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."

Comments

I ran benchmarks too

[Dancin_Santa]

Neither runs faster than the Orchid Farenheit 180.

I used Lotus 123 and WordPerfect 5.1 as the test applications.

Re:I ran benchmarks too

[WIAKywbfatw]

Neither runs faster than the Orchid Farenheit 180.

I used Lotus 123 and WordPerfect 5.1 as the test applications.

Aah, but do you have the ISA, EISA or MCA version of the card? That EISA version really kicks ass, especially on a system running MS-DOS 5.0 and tweaked with 386MAX.

Re:I ran benchmarks too

[palp]

These cards are not for games, they are for professional graphics work. I know this is slashdot, but please try to have a clue before you knock hardware that is useful.

Re:I ran benchmarks too

[Second Vampyre]

No you must own this card to run Doom 3! LoL LOL OLO L LEO LOLERZ!

Re:I ran benchmarks too

[idontrtfm]

qwerty!

Re:I ran benchmarks too

[Anonvmous Coward]

"the only game i ever play is skifree and i dont need a $500 video card to do so."

You'd pay >$500 for a vid card if your job was to build a 3d model consisting of 100's of thousands of polygons. Oh, coincidentally, that type of work is what this story is about.

Re:I ran benchmarks too

[sl3xd]

I'm not sure if I should laugh or cry; the fact that I know exactly what you're talking about is disquieting. I remember spending hours talking about the virtues QEMM/Memmaker/386MAX; and why OS/2 was gonna beat them all. I'm startin to feel really old. There's some really snazzy new-fangled things a comin these days, ain't there? Then I hear 'bout this new-fangled penguin thing made by a bunch of young whipper-snappers... who are older than I am...

Maybe I'm spending too much time talking to the old folks...

On the other hand, they whine a lot less than my parents' generation...

Re:I ran benchmarks too

[DAldredge]

Clue? Slashdot?

Re:I ran benchmarks too

[palp]

Agreed, but *not* playing games is probably not a sufficient reason to not buy one, either.

FireGL has MUCH better Price / Performance

[Gherald]

> 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."

Not really. The benchmarks were very close in most of the tests and if you consider what the end of the article says:

At this point in time, various price search engines have the ATi FireGL X1 listed at or around $530. Conversely, the NVIDIA Quadro FX 2000 is listed at no less that $1250 and that's in the 128MB variant, not the 256MB model we tested. So with this in mind, the FireGL X1 price/performance ratio is rather compelling, at less than half the cost of the competing NVIDIA product.

...The FireGL looks like a much better value.

Re:FireGL has MUCH better Price / Performance

[1000101]

Even though $720 seems like a huge price difference, even small gains in performance can result in thousands of dollars in saved time. I would think the better performance of the Nvidia would end up costing far less than the ATI in the long run.

Re:FireGL has MUCH better Price / Performance

[Anonymous Coward]

But "the long run" doesn't exist when you can buy another much faster card in 6 months.

Re:FireGL has MUCH better Price / Performance

[Anonvmous Coward]

"Even though $720 seems like a huge price difference, even small gains in performance can result in thousands of dollars in saved time."

Not as much as you'd think. These cards are for the UI to the 3D app, not for rendering. The difference between 30 fps and 60 fps isn't going to save any significant amount of money.

Re:FireGL has MUCH better Price / Performance

[Glock27]

These cards are for the UI to the 3D app, not for rendering. The difference between 30 fps and 60 fps isn't going to save any significant amount of money.

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.

Re:FireGL has MUCH better Price / Performance

[GoSpeedRacerGo]

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).

Re:FireGL has MUCH better Price / Performance

[NanoGator]

"Speeding up a complex model from .5 FPS to 10 (or 100) FPS can result in big productivity gains. "

Yes, we all know that speeding up something 20x will speed up productivity. I don't remember my vid cards ever getting that big of upgrade in one generation.

Re:FireGL has MUCH better Price / Performance

[CrowScape]

Except nowhere in the benchmarks do you hear a peep about the hardware rendering capabilities of either card. In the one benchmark I found that covered it the ATI absolutely CREAMED the Quadro FX, and that's where you can see massive $$$ savings.

Re:FireGL has MUCH better Price / Performance

[Methuseus]

Also, the R300/350 cores can be linked together in up to 256 GPUs interconnected, running basically in SMP mode for graphics. ATI wins in that respect as far as base hardware works. Too bad ATI isn't making any of these multiple GPU cards themselves.

Re:FireGL has MUCH better Price / Performance

[zangdesign]

Speeding up a complex model from .5 FPS to 10 (or 100) FPS can result in big productivity gains.

Not to mention huge productivity losses the moment someone comes up with a game that really pushes these cards to the limit!

Re:FireGL has MUCH better Price / Performance

[Outland Traveller]

If you actually look closely at the article's benchmarks, the ATI card is ahead more often than not..

Remembers, for a lot of the graphs lower numbers are better.

Re:FireGL has MUCH better Price / Performance

[dnoyeb]

Once again this is a fallacy. Speeding up computer performance does not result in higher user performance. it results in higher quality output as the user will tend to examine much more than otherwise.

When the computers are slower, the users simply skip through tasks, but they get them done is generally the same amount of time.

