Slashdot Mirror
ATi FireGL X1 Vs. NVIDIA Quadro FX 2000
SpinnerBait writes "The professional graphics card arena has been heating up as of late, with new
products from ATi and NVIDIA hitting the streets on the heels of SIGGRAPH
unveilings. In a first of two article series,
HotHardware has a showcase with benchmarks on the ATi FireGL X1 and NVIDIA
Quadro FX 2000. It seems as though NVIDIA still has a stronghold in
this market, as their card seems to dominate many of the benchmark runs shown
here."
It doesn't bother me that much who has the fastest card. All I know is that this sort of competition is great in the Linux arena. With the recent trends in 3d animation studios transition to Linux, they can't ignore the need for high quality drivers.
Nvidia has really polished up their Linux drivers recently, and in response ATI has done the same.
This means Linux is one step closer to gaining a foothold on the desktop. Hopefully this will will spur interest 3D gaming on the linux platform.
One can dream of the day of playing Battlefield 1942 on Linux. I'm using the Liux FireGL drivers on my Radeon 9700 Pro, and so far, they work great for playing RTCW ET.
"Why is it that both products we'll be looking at today, the [$725--128MB]ATi FireGL X1 and [$1500--256MB]NVIDIA Quadro FX 2000, share nearly identical hardware with their consumer counterparts, yet cost 3 to 5 times as much? The answer goes back to those highly specialized applications again, and optimizing the hardware and drivers to accelerate performance to the best of the core Graphics Processor's ability"
---
It would have been nice if they also benchmarked a $400 GeForceFX5900-256MB and a $425 Radeon 9800Pro-256MB then. (current prices from pricewatch)
Anyone have a link to another review that includes these?
Whenever the offence inspires less horror than the punishment, the rigour of penal law is obliged to give way...
Have we, the public, not learnt yet that ATI and Nvidia have "optimized" their drivers for whatever benchmark.
Unless reviewers compare same motherboard, same amount of ram, same processor, same bios version, same version of the motherboard, as what their audience has then the numbers are MEANINGLESS.
My impression from the Anandtech review is that the visual quality of the FX 5900 vs. ATI's 9800 were quite comparable, and at times, indistinguishable.
*shrug*
ATI is going to have a hard time in the developer market ---
"According to Carmack, nVidia is among the best in the business at writing drivers. He went on to explain that whenever he runs into a driver-related bug with nVidia, he assumes the problem is with his own code. With ATI or other card manufacturers, he assumes the problem is with the driver. Extremely high praise for the driver engineers at nVidia."
[cited from http://www.dallasnews.com/, a review of QuakeCon 2002]
The only reason I'm still buying nVidia is because it's setup under linux is very well documented, whereas I've had a bugger of a time getting a Radeon 9000 Pro to run accelerated on linux. And I've not found anything terribly useful when asking google.
Inconceivable!
I respect John Carmack, but I am not a picky developer like him. ATIs drivers seem fine to me and the cards perform great.. and thats all the criteria I have.
My experience with ATI and Linux is limited to Gentoo:
emerge ati-drivers
Works like a charm...
The unofficial
I don't see dual monitor setups mentioned in the article. Does ATI's output quality stand up to NVidia's?
I have a Radeon 8500, and I can tell you that ATI has some serious issues with output signal quality. On my main crt monitor, I can still see occasional sheering and small display glitches. The 2nd monitor quality was even worse. I am using a pci TNT card to get 2nd monitor suupport.
Judging by the picture of the ATI card, the second DVI connection may have problems. It is an extra board so there is not a continuous trace which can introduce all sorts of problems (like contact resistance, oxidation, etc.) Yes, it is a digital signal, but it's like putting an ide ribbon cable with really short wires. You are going to get all sorts of problems...
While this may be brushing 'redundant/offtopic', I have to say that getting one of those may cost you a bit more, but it's much nicer than a consumer graphics card.
/. article on it), while the latter have been traditionally good with it.
What the author fails to mention is that there's better R&D (build quality?) put in there. Not just application-specific optimisations. If they *had* tested the consumer equivalents, they'd see them outperformed, methinks. That's my experience, anyway.
Back in '98 I had a Diamond FireGL 1000 Pro (yes, the FireGL series was owned by Diamond then), which was matching/outperforming many 'new' gaming cards my mates were buying (it was a fairly old model at the time, IIRC). Thing is, I hadn't paid a fortune for it, as you might think. It was a bit expensive, but not *that* different from what my mates were paying.
Now I have a FireGL 8800 and again the performance is there. Gaming-wise, I can play GTA3 and CMR3 at resolutions previously undreamt of with the 9500 (1600x1200).
Having said that, it's a pain to get (linux) support by ATI. Ever tried emailing them? Up 'till March (IIRC) things were OK and they even had good drivers. But now it's all shaky and iffy, as we all know.
