Workstation Graphics: AGP Cross Section 2004
by Derek Wilson on December 23, 2004 4:14 PM EST- Posted in
- GPUs
Shader Analysis
To open this section, we would first like to start by stating that we wish we could have found a suitable benchmark to test GLSL performance in a similar way that Shadermark manages to test HLSL performance. OpenGL fragment shading performance under which we ran the demos on the Wildcat Realizm part is much higher than its DirectX pixel shading performance under Shadermark. In fact, even in playing with ATI's own Rendermonkey, it was apparent that the 3Dlabs card handled GLSL shaders better than the FireGL X3. Since OpenGL is the language of the workstation, it makes sense that the workstation is only where 3Dlabs would focus its efforts first, while ATI's consumer oriented approach would lend it the clear upper hand in DirectX HLSL benchmarks like Shadermark.But DirectX and HLSL is still a very relevant test and is supported on all these cards. Of note is the fact that Shadermark would not run PS 3.0 code on the Wildcat Realizm. Shadermark chose to use the PS 2.0a path, which supports a wider range of features than the PS 2.0b path used for both of the ATI cards. Shadermark has been known to be very picky about code paths that it runs, and it's possible that there is an issue with the fact that this 3Dlabs part is simply not on the Shadermark map. But part of the point of HLSL is that the code should still run with no problems. We did get the option of creating an A2R10G10B10 hardware device on the Wildcat Realizm in shadermark where no other card presented such a feature. But let's take a look at what the numbers have to say.
| Shadermark v2.1 Performance Chart | |||||
| GeForce 6800U | Quadro FX 4000 | Radeon X850XT | FireGL X3-256 | Realizm 200 | |
| shader 2 | 893 | 596 | 996 | 731 | 41 |
| shader 3 | 736 | 493 | 735 | 531 | 28 |
| shader 4 | 737 | 493 | 732 | 531 | 28 |
| shader 5 | 669 | 448 | 608 | 438 | 16 |
| shader 6 | 680 | 467 | 735 | 530 | 28 |
| shader 7 | 631 | 417 | 654 | 485 | 23 |
| shader 8 | 383 | 255 | 406 | 301 | 11 |
| shader 9 | 894 | 630 | 1263 | 977 | 55 |
| shader 10 | 807 | 553 | 819 | 617 | 43 |
| shader 11 | 680 | 467 | 694 | 509 | 27 |
| shader 12 | 446 | 319 | 263 | 186 | 13 |
| shader 13 | 383 | 276 | 361 | 252 | 13 |
| shader 14 | 446 | 316 | 399 | 280 | 18 |
| shader 15 | 328 | 244 | 285 | 206 | 21 |
| shader 16 | 314 | 224 | 336 | 244 | 8 |
| shader 17 | 425 | 309 | 429 | 315 | 8 |
| shader 18 | 56 | 39 | 40 | 30 | 2 |
| shader 19 | 180 | 134 | 139 | 99 | 6 |
| shader 20 | 57 | 41 | 47 | 33 | 3 |
| shader 21 | 90 | 63 | - | - | - |
| shader 22 | 119 | 96 | 204 | 154 | 14 |
| shader 23 | 133 | 106 | - | - | 15 |
| shader 24 | 80 | 67 | 143 | 108 | 118 |
| shader 25 | 97 | 69 | 118 | 86 | 6 |
| shader 26 | 93 | 67 | 123 | 89 | 6 |
Not surprisingly, the consumer level parts are the top performers here. The Quadro FX 4000 and FireGL X3-256 don't do a bad job of keeping up with their desktop counterparts. However, the Wildcat Realizm 200 puts in a very poor showing. In addition to this, the Realizm didn't render many of the shaders correctly. Granted, the Microsoft reference rasterizer does not create correct images, but they are close in most cases. Shadermark generates MSE (mean squared error) data for screenshots taken and compared against reference images. Both ATI and NVIDIA hit between 0.5 and 1 in most tests. There is not a single shader rendered on the Wildcat Realizm 200 with an MSE of less than about 2.5. Most shaders show very clear image quality issues.
With the image quality of Wildcat Realizm in Doom 3 and Half-Life 2 being dead on with the other cards, and performance under Half-Life 2 not being as bad as we expected, we have to wonder how much of the issues with Shadermark would translate into actual applications. And by applications, we mean any application that allows the creation and/or visualization of HLSL or GLSL shaders. DCC workstation users are becoming more and more involved in the process of creating and designing complex shader effects. In order to maintain a firm position in the future of DCC workflow, 3Dlabs will need to assure smooth, accurate support of HLSL, no matter what the application running the code.
We will continue to evaluate programs for benchmarking GLSL performance. Through observation, the NVIDIA and 3Dlabs parts have an advantage over the ATI parts in GLSL performance. Unfortunately, we don't have any quantitative tests to bring to the table at this time.
Facebook
Twitter
RSS
25 Comments
View All Comments
Jeanlou - Thursday, December 1, 2005 - link
Hello,I just bumped into AnandTech Video Card Tests, and I'm really impressed !