Re:FireGL has MUCH better Price / Performance

[arkanes]

That's both wrong and stupid. If you're designing a massive 3d model, but your card & hardware are pushed to the limits rendering it so you can't easily manipulate it and operations affecting it are time consuming, it will take you longer to complete this model. If working on these models is what you do all day, then you'll be more productive with high end hardware.

Re:FireGL has MUCH better Price / Performance

[Glock27]

Yes, we all know that speeding up something 20x will speed up productivity. I don't remember my vid cards ever getting that big of upgrade in one generation.

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.

Re:FireGL has MUCH better Price / Performance

[dnoyeb]

No. If your system does not support what you do on the basic level, then you system is essentially broken for your task anyway.

I am talking about speeding up a functioning system. not making a functioning system out of a broken un-performing one.

Re:ati vs nvidia

[rokzy]

you've obviously been hiding in a cave the last year. ATI is kicking the crap out of nVidia in all departments except market share, number of cheats in drivers, lies told to consumers, and arrogance.

Re:ati vs nvidia

[Gherald]

Well I have used a Geforce MX 200, Geforce 3, Radeon 8500, Radeon 9500 non-pro, and just bought a Radeon 9800 non-pro (will be flashing with pro BIOS).

The Geforce 3 was a good card, but its the only one that has died on me.

No problems with any of the Radeons, and they sure are fast!

IMO, Nvidia's only good desktop offering right now is the FX 5600 Ultra, which has a value comparable to ATI.

The 5900 has a few more frames than the 9800 in UT2K3, but its image quality with is noticeably inferior to the Radeon.

Didn't RTFA

[EmagGeek]

Didn't have to.

Do I trust benchmarks? No.

Will I ever trust benchmarks? No.

Are benchmarks meaningful in any way? No.

Do benchmarks have any credibility whatsoever? No.

'nuff said

Re:Didn't RTFA

[MoThugz]

Addition:

Do I play games? Yes, I'm a level infinity Elf Warrior in Nethack and I don't need no stinkin' 3D acceleration.

FPS are for wusses!

Re:Didn't RTFA

[EmagGeek]

And when are comparisons of two cards ever different by more than +/- 10% in the benchmarks?

Re:Didn't RTFA

[Gherald]

RTFA? There were differences of 25%-60% in several important benchmarks.

Re:Didn't RTFA

[VoiceOfRaisin]

"Are benchmarks meaningful in any way? No.
Do benchmarks have any credibility whatsoever? No."

are you insane? are the people that modded you up insane?

so a company that does 3d design needs new cards for their systems.. what do you suggest they do? buy the most expensive card on the market and hope its the best? buy the most expensive one they can afford? buy an assortment of cards and try every one themselves then decide which is best then try to return the other cards?

or be sane and read a review of various cards that TEST them, with something thats called BENCHMARKING.
is benchmarking a perfect science? no. will it ever be? no. but they are of use like it or not.

Re:Didn't RTFA

[arthurh3535]

Nevermind that it can be incredibly cheaper than the alternative means of buying the most expensive ones! And you're right, who the hell modded him up as insightful?

Benchmarks aren't perfect, but in a perfect world God would just implant the knowledge of which card was perfect for you into you're brain.

But some of us don't have God on the brain and have to make the best of our own choices. Choices that are made better when you at least check out the different options and don't automatically totally distrust anyone but your own testing.

Arthur Hansen

Re:Didn't RTFA

[EmagGeek]

"are you insane?"

Nope, not in the least

"are the people that modded you up insane?"

I think they must be - I was fully expecting a "Troll" rating on this one :) There was definitely a karma battle going on, but since they wienered out and changed to percentages, I don't know how many points were expended on each side.

That said, I still think benchmarks are complete BS. When they start "optimizing" things for benchmarks, their credibility and technical relevance are GONE.

I want them optimizing their stuff for MY APPLICATIONS, not the damn benchmarks.

Re:Didn't RTFA

[virtual_mps]

so a company that does 3d design needs new cards for their systems.. what do you suggest they do?...buy an assortment of cards and try every one themselves then decide which is best then try to return the other cards?

Well, yes. You don't make that kind of investment (you're probably talking a bunch of cards, not just one to outfit a whole shop) without testing how it works in your environment. Hell, if you're going to buy enough of them you can probably get the vendors to loan you a test sample. To buy on benchmarks is just nuts.

Re:Didn't RTFA

[EmagGeek]

Well, it's official, 14 moderator points used on my post. It's a record. In the past, I have only been able to garner 11.

Insightful: 7
Funny : 1 (oddly, the original intent)
Troll : 4
Overrated : 2

Looks like I win this round :)

Re:ati vs nvidia

[Feyr]

i have a radeon 9000, which i bought lately because exactly, i've been seeing those neat numbers about ati.

guess what? IT STILL SUCKS. the crappy drivers are STILL making the computer freeze (and NOTHING is overclocked)

Uhm

[Dark Lord Seth]

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.

Apart from the fact nvidia got their asses kicked in most benchmarks it does indeed rock, yes. Especially the bit which claims the price for the damned thing is over 1200 USD a piece. Ah well, next time it will be an Ati I guess, considering they both fecking cheat with benchmarks these days I might as well go for the cheapest cheater.

Fire GL1 is a bad video card...

