Revisiting Linux GPUs: Playing the Driver Game
by Kristopher Kubicki on December 17, 2004 12:49 AM EST- Posted in
- Linux
The Test
We will run a few different configurations of GPUs and operating systems just to give everyone a point of reference. The graphics cards that we have selected today should strongly correlate the same market segments - the Radeon X800 Pro and the GeForce 6800 generally both run at or over $300. Unfortunately, there are very few cards still available today that compete on the 9800 Pro level, so we were limited to including a GeForce 5700 Ultra for comparison (both cards run around $200.| Performance Test Configuration | |
| Processor(s): | AMD Athlon 64 3800+ (130nm, 2.4GHz, 512KB L2 Cache) |
| RAM: | 2 x 512MB Mushkin PC-3200 CL2 (400MHz) |
| Motherboard(s): | MSI K8T Neo2 (Socket 939) |
| Memory Timings: | Default |
| Video Card(s): | GeForce 6800 128MB GeForceFX 5700 Ultra 128MB Radeon X800 Pro 256MB Radeon 9800 Pro 128MB |
| Operating System(s): | SUSE 9.1 Professional kernel 2.6.8-14-default x86 SUSE 9.1 Professional kernel 2.6.8-14-default x86_64 |
| Driver: | fglrx 3.12.0 fglrx 3.14.6 NVIDIA 1.0-6111 NVIDIA 1.0-6629 |
| Compiler: | GCC 3.3.3 |
We will try to demonstrate the performance between the 64-bit and 32-bit kernel on Linux with the two driver sets. The NVIDIA driver that we are looking at today is already a month old - but most Linux users take a very conservative approach to inserting binary code into their kernels and we wouldn't be surprised if a strong majority of users have not updated to this release yet. ATI's 3.14.6 driver is also about a month old, which gave us some lead time to test the newest build of Doom3 with the newest drivers.
You will notice the test configuration that we are running looks a little old - SUSE 9.2 has been available for a few weeks now. However, since we had some of these benchmarks completed in October on the SUSE 9.1 platform, we opted to continue this analysis with the same platform to illustrate the differences on a stable platform. Feel free to check out the previous analysis test page here. The kernel used in this benchmark is still fairly current, but the userland is older.
Our test methodology is very straight forward, since we will only be benchmarking three games for this analysis. For Enemy Territory, Unreal Tournament and Doom3, we will run a basic set of timedemos along with two dimensional data from the AnandTech FrameGetter utility that we unveiled several weeks ago. We will be watching for specific fluctuations between the two driver sets on each video card. Furthermore, we will use Unreal Tournament 2004 in 64-bit and 32-bit mode to compare performance between the two sets of drivers in 64-bit mode. Unfortunately, there are still no 64-bit ATI binaries, so those tests are limited to NVIDIA only.
Facebook
Twitter
RSS
23 Comments
View All Comments
DarkTrooper - Tuesday, December 6, 2005 - link
First the good part, I am glad that ATI has finally managed to get a good working installer. Now the bad part for me: since I switcht to an x850 (R481 Core) I cant use the Driver, since it seems that the specifiy PCI-ID is not in this driver, which results in an "no matching pci ID found" after running fglrxconfig (which would be nice to have in a GUI version, or at least a menue based text version). Hope they manage to add this product soon... gets kind of anoying always having to wait month over month until they manage to fix things like this. Also the missing 64bit driver is some what disapointing (even though I only own 32Bit maschines).mpineiro - Thursday, May 5, 2005 - link
While I will admit that I only skimmed through the article, I think you failed to mention that it is nearly impossible to install the ATI drivers, especially with 3d Acceleration on most distributions.momenman - Friday, December 31, 2004 - link
Though the statements here regarding NVIDIA is generically true - I am personally unhappy with the "mouse moves but screen frozen" bug - due to which I can't use the 3D acceleration on FX-5700 ( though the problem is not there with GeForce2 MX cards, for example. ) . Its been almost a year and half since this bug surfaced on the nvidia linux board but the problem persists - thanks to the closed nature of the nvidia drivers.Let's just wait and see .....
svartalf - Sunday, December 19, 2004 - link
In regards to the DRI project drivers "not being there", I will offer that you've been trying the official released drivers that are typically included with the distributions. (By the way, you can _almost_ play UT 2k4 with the drivers- I know, I recently tried it on a Centrino based laptop and an r200 based GPU.) Recent improvements include 3D accel across Xinerama screens and TCL support- these are betas in the version control system and should be showing shortly in the distributions as they're wrung out.SLIM - Saturday, December 18, 2004 - link
You can add one more dissenting opinion to the growing list in regard to the comment about 64/64 being a meager boost compared to 32/32 UT2004. A 15% boost in performance is better than a processor upgrade, some vid card upgrades and certainly better than anything expensive ram can do compared to value ram (2-2-2 vs 3-3-3 @400). I just hope 64bit windows will be able to show the same level of performance gains.Saist - Friday, December 17, 2004 - link
I'm just going to say this to the author:try to install ATi drivers into Debian.
if you get that to succeed. Please do an article on how you did it.
Rand - Friday, December 17, 2004 - link
I'm inclined to agree with the opinions of thoe others, I felt the relative gains from going to 64bitwas quite appreciable.At this point I'm not sure I would even have expected the ~10% gain seen, so it's quite decent from my perspctive.
ATI's performance under Linux still remains wholly unimpressive at best. Seeing the less then impressive 5700U beating the X800 Pro in a number of cases only serves to underline how poorly ATI's drivers perform.
It doesn't impact me personally as I rarely use Linux and not at all for gaming, but for those that do nVidia is still clearly the manufacturer of choice.
mlittl3 - Friday, December 17, 2004 - link
Oh and one other thing.Intel said that going from 32 to 64 bits will just allow one to add (address) more than 4 GBs of RAM to a system and that's all.
AMD always said that the added registers would allow for improvements no matter how much RAM you have.
I guess AMD was right.
PS. How much improvement is necessary to go from meager to the next level? The way I see if from the article it's something like this.
0-10% Abysmal
10-20% Meager
20-30% Poor
30-40% Average
40-50% Good
50%> Excellent
I don't think we will ever see a greater than 30% increase going from 32 to 64 bit in order to get a positive rating, but hey, some people are harder to please than others.
mlittl3 - Friday, December 17, 2004 - link
I have to agree also with #5,8,10. I must not know what the definition of "meager" is. Any improvement over 10% is totally significant by any measurement.Just looking back at an old Anandtech article comparing Unreal 32 to 64 bit performance in SuSe 9.1 shows 32bit performance at 26 fps and 64bit at 25 fps. Now a 1fps difference is meager and shows 32bit winning over 64bit.
http://www.anandtech.com/linux/showdoc.aspx?i=2114...
I know this was a very different system but the absolute frame rate is not what's important. Its the difference between 64bit vs. 32bit. Driver development has come a long way and now we are seeing a 13 to 15% increase in fps going from 32 to 64bit.
And you can't say that its just the drivers because the same drivers are being used in 32 and 64 bit modes in this new comparison and we are still seeing a "huge" (not meager) gain in performance going from 32 to 64bit.
Framing articles in reference to old data is a good way to measure improvements. I think that is how we know if an older video card is outperformed by a newer one, a new operating system works better than an older, etc.
deathwalker - Friday, December 17, 2004 - link
yawwnnn!!!!