AMD's A-Sync DMA Code Makes For Fast Performance

[Hamish Wilson]

Fun times ahead it seems. Will be interesting to see where things are at when the dust settles on all of these patches.

[Veerappan]

I've finished the tests that I was planning on running. Heaven improved the most, but when I ran Xonotic at 1920x1200 at the highest settings, the framerate went from 9fps to 29... Not bad for just a kernel upgrade with a few hundred lines of changed code. I'm assuming that running Xonotic at max settings caused my GPU to run out of available VRAM and cause lots of swapping between system/vram, which is something that this new feature should help out with.

The tests include:
Heaven
Nexuiz
OpenArena
World of Padman
Smoking Guns
Tremulous
Urban Terror
Warsow
Reaction Quake 3
Xonotic

http://openbenchmarking.org/result/1...SU-ASYNCDMAT84

[crazycheese]

**tests**

Correct URL: http://openbenchmarking.org/result/1...SU-ASYNCDMAT84

Thanks for testing!!

It seems opensource has ~60% performance already.
Does your driver already have Marek HyperZ patches?

I assume Xonotic uses more modern OpenGL features, that are done in software or ignored now - hence low fps.
You can prove me right or wrong, if you test OpenArena 0.8.5 (old GL) and compare how it scales to 0.8.8 (more modern GL).

PS
You can safely drop Urban Terror, its bottlenecking piece of dust.

PPS
I wonder how it scales on 5850/70 // 6950/70 hardware...
Thanks again!

[Hamish Wilson]

It seems opensource has ~60% performance already.

60% of what? What baseline are you using?

[Veerappan]

Correct URL: http://openbenchmarking.org/result/1...SU-ASYNCDMAT84
It seems opensource has ~60% performance already.
Does your driver already have Marek HyperZ patches?

No, this doesn't have the HyperZ patches applied. This was as of the following commit (http://cgit.freedesktop.org/mesa/mes...it/?id=3392f2f)

I honestly don't remember if this had the suballocator patches hand-applied before I started, but I think it did. I can check again by re-running the reaction quake tests on the old kernel and checking the performance against my original reaction quake 3 tests.

I assume Xonotic uses more modern OpenGL features, that are done in software or ignored now - hence low fps.
You can prove me right or wrong, if you test OpenArena 0.8.5 (old GL) and compare how it scales to 0.8.8 (more modern GL).

Running that currently. Probably won't have the results posted until morning.

PPS
I wonder how it scales on 5850/70 // 6950/70 hardware...
Thanks again!

Couldn't tell you. The only other radeon hardware I've got is the HD4200 IGP built into my motherboard, and the Llano chip in my HTPC (which I'm not especially thrilled about messing with the software installation on).

[Veerappan]

http://openbenchmarking.org/result/1...SU-ASYNCDMAT84
I assume Xonotic uses more modern OpenGL features, that are done in software or ignored now - hence low fps.
You can prove me right or wrong, if you test OpenArena 0.8.5 (old GL) and compare how it scales to 0.8.8 (more modern GL).

OpenArena 0.8.5 actually regressed slightly (from 154 to 147 fps, not enough to worry too much about).

I'm downloading pts/etxreal at the moment, and we'll see how that works out. once that's done running and I've gotten some sleep, I'll upload an updated result set.

[F i L]

Nice! Thanks for the benchmarks!

[crazycheese]

OpenArena 0.8.5 actually regressed slightly (from 154 to 147 fps, not enough to worry too much about)

Is within acceptable fault rate. Yes, we have 24 fps for 0.8.8 and 150 fps for 0.8.5. Its ~84% speed loss. This would prove my point that OpenArena 0.8.8 relies on features that on MESA side cry for improvement.
Compared to Catalyst ratings as baseline, if we watch 6950 fps drop, the ideal performance loss is only 75%(instead of 84%). Thats the amount of deficiency in MESA.

This would mean, if GLSL and various other adjustments in MESA are done, your GPU would theoretically top at 37 fps for OA 0.8.8 profile.
Of course, this is projected from current baseline OpenGL2.0 performance - if that improves too, fps would climb at same rate.
For example, in link above, watch a fps "show" with Warsow 1.0. This game heavily uses shaders and they are bottlenecking to such a great degree, that under opensource 6570 (50% to catalyst) runs same to 6950 (15% to catalyst)!

But hey, only 3 years ago, I had 5 fps in 0.8.5 using 4670...

60% of what? What baseline are you using?

Baseline is ....catalyst in high resolutions. Baseline is practical maximum, as achieved by catalyst.

[Veerappan]

ET:XReal wouldn't run on this machine, so I'm calling it quits for now with this latest upload:

http://openbenchmarking.org/result/1...SU-ASYNCDMAT52

If I feel really ambitious over the weekend, I'll install catalyst and get some comparative numbers.

[Hamish Wilson]

Baseline is ....catalyst in high resolutions. Baseline is practical maximum, as achieved by catalyst.

Where are you getting your Catalyst data?