[John Seminal]

I own the GL1 and it is a horrible video card. It was a pain in the rear to get the driver working. Every time I ran the install, it had to reboot and then it did not recognize the card in anything but 800*640. Once I got the card working, some software would lock up when it changes resolution, like what games do. My experiance with the Fire GL1 is it is a horrible card.

Re:Fire GL1 is a bad video card...

[cptgrudge]

Counter-point.

We have this card in a lab of Dell Precision 450's. I have never had any problems with it. I've done driver updates, loaded the ATI hydravision software, and generally pushed the card. I have had no complaints, or troubles with card.

On another note, I not surprised that the Nvidia card beat the ATI when it is over twice as much. Duh.

Re:Fire GL1 is a bad video card...

[SynKKnyS]

You are just a little bit off-topic. The comparison was between the FireGL X1 and the Quadro FX 2000. The FireGL 1 was a card from yesteryears. Declaring problems about the FireGL 1 is like telling people Datsun is a horrible car company in an article about Nissan cars.

Re:Fire GL1 is a bad video card...

[razjml]

Does anyone else feel like they're reading a 6th grade essay while reading this post?

Hmm?

[Dave2 Wickham]

GeForce FX? That'll heat both cards up nicely :).

From a Linux Perspective

[niko9]

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.

Re:From a Linux Perspective

[Distinguished Hero]

Battlefield 1942 is a Direct3D game; OpenGL drivers won't really help. Till game developers drop Direct3D completely, which I don't foresee happening in the near future, Linux (& Mac) gamers will always be missing a few titles.

Re:From a Linux Perspective

[niko9]

Alas, I wasn't sure which one, but there's always Carmack's games.

Re:From a Linux Perspective

[Glock27]

Battlefield 1942 is a Direct3D game; OpenGL drivers won't really help. Till game developers drop Direct3D completely, which I don't foresee happening in the near future, Linux (& Mac) gamers will always be missing a few titles.

Yep, too bad that author was shortsighted and used sucky APIs that locked 'em to a single platform.

Still, doing an OpenGL/SDL port probably wouldn't be too hard...

Re:From a Linux Perspective

[13Echo]

Wow! Thanks for the link to that Catalyst thread. I may very well be buying an ATi card soon. ;)

That's just the news that I wanted to see. I'm glad that they're still serious about Linux drivers.

Re:From a Linux Perspective

[Glock27]

Yeah, I bet when he's not counting his enormous stack of money, he has great regrets about not porting his game to Linux.

I think he'd be happy to buy an extra yacht with the 100,000+ copies he would have sold for Mac though.

Linux support is simply going for geek-factor cool...which is worth a lot in word-of-mouth advertising. ;-)

Re:From a Linux Perspective

[13Echo]

Well, had they used OpenGL, a Linux *AND* MacOS port would have been a simple task... Even a little profit is still profit.

Just look at the names

[Anonymous Coward]

FX 2000? What? Are they on crack? That says to people "This is the card from the year 2000! It's slow!!!" but X1? WOWOW!!! Holy jeez, that sounds fast!

Re:ati vs nvidia

[kannibal_klown]

I don't blame him. It's all opinion. I've been completely turned away from entire brands before because of a couple of consecutive problems (rather major ones).

Case-in-point, I will NEVER by Dell again because my last two purchases were utter garbage. Does this mean that Dell sux? No, they're probably one of the better PC manufacturers out there. I'm sure my experiences were in the small minority. But that doesn't change the fact that they've lost me as a customer forever.

It doesn't take much for a company to permanently leave a bad impression on one's mind. And a person shouldn't be blamed for haiting a company that's given them a bad experience, more than one.

I don't mind the benchmarks

[ReyTFox]

I've always gone for featureset when looking at graphics cards. Speed is a secondary and usually fairly costly function.

If I need the speed, I turn off AA and lower the resolution and game detail settings. But if it's fast enough for me as is and looks like it'll suffice for a couple of years, I don't care about the benchmarks.

Re:I don't mind the benchmarks

[NanoGator]

" But if it's fast enough for me as is and looks like it'll suffice for a couple of years, I don't care about the benchmarks."

I have a similar philosophy about cards. I also do 3D work. Here's my requirements:

1.) Does it support dual monitor? (not only support it, but good desktop support as well.)

2.) Does it do decent anti-aliasing? In 3D modelling, anti-aliasing makes a huge difference. When you're building your object, this graphics mode can reveal more about how your model will look when you go to render. It's worth the hit in FPS.

3.) Does it offer enough of a boost over the card I have now? Can I spend $200 and get 2x the speed?

I used to want a professional card. However, by limiting myself a bit, I've grown to become more efficient in handling the 3D assets. I know have a more structured workflow than I would have had if I had a much much quicker machine. I'm insanely curious what a Quadro would do for me, but man I have a hard time imagining it's worth the price tag.

However, I will likely change my tune soon. All the major 3D apps are taking much more advantage of the cards, previewing more and more of what the renderer will do. Modelling is anti-aliased, lens flares show up in real time, texturing, depth of field, motion blur, you name it. I can't help but think one day I'll be buying rendering cards instead of real time 3D cards.

