Revisiting Linux GPUs: Playing the Driver Game
by Kristopher Kubicki on December 17, 2004 12:49 AM EST- Posted in
- Linux
>Unreal Tournament 2004 32-bit Linux
Unreal Tournament has the unique advantage of being a 64-bit and a 32-bit benchmark. We will use the 3334 demo binaries here - the 3339 demo was installed, but provided no performance benefits. We stayed with the 3334 binaries so that the previous analysis benchmarks were still relevant. Don't forget to check out our previous analysis here.
Things pretty much replicate what we saw in the Wolfenstein benchmark - small to moderate performance gains by both cards with ATI looking at a lot of ground to catch up on. Below, you can see a portion of the Unreal Tournament 2004 "assault" benchmark.
Also, you can see what affect the new driver set had on this same timeframe (the CSV file is available for download here.
Pay very close attention to the local maxima and minima. Comparing the second graph with the first, and not just the two cards against each other, reveals that both cards received some serious gains in performance without losing a step. NVIDIA's performance is very rapidly approaching the Windows performance, and we wouldn't be surprised if NVIDIA could actually pull off symmetric performance on its Linux and Windows platforms should the next few driver updates provide this sort of a performance gain.
Unreal Tournament 2004 64-bit Linux
As we mentioned before, we are limited to NVIDIA graphics cards for this portion of the test - there are no 64-bit ATI drivers yet. Here is a quick and dirty comparison of the two driver sets on the video cards that we selected for this analysis. The first graph shows the 32-binaries running on the 64-bit kernel. The second graph shows the 64-bit binaries running on the 64-bit kernel.
As expected, we noticed virtually no difference between the 32-bit binaries on the 32-bit kernel or 64-bit kernel. We see an occasional single FPS bump on the higher end cards, but there really isn't much advantage to running the 32-bit binaries on the 64-bit world. The 64-bit binaries yield better results, although there is nothing really outstanding about the performance gains going from the 32-bit binaries to the 64 ones, or going from the August driver set to the November one.
Facebook
Twitter
RSS
23 Comments
View All Comments
KristopherKubicki - Friday, December 17, 2004 - link
Pannenkoek: Unfortunately I have large doubts about ATI or NVIDIA ever opening up their drivers. Both companies have more software engineers than hardware engineers. They spend a *lot* of time and money reinventing the wheel between the two of them - and I think they want to keep it that way.Kristopher
KristopherKubicki - Friday, December 17, 2004 - link
MooseMuffin: It's SUSE 9.1 - i think i might have a typo in there somewhere. We kept it at 9.1 instead of 9.2 just for that reason (the kernel is very updated though).Hope that helps,
Kristopher
Pannenkoek - Friday, December 17, 2004 - link
I back up #5: 10-15% gain from 32bit to 64bit is not "meager"...Before you ditch the open source 3D drivers for the newer videocards (if any exist...), please keep in mind they have to reverse engineer the cards, as NVIDIA and ATI don't co-operate and no hardware spec's are available. As far as I know only serious 2D OSS drivers are in development.
Also, we should not applause ATI's gains in performance, as they were abominable to start with. However, we should applause their changing attitude towards open source platforms. Let us hope it will continue to improve!
And let us hope NVIDIA and ATI will open their hardware in the future, so open source drivers can be made for them. No buggy proprietary drivers tainting the kernel anymore. But I fear we may wait a long time for that to happen. ATI is hugging Direct3D and MS too closely to encourage development of good OSS drivers as a way to counter NVIDIA's lead in OpenGL. And NVIDIA won't open up as long as that is the case.
Nevertheless, hereby I beg NVIDIA and ATI to design their future generation cards in such a way that opening the spec will not expose their holy IP.
Myrandex - Friday, December 17, 2004 - link
And that should only increase with time with optimized 64bit code, drivers, improved operating system components, etc.icarus4586 - Friday, December 17, 2004 - link
...performance gains between 32-bit and 64-bit distributions on Unreal Tournament 2004 were meagerI agree with Icehawk, and beg to differ with the author. >10% performance gains are not "meager" by any stretch. Imagine NVidia/ATI releases a new Windows driver that increases performance 10%. I'm pretty sure nobody would say that was a "meager" improvement.
R3MF - Friday, December 17, 2004 - link
thank god i have an nForce2 and Ti4200, SUSE 9.1 runs like a dream.i have just bought an nForce3Ultra and 6800GT for the parents.
i will upgrade to an nForce4 and 6800GT early next year.
notice a trend? you would have to be daft using ATI silicon in a machine you intend to install Linux onto.
Icehawk - Friday, December 17, 2004 - link
I should say total delta from 32:32-bit to 64:64-bit .Icehawk - Friday, December 17, 2004 - link
I don't know why they say the bump in 64-bit UT performance is minor - if you look at the total delta from 1.0-611 32-bit -> 1.0-6629 64-bit it is a ~13% increase on the 6800 and ~15% on the 5700U which is pretty darn good IMO.MooseMuffin - Friday, December 17, 2004 - link
How did you guys get these drivers installed on Suse 9.2? As far as I can tell suse 9.2 uses xorg and ati only supplies xfree drivers.mickyb - Friday, December 17, 2004 - link
ATI's performance is shameful on linux. They have some serious work to do.