Now I'm looking for a 3DLabs/NVidia. The former are increasing their linux support (I even recall a
It would have been very interesting if they'd included the VP990 Pro or the VP970 in the comparison...
Trian
I'm no longer fed up with MS Windows: I go rid of them
How often does it fsck up a render? With consumer cards, who cares if you mess up a render, because it may just be a temporary jaggy, they just want to be all out speed-demons. But, with these corporate cards, a messed up render could be a misplaced weld, or something along those lines.
I hate sigs.
First of all, that isn't true even with last-generation hardware - CAD apps sure need realtime rendering. Speeding up a complex model from .5 FPS to 10 (or 100) FPS can result in big productivity gains.
Even the VR apps like 3D Studio can use the programmable shader features of these newest cards to render production quality scenes in realtime or near realtime. One of the big presentations at E3 *last year* was a realtime rendering of one of the big scenes in the Two Towers - in full cinematic quality on a GeForce FX.
3D graphics is a tremendously exciting area.
Galileo: "The Earth revolves around the Sun!"
Score: -1 100% Flamebait
This article doesn't factor in a number of important things such as extreme model sizes (that kill the FireGL), driver quality (ATI Radeon is historically bad, New FireGL (ATI parts not IBM parts) is worse), and precision. The FireGL X1/X2 have horribly low precision. The biggest area that shows this is in their sub pixel precision which results in many rendering errors per frame (holes, spots, tears, speckles).
With Quadro4 it actually became a different chip so you can't turn the workstation specific parts of the chip on with softquadro any more.
The (true) Wildcat 4 cards are much slower than the Quadro FX 2000 and Quadro FX 3000. They are also slower than the ATI FireGL products. The Wildcat VP products are slower than just about anything you put next to them. 3dlabs is no longer a significant player in the workstation graphics space.
That board (and epoxy/heatsink) was NOT made and assembled by NVIDIA. NVIDIA does not produce their consumer boards. All the boards are produced by other vendors such as PNY, Asus, MSI, Chaintech, Leadtek, Gainward, etc. using chips they buy from NVIDIA. NVIDIA does produce and tightly control the manufactur%3of their Quadro cards.
I am a CAD administrator, and use several different CAD packages. The problem I have with most graphic card now isn't performance - it's accuracy. When you zoom up on an intersection and the lines "move" at different zoom levels, it becomes impossible to know which surface is which.
I have had this problem with Quadro cards. I have not had a chance to try ATI cards. I have had the best results with older 3D labs card (gx1 pro and gmx 2000.) Those cards did not offer the fasts performance, but were better for surfacing.
Where performance matters is when I'm working on large assemblies. Some of the repaint/redraw times can be as high as 15 minutes (1.8 GHz, 1 GB RAM, Quadro 550, Pro/Engineer, Windows NT)
I would love to give Pro/Engineer a try on Linux (It's available - web site is ww.ptc.com) But, our PDM package (Pro/Intralink) is not. Does anyone know how the performance compares to Windows?
Man, trying to keep up with all the things you know nothing about is hard but I'll try.
NVidia can't provide hardware interface specs for their boards because they don't own the rights to do so. That's why they go out of their friggin way to make the best 3d drivers and experience for games and graphics for linux out of the kindness of their hearts on their time on their dime. It's not like they are making money doing it. Me and some other guys started a petition a while back, and after about a year they started paying attention.
I wouldn't say NVidia services the market in a limited way given they walked away the hands down performance winner in at least this poorly conducted benchmark. There are plenty of people that trust and use their products. It has a lot to do with driver quality. Don't fall into the pathetic fanboy trap that a lot of people have and start predicting their demise. They are highly diversified. the NForce mobo chipsets alone have made them a fortune. It was pretty nice of them to release Linux drivers for those too. It's not like they had too. They also didn't really have to bother releasing AGP support for other peoples cards either.
If you'd do a little bit of research into how SCSI works, and how it is manufactured, you'd probably understand why it is priced the way it is. You'd probably eventually understand why it's the only choice in a number of situations. Lets see you put together a 60 drive rack mount EIDE raid 5 solution. For a home user? EIDE or SATA is probably going to be fine. Someone doing massive amounts of graphics processing and realtime video editing? The drives are definitely going to end up being a bottleneck if they aren't fast as hell. Transfer and seek time. Raid 10 with 15,000 rpm drives dangling off of 3 channels doing reads and writes fast enough to keep up with the rest of your system. A friggin boatload of cache with speculative read ahead caching. Godlike control over every switch and variable you need to maximise throughput based on the size of the files you are working on. Basically, things you can't commonly do well with any IDE raid solution at this point.
For every annoying gentoo user, are three even more annoying anti-gentoo crybabies. Take Yosh from #Gimp for example.
For the 3DS type apps, that big of a jump will happen in one generation.
The reason being, the older cards simply couldn't render the bulk of a 3DS scene in hardware...the new ones can.
Galileo: "The Earth revolves around the Sun!"
Score: -1 100% Flamebait