Re:I don't mind the benchmarks

[paradesign]

render cards are old hat.

art

my school (CCSCAD - google it) has a renderdrive in our renderfarm, and all i have to say is damn! thats fast shit. but i dont know how upgradeable it is, last i used it i was on Maya 4.5, i dont know if it suports 5 or mental ray.

Re:I don't mind the benchmarks

[paradesign]

hmm, that link didnt seem to work out so well...

http://www.art.co.uk/

try that

Re:I don't mind the benchmarks

[mnemonic_]

You might want to upgrade your requirements a bit. Nearly every mid-range card in the past 2 years can do the first 2 items (dual head and decent AA).

Failing cards...

[WIAKywbfatw]

even if the ati card WAS faster, hell will freeze over before i ever buy another card from them!

they have crappy support, crappy hardware (as in reliability) and crappy drivers. i've had so many ATI cards die on me it's not even funny.

on the other hand i've had only one nvidia card die, due to rough handling and no fan (it came loose somehow and i didn't notice it, probably in transport)

Man, I've handled well over 100 different models from at least a dozen manufacturers over at least as many years and I've never had a card die on me. If all these cards are dying on you then there's got to be a reason - just what the hell are you using them for and in what environment? Unless you're a big time overclocker,video cards are pretty damn sturdy and the odds are that a card will outlast your use for it, so perhaps you need to re-examine just how you handle your cards and how much abuse that they're taking?

Having one card fail on you is unfortunate. Two, three or more smacks of carelessness.

(I'm not looking to troll here. I'm just comparing my extensive experience with yours.)

Re:Failing cards...

[bugnuts]

Could be crappy support hardware, or even the monitor doing something funky.

I've never had a vid card die on me either. And I also recently bought two 9800pros, replacing my geforces.

ATI hardware is far from crap. Software *WAS* crap but has come about 180 degrees.

Re:Failing cards...

[Feyr]

i'm no overclocker, the environment they were in is completly different (humid, dry, cold, hot), some were in work machines, some in my personnal machines (handled by the family) some in friend's .
none of the other users were overclockers either (hey, some of them can't even find the start menu)

that's only for the ones that physically died.

performances on everything i've tried was abysmal (including my lastest radeon 9000 with the lastest drivers) when the drivers don't crash the computer (works fine with a geforce)

the tech support is horrid, just getting an rma # takes ages.

dunno, some people are more lucky than others :)

Re:Failing cards...

[1000101]

Just my 2 cents:

I just requested an RMA from ABIT for my GeForce4 Ti 4200 128 MB today. I bought the card in March of this year and this is the second time I've had to return it. The first time took 3.5 weeks and I was stuck with my GeForce 2. I don't know if it's an ABIT issue or an Nvidia issue, but two times in three months (since the last failure) is unacceptable. Just goes to show that video cards do go bad.

Oh yeah, I don't overclock and have default mobo settings. The GeForce 2 has worked fine for years.

Re:Failing cards...

[andrewski]

ATI hardware is far from crap. Software *WAS* crap but has come about 180 degrees.

When I shit on a plate and rotate it 1/2 turn, it still stinks pretty bad. ATI are arrogant, provide half-assed support, and don't acknowledge that their drivers need serious work. NVIDIA are prompt, polite, and will actually take user suggestions.

Re:Failing cards...

[woodhouse]

You bought a radeon 9000 - ATi's most low budget card - and were surprised when the performance wasn't great? I have even less sympathy for you than all the people who bought geforce 4 mxes only to discover they were just Geforce 2s. In future, read a review or two before you pay for a video card.

Re:Failing cards...

[ix42]

I would have been surprised if the r9k was a slow card. Fortunately, my radeon 9000 is way faster than the Voodoo Banshee it replaced.

What? You mean there's people that upgrade more often than once every half decade? Weird.

Re:Failing cards...

[benzapp]

I concur. In fact, I have always used ATI cards since I got my first VGA card because they always worked well. Gave up the Mach8 a while back in my long lived AMD 386 DX/40, but the Mach64 still runs in a NexGen Nx586/100 box... It only runs as a firewall these days though.

Re:Failing cards...

[Kolgoth]

I'm losing ALL faith I have in ABIT lately... I went through 3 ABIT NF7-S mobo's this year - trying to get my primary PC up and running again... So ya, I seriously would reccomend something OTHER than ABIT - their latest products have been the pits...good specs/features/etc...but CRAP reliability... Back to the vid card topic - I've never had one die on me either...I've been an ATI customer since my Radeon 64mb DDR Vivo (aka the 7200 now)

Re:Failing cards...

[Forkenhoppen]

What type of power supplies do you have in these machines? I ran my 8500le with the one that came with my case for almost a year, and it crashed on me constantly. Upgrading to a 450W power supply fixed all my problems. Perhaps you're just not letting the card have enough juice?

Re:Failing cards...

[Feyr]

when i buy a video card these days, i expect them to outperform a geforce 256. the radeon 9000. as low budget as it is, has NO FUCKIN REASON NOT TO. i'm no performance whore, i don't give a shit if my unreal tournament doesn't do 160fps. but damnit, to hell if i'm gonna accept choppyness at 800x600

i'll stick to nvidia, from which i can get a 50$ card three times better than my old one

Re:Failing cards...

[Methuseus]

You get choppiness? never had choppiness on anything that wouldn't be choppy on a GF3 or lower.....

