics card, here are benchmarks comparing the modern Linux 3.1 to 3.8 releases.

http://www.phoronix.com/vr.php?view=18495

[F i L]

well that's discouraging...

[chris200x9]

well that's discouraging...

not really...just kind of pointless IMHO. Isn't the x1800 using r300g which has been pretty much feature complete for a while making this test all but useless? It would have been interesting to see a 5870 or something r600g (where improvements are actually still being made)

[randomizer]

If you still use a card this old then you do not play games where further driver improvements will help you.

[Michael]

not really...just kind of pointless IMHO. Isn't the x1800 using r300g which has been pretty much feature complete for a while making this test all but useless?

As results next week will show, there's still OpenGL cases where Catalyst is much faster than R300g.

[F i L]

not really...just kind of pointless IMHO. Isn't the x1800 using r300g which has been pretty much feature complete for a while making this test all but useless? It would have been interesting to see a 5870 or something r600g (where improvements are actually still being made)

Oh whoops, i read it wrong and thought he was using a newer card than that.

[necro-lover]

As results next week will show, there's still OpenGL cases where Catalyst is much faster than R300g.

You post fast FPS result in the past for the catalyst but the mouse input lag was 2seconds or higher means unplayable-

I stop playing HON with the catalyst because the input lag if i try to drop a item was so high no one can play a advance bottle master support with that kind of shit driver.

But you are not a gamer you don't have any clue what gamers really need.

It just doesn’t matter if your driver do 60fps or 400fps its more important do have a constant reaction time.

that’s the problem of crossfire if you do have 4 cards your reaction lag time of mouse and key-port input is so high your 10000fps are just pointless.

But hey you are the "Professional" who cares about the truth's anyway the FPS is higher go catalyst go! LOL!

[Ericg]

This is one thing that I really appreciate with the open source drivers vs the closed source drivers....closed source drivers are millions of lines code due to per-generation-optimizations. I cant even begin to image what its like to debug the thing...

open source drivers are more "general, over all optimzations" and more debuggable. The difference is... Radeon gives less total performance, but more consistent performance. (Input lag came to mind, or lack there-of, which the user above covered) Which you just don't get with Catalyst it seems.

[smitty3268]

As results next week will show, there's still OpenGL cases where Catalyst is much faster than R300g.

Maybe so, but i don't think anyone is working on it for r300g anymore. Didn't Marek basically say the driver was done, and he wouldn't be doing anything but adding bug fixes to it? A long time ago, like maybe even before the linux 3.1 kernel that you started these tests with?

I guess it's fine to confirm that's happening, and at least there haven't been any regressions. But these results are expected.

[curaga]

This just means we need frame latency benchmarks in PTS