As a Belgian Vision Systems Integration Consultant (since 1979), I'm very interrested about the ability to compare these 3 cards (Realizm 200 vs FireGL X3 256 vs NVIDIA Quatro FX 4000).
I just had a bad experience with the Realizm 200 (!)
On a ASUS NCCH-DL motherboard, Dual Xeon 2.8GHz, 2GB DDR 400, Seagate SCSI Ultra 320 HDD, 2 EIZO monitors (Monitor N°1= L985EX at 1600x1200 px), (Monitopr N°2= L565 at 1280x1024 px), Windows XP Pro SP2 x32bit partition C:\ 16GB, Windows XP Pro x64bit edition partition D:\ 16GB, plus Extended partions (2 logical E:\ and F:\). All NTFS.
Using the main monitor for images analyses (quality control) and the slave monitor for tools, I was unable to have a stable image at 1600 by 1200 pixels. While the Wildcat4 - 7110, or even the VP990 Pro have a very stable screen at maximum resolution. But the 7110 and the VP990 Pro don't have drivers for Window XP x64bit.
Tried everything, latest BIOS, latest drive for ChipSet...
Even 3Dlabs was unable to give me the necessary support and do not answer anymore !
As soon I reduced the resolution from the main monitor to 1280 by 1024, was everything stable, but that's not what I want, I need the maximum resolution on the main monitor.
The specs from 3Dlabs resolution table is giving 3840 by 2400 pixels maximum!
I send it back, and I'm looking for an other card.
I wonder if the FireGL X3 256 will do the job ?
We also use an other monitor from EIZO (S2410W) with 1920 by 1200 pixels !
What are exactly the several resolutions possible with the FireGL X3 256 using 2 monitors ? I cannot find it on the specs.
Any comment will be appreciated,
Best regards,
Jean
kaissa - Sunday, February 20, 2005 - link
Excellent article. I hope that you make workstation graphic card comparision a regular article. How about an article on workstation notebooks? Thanks a lot.laverdir - Thursday, December 30, 2004 - link
dear derek wilson,could you tell us how much is the performance
difference between numa and uma in general
on this tests..
and it would be great if you could post maya
related results for guadro 4k with numa enabled..
seasonal greetings
RedNight - Tuesday, December 28, 2004 - link
This is the best workstation graphics card review I have read in ages. Not only does it present the positive and negatives of each the principal cards in question, it presents them in relationship to high end mainsteam cards and thereby helps many, including myself, understand the real differences in performance. Also, by inovatingly including AutoCAD and Gaming Tests one gets a clear indication of when the workstation cards are necessary and when they would be a waste of money. ThanksDerekWilson - Monday, December 27, 2004 - link
Dubb,Thanks for letting us know about that one :-) We'll have to have a nice long talk with NV's workstation team about what exactly is going on there. They very strongly gave us the idea that the featureset wasn't present on geforce cards.
#19, NUMA was disabled because most people running a workstation with 4 or fewer GB of RAM on a 32 machine will not be running with the pae kernel installed. We wanted to test with a setup most people would be running under the circumstances. We will test NUMA capabilities in the future.
#20,
When we test workstation CPU performance or system performance, POVRay will be a possible inclusion. Thanks for the suggestion.
Derek Wilson
mbhame - Sunday, December 26, 2004 - link
Please include POVRay benchies in Workstation tests.Myrandex - Saturday, December 25, 2004 - link
I wonder why NUMA was fully supported but yet disabled. Maybe instabilities or something.Dubb - Friday, December 24, 2004 - link
http://newbietech.net/eng/qtoq/index.phphttp://forums.guru3d.com/showthread.php?s=2347485b...
Dubb - Friday, December 24, 2004 - link
uhhh.. my softquadro'd 5900 ultra begs to differ. as would all the 6800 > qfx4000 mods being done by people on guru3d's rivatuner forum.I thought you guys knew that just because nvida says something doesn't mean it's true?
they must consider "physically different sillicon" to be "we moved a resistor or two"...
DerekWilson - Friday, December 24, 2004 - link
By high end features, I wasn't talking about texturing or prgrammatic vertex or fragment shading (which is highend in the consumer space).I was rather talking about hardware support for: AA lines and points, overlay plane support, two-sided lighting (fixed function path), logic operations, fast pixel read-back speeds, and dual 10-bit 400MHz RAMDACs and 2 dual-link DVI-I connectors supporting 3840x2400 on a single display (the IBM T221 comes to mind).
There are other features, but these are key. In products like Maya and 3D Studio, not having overlay plane support creates an absolutely noticable performance hit. It really does depend on how you push the cards. We do prefer the in application benchmarks to SPECveiwperf. Even the SPECapc tests can give a better feel for where things will fall -- because the entire system is a factor rather than just the gfx card and CPU.
#14, Dubb -- I hate to be the one to tell you this -- GeForce and Quadro are physically different silicon now (NV40 and NV40GL). AFAIK, ever since GF4/Quadro4, it has been impossible to softquadro an nvidia card. The Quadro team uses the GeForce as it's base core, but then adds on workstation class features.