Re:Failing cards...

[Doppleganger]

Heh.. I finally upgraded my Voodoo 3000 to a GeForce 4 when Neverwinter Nights came out. Just hadn't had a need to do it earlier, it ran everything I'd thrown at it just fine until then. Meanwhile, I think my roomate had gone through three or four video card upgrades...

Re:Failing cards...

[ukyoCE]

I've had two cards die on me. One TNT2 that was about 2-4 years old. One day during finals when I was trying to play Serious Sam I started seeing wireframes shooting through the scene all over the place. That would get worse and the comp would freeze/die within a minute or two. If I didn't use 3D it worked fine. I never figured out why that happened...and I believe I still have it lying around if anyone thinks they know a way to fix it.

After that I bought a asus geforce2 with TV-card. I got the card home and when I got around to playing with the TV programs I found out even though the box says "windows 2000" on the side, none of the TV features run in windows2k. fuck that. and the store wouldn't take it back. I was pretty pissed, that being a 200$ video card.

So I put the Geforce2 in my second computer (where it did get some use at lan parties) and bought an ATI Radeon All-In-Wonder. MUCH better TV programs, by a mile. But shortly after I got the radeon home it died. I'm pretty sure it got fried in a small thunderstorm. But since it was only about 2 days old I took it back to Best Buy and said "this didnt work" and got a new one :)

That said, i'm back to using the geforce2 at this point because WineX only really supported Nvidia cards. (havent used it in a while though so that might have changed).

As far as running native apps like Quake3 in linux, the Radeon and Geforce2 had basically identical framerates.

Re:Failing cards...

[mwood]

Indeed, his failure rate seems excessive. My decade-old Orchid Fahrenheit 1280 is still going strong, even after I risked destruction by filling the open DIP memory sockets myself without any discernible static protection. :-O

Whatever happened to independant benchmarking?

[binaryDigit]

Why do they bother with these "standardized" benchmarks. We already know that the manuf. tend to gear their products towards scoring well on these things. That and from a content pov, anybody with the requisite hardware could do what they did. Whatever happened to the days when a group with solid domain knowledge would take some products and run it through their "own" benchmarking? Instead of using some canned 3DStudio simulation benchmark, find a bunch of models you've created and test them out. Run the cards through tests that YOU (not YOU the reader, YOU the ficticious reviewer) know are important. In this way people get a MUCH more realistic feel for what type of performance they can expect and the reviewer actually has some value added to doing the review in the first place (not just running the same thing that the eight "other" benchmarking sites do).

Consumer Card Comparison

[heli0]

"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?

Re:Consumer Card Comparison

[niko9]

The cards, as per the article, point out that they are identical (besides the DVI daughterboards) to the consumer versions, with "optimizations" done to the hardware, plus the optimized drivers. You'll notice that neither card has a beefier heatsink

I wonder if the "hardware optimizations" are FUD, since you can get the FireGL and Quadro orkstation drivers from their websites for free.

Re:Consumer Card Comparison

[heli0]

"I wonder if the "hardware optimizations" are FUD"

I had similar thoughts which is why I would be curious to see their consumer level counterparts benchmarked along side them.

Re:Consumer Card Comparison

[Glock27]

It would have been nice if they also benchmarked a $400 GeForceFX5900-256MB and a $425 Radeon 9800Pro-256MB then. (current prices from pricewatch)

Just for grins, it'd be great to see the Radeon 9600/9800 benchmarked in the new Apple G5s. Those also claim to support pro level applications.

The dirty little secret of the graphics chip industry is that these cards are really no different from the 'consumer' versions. It is simply a matter of the pro driver sensing a firmware dongle, then enabling the pro features like fast antialiased linedrawing for CAD.

So, I'm curious to see how Apple CAD applications fare with the 'consumer' cards in the G5s. The G5s should run these OpenGL benchmarks just fine (so should the G4s for that matter).

Re:Consumer Card Comparison

[paradesign]

Gamer boards work fine in DCC apps, as long as you dont push them too hard. Its not that they dont work, i mean openGL is still openGL regardless the card. BUT, there may be bugs in the app that only appear on the non 'pro' boards.

how do i know? well i run Maya, 3D S.Max, SolidWorks and Alias StudioTools on both sets of cards. at work on quadros, at school on firegls, at home on a geforce 4. the only thing ive noticed are small bugs at home, like funny screen draws, and wireframes that dont dislpay 'perfectly' (i love line AA, a 'pro' feature.)

well, why dont you softquadro? you ask me. cause the apps run well enough to make it not worth teh trouble. sure id love a quadro of my own, but a students wallet is always empty. well that and FPS arent 'life' in DCC apps like in games.

Re:Consumer Card Comparison

[mnemonic_]

Price has never been only about the work required to produce the product. Price is determined by the amount people are willing to pay and the value which they perceive in the product. In the high-end 3d graphics industry, cards costing several hundred or a thousand dollars are the norm, so that is why ATi and nVIDIA chose to price their cards the way they are.

Re:Consumer Card Comparison

[gl4ss]

yes, at least with earlier nvidia boards you could simply solder one resistor(or unsolder, can't remember which way) to trigger it into being a quadro.

sounds like a scam? yeah.

Benchmarks mean nothing

[Nishida]

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.

Re:Benchmarks mean nothing

[mnemonic_]

Benchmark optimization can only go so far. With non-synthetic benchmarks, manufacturers must attempt to cheat without it being noticed in normal usage of the software. Synthetic benchmarks have only one or maybe a few ways they can be run, so for example objects that are obscured could be omitted from a full rendering, without the viewer noticing. Non-synthetic benchmarks have infinite ways they can run and analyzed, generally preventing such cheating (when's the last time we've heard of cheating in a game benchmark? quack3.exe which was years ago?). If optimizations can be made without any noticeable effect in any benchmark run in any way, then I see those as standard driver efficiency improvements that are useful in the real world and not driver cheating at all.

Also, why are the numbers meaningless if all of the hardware used isn't the same as that of their audience? The purpose of the review was to compare the performance of the two video cards. As long as every other piece of hardware is the same in the systems used (assuming there were 2 computers and not just 1 with the vid card swapped), the video card is the only variable and is the only thing that will significantly affect performance.

Um.... WTF?

[Raxxon]

I thought FireGL was from Diamond, not ATI. :p

In fact, I'm holding a FireGL right now... I'll sell it for $200! Less than what ATI charges.

god I wish peole would think before naming things sometimes.... USB2.0 vs USB2.0 "Hi-Speed"....

Re:Um.... WTF?

[Quarters]

Wow. I know that Slashdot readers don't read the linked articles, but you take the cake. Not only did you not read the article synopsis, you didn't even read the headline.

It's the new ATi FireGL X1 that the review is using, not the extremely old and outdated Diamond FireGL1.

Re:Um.... WTF?

[Methuseus]

Moderators and anyone else who takes this seriously: IT IS A JOKE!!

now, that said, please stop thinking this guy was being stupid. He wasy making a funny based on the names of the cards.

Re:Um.... WTF?

[soulsteal]

Diamond's was bought by S3, who split it and sold the video portion to ATi.

'Nuff said.

Re:Um.... WTF?

[Raxxon]

I'm sorry, you don't seem to recongise sarcasm when you see it nor do you remember the card I'm talking about. :p

The Diamond FireGL 1000 and the 1000Pro were the cards I was refering to.

Thank you, drive through.

Re:Um.... WTF?

[Raxxon]

I knew that Diamond was bought by S3, but I don't remember S3 selling that aspect of the business to ATi.... Hrm. Learn something new every day.

ATI Catalyst Beta Program

[niko9]

I'm sorry for having to reply to myself, but the link to rage 3d shows that ATI has recently called for Linux Catalyst driver beta testers.

They're not saying if they are going to support all the multimedia features, like TV Out, capture, but it lookm s like it might be going in that direction. You have to sign a NDA to be elgible.

People have always bitched on the mailing lists that their AIW cards were half-ass supported, so this might be the turning point as far as these cards are concerned.

Re:ATI Catalyst Beta Program

[TrekkieGod]

the link to rage 3d shows that ATI has recently called for Linux Catalyst driver beta testers.

Recently? That post was done in October last year according to the time stamp.

And ATI Linux drivers are *still* horrible. I'm yet to be able to properly play a game on my 8500. Neverwinter Nights client locks up at the title screen, Unreal Tournament has some really weird artifacts...hmm...I haven't tried any other games, but as we all know, the choices are not plentiful. Not to mention I can't use tv out...

Re:ATI Catalyst Beta Program

[niko9]

The time stamp clearly shows August 11 2003.

Now, as far as your driver troubles, are you using the November 2002 drivers? If so, please, download the FireGl drivers. Click on Linux, then FireGL, then the first card on the list. You could also follow this link.

They were updated last month, will work with your Radeon 8500 (I have both a 9700 Pro and a 8500), perform better, and have fixed Xv video support. They also convert nicely with alien if your using Debian (which I am).

Re:ATI Catalyst Beta Program

[mczak]

Recently? That post was done in October last year according to the time stamp.

No, you got that wrong. Oct 2002 is the date Catalyst Maker registered, the post was made Aug 11, 2003 - I'd call that recently...

And ATI Linux drivers are *still* horrible. I'm yet to be able to properly play a game on my 8500. Neverwinter Nights client locks up at the title screen, Unreal Tournament has some really weird artifacts...hmm...I haven't tried any other games, but as we all know, the choices are not plentiful. Not to mention I can't use tv out...

You should try the drivers here: http://www.schneider-digital.de/html/download_ati. html . This site has always newer drivers than ATI's own site, and the latest driver apparently fixed the nwn lockups (on 8500/9000 cards) for some people. (Though currently there is an even newer driver available than that at the suse download site.)
That's certainly not to say the drivers are perfect, however they do get better.

Re:ATI Catalyst Beta Program

[TrekkieGod]

The time stamp clearly shows August 11 2003.

Ack...yeah, sorry about that. As another poster who replied to my message said, I got confused when looking at the date that the poster joined the forum

Now, as far as your driver troubles, are you using the November 2002 drivers?...

Actually, I am using the latest ones. The November 2002 ones don't work with XFree86 4.3. I still have issues though, but it makes me happy to know they're still working on it. I thought ATI had pretty much abandoned the whole Linux effort, when I noticed that they didn't even bother to include the firegl drivers on the download page for 8500 drivers

Re:ATI Catalyst Beta Program

[FroMan]

I have a 9000 pro which seems to run NWN decently. I haven't played more than an hour at a time though. I was unable to get the game to work with drivers less than 3.0. If you are using gentoo try:

ACCEPT_KEYWORDS="~x86" emerge \=ati-drivers-3.2.4

I have only been using them with the most recent gentoo-sources.

The 3.0 drivers had a pretty bad bug in them for RTCW-ET which made it completely unplayable. The newest drivers though I can usually play for an hour+ without any lockups.

Obligatory grammar nazi comment

[Wesley Felter]

I Love it when a Writer decides to capitalize Random Words when writing an Article.

Re:Obligatory grammar nazi comment

[Coneasfast]

hehe, right in the first 2 sentences it says:

Professional Graphics Cards are sort of an oddity to us here at HotHardware.com. Admittedly, we don't get to tinker enough in this area, since frankly the market that is served by these products is a specialty niche, where there just isn't the same level of Marketing buzz, as in the Consumer Graphics space.

You are certainly right, no need to capitalize 'Graphics Cards', nor 'Consumer Graphics', nor 'Marketing buzz'. It's pretty bad throughout. For some reason my respect for this article has dropped because of this. If it was one or two silly mistakes it's ok but this is quite a lot of false capitalization.

Re:Obligatory grammar nazi comment

[Methuseus]

Maybe the writer's native language is German or something comparable? I know when I was taking German in high school that I would Randomly capitalize Nouns because you have to capitalize all nouns in German, but only certain ones in English, so sometimes a capital would slip through.

The obvious question:

[Michalson]

Where exactly are they getting "new" from. The FireGL X1 card may as well have cobwebs on it. The current workstation cards being pushed by ATI are the FireGL X2 and FireGL T2 (X2 being highend as the X1 was, T2 being targeted at the budget market). Claiming "NVIDIA still has a stronghold in this market" is deceptive at the very least. Would you find a CPU benchmark accurate if they compared an Athlon XP 3200+ with a Pentium 4 @ 2.4GHz and concluded that AMD was leading the market?

Re:The obvious question:

[GoSpeedRacerGo]

The artical is flawed in many ways but the Quadro FX 3000 is shipping now and it thumps the FireGL X2 (as does the FX 2000). It is pretty hard to argue that NVIDIA isn't dominating the workstation market at this point.

Re:The obvious question:

[Methuseus]

I want some proof goddamnit!!

gimme a link or something, I'm hungry for some benchmarks and intelligent (and I mean intelligent) rhetoric on each card.

Re:The obvious question:

[GoSpeedRacerGo]

http://www.spec.org/gpc/opc.data/vp71/summary.html

3Dlabs

[Distinguished Hero]

They should have also benchmarked the latest 3Dlabs cards in order to give us a proper frame of reference. For all we know, both these cards could be providing inferior performance compared to the latest Wildcat; good gaming performance doesn't necessarily translate into a good professional video card.

The Wildcats are also cheaper: $899 for the 512 MB VP990 Pro and $499 for the 256 MB VP880 Pro or the 128 MB VP970 (from the 3Dlabs eStore) compared to $530 for the cheapest 128 MB ATi FireGL X1 and $1250 for the cheapest 128 MB nVidia Quadro FX 2000 (the 256 MB variant was used for benchmarking).

Anyways, these aren't even ATi's and nVidia's top of the line cards; ATi's is the FireGL X2-256 and nVidia's is the Quadro FX 3000.

Re:3Dlabs

[GoSpeedRacerGo]

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.

Re:3Dlabs

[Compuser]

That's news to me. Can you give me reference to
a Pro/E or 3DS benchmark site for this? I would
be very curious. I am looking for highest
quality rendering with 1 Million polygons and
8 lights. None of those NVidia tricks with
reduced quality texturing and what not. Please,
if you know of a good reference on this it would
help me for real. Thanks.

Re:3Dlabs

[tinrobot]

I test a lot of graphics cards for various magazines and publications. The Wildcats lost the performance edge right around the time Intergraph sold out to 3DLabs. They seem to lag behind or at best are equal to the top nVidia cards. Right now, they're lagging until the next generation of the VP series comes out. The VP cards were also notoriously unstable, though recent drivers have corrected most of that.

The new nVidia cards seeme to be filling the niche that the original Wildcats used to fill - they even now have genlock, which used to be the high end Wildcat's big advantage.

Re:3Dlabs

[GoSpeedRacerGo]

http://www.spec.org/gpc/opc.data/vp71/summary.html Note that there is only one system submitted with Wildcat 4 and it gets significantly beat by ATI and NVIDIA, especially by NVIDIA and especially in the Pro/E and 3ds max components of the benchmark suite.

Also note that there are ZERO systems submitted with Wildcat Pro.

What systems are submitted by the workstation vendors is almost as significant as the scores that are posted. Any time a system/card starts getting beat too bad on viewperf the results for that system are pulled. i.e. the posted solutions are the cream of the crop.

Re:3Dlabs

[PCBman!]

Ace's Hardware has a review you might be interested in reading:
Professional Grade Revisited

To The Misinformed

[palp]

Since I'm seeing a lot of this, a note to the uninformed/misinformed who didn't RTFA or even much of the blurb:

These are NOT cards for gaming, they are for professional graphics work. Very different market, so please refrain from telling us how you don't need a high end video card or don't play video games. It's of no consequence.

Re:To The Misinformed

[niko9]

Since I'm also seeing a lot of this, a note to the uninformed/misinformed who did RTFA or even much of the blurb, please don't point out the obvious.

It's as about as redundant as asking if a one legged duck swims in circles.

Re:To The Misinformed

[palp]

It's as about as redundant as asking if a one legged duck swims in circles.

It doesn't if it's just sitting there with a dumb look on its face.

Re:To The Misinformed

[palp]

You haven't seen the ducks we have here in south FL, then..

Look at this ugly fucker.

Re:ati vs nvidia

[DataPath]

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.

Re:ati vs nvidia

[Feyr]

yep, that's the first thing i did

Re:ati vs nvidia

[Gherald]

> What are you basing that conclusion on, cards from 4 years ago?

FYI:

Hardocp

Tomshardware

A relevant forum discussion

Re:ati vs nvidia

[xSauronx]

i loved my 8500...the only reason i switched to a geforce 3 was drivers, when i installed one series of the catalyst 3.x drivers, it literally ruined my system the same drivers ruined my brothers as well, we both had to reinstall. the 8500 was a tad better with aa/af i think than this gf3, otherwise the performance is so close i cant tell, and the nvidia drivers have never ruined a thing for me

Re:ati vs nvidia

[Gherald]

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...

Re:ati vs nvidia

[mojowantshappy]

I have nothing against ATI, but the reason I buy nVidia cards is because of the linux support. Last I checked the ATI linux support wasn't very good. Then again, they may have been improved in the last year. Doesn't really matter because the last card I bought was my GeForce 2 two years ago.

Important tests missing.

[DraconPern]

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...

Re:Important tests missing.

[Forkenhoppen]

At 1280x960, this is a known issue with CRTs. At all other resolutions .... that's weird.

Re:Important tests missing.

[Methuseus]

I've tried my AT Radeon 9000 pro at all resolutions, and the only time I had any bad signal quality was with a card with a bad DVI output. Once I RMA'ed that board and got a new one (the new one looked a lot cleaner and more professional than the first one I got) I had perfect quality on both monitors, except when I plugged 2 monitors, as well as a TV into the card. But this is a known issue as ATI didn't build the chip to push enough power to drive the TV and both monitor outputs at the same time.

Awesome!

[john_smith_45678]

Now all my DirectX 9 vaporware games will run even faster! (300 fps wasn't fast enough anyhow).

Re:ati vs nvidia

[EverDense]

The article you are quoting turned 1 year old yesterday.
You should find a more recent statement to cite.

ATI drivers have come on in leaps and bounds in the last couple of years.

ati firegl x1 didn't work for me.

[yo5oy]

just chiming in with some news on ati firegl x1 card and how it just didn't work for me under linux or win32. the drivers are just not up to production level. i would call their tech support, which in my opinion is very good, and implement all their suggestions and the machines would just randomly reboot in both 2d and 3d applications. The same workstations had 30 day uptimes under win2k with sp3. Softwares used were 3dsmax5, photoshop 7, and lightwave 7. we ended sending the cards back for a refund. we lost a week of two machines being down and all my and another techs time trouble shooting and recreating problems.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Re:ati vs nvidia

[Gherald]

On the software end of things? Never. He's a developer. If he thinks Nvidia is easier to work with, great. But he damn well better support my 9800 well if he expects me to run games that use his engine.

Re:ati vs nvidia

[onash]

my Compaq Presario kept giving me BSOD and sometimes rebooting while playing games that use DX/9 with ATI Radeon Mobile 7500 (eve-oneline).. and after one month of e-mailing ATI support (with over 20 e-mails of mostly bullshit from ATI and Compaq) i found a fix.. unofficial ATI drivers that are not from the laptop vendor (Compaq).

after this, i am never ever buying ATI again.. even though its not just their fault, Compaq should had supplied new drivers months ago, it just showed me that i shouldn't buy things from a company with bullshit drivers from vendors only policy.

I've missed the Radeon drivers since i bought this machine.. i guess i will miss them too when i'll try to switch to Linux next week!

My hat is off to you, sir

[sbszine]

You may the be the first ever /. grammar nazi who can spell the word 'grammar' properly.

Three cheers!

Comment removed

[account_deleted]

Comment removed based on user account deletion

Re:ati vs nvidia

[Gherald]

one year ago

Leave reviewing to the big boys

[SuperBanana]

HotHardware

Um, pardon me, but...who?

Call me when you've got benchmarks from a real magazine(say a CAD/CAM, 3D graphics and/or animation, etc related magazine), and not two-guys-in-a-dorm-room-who-write-reviews-for-kick backs websites who run Unreal Tournament to benchmark professional graphics cards.

Case and point is their 'testbed' system: they used a "DFI LAN Party 875Pro" motherboard. They used Pentium 4's instead of workstation-class Xeon processors. IDE drives instead of SCSI. Folks, that's NOT a "workstation". A dual Xeon cHomPaq is a workstation.

Oh, and the benchmarks? One no-name benchmark, and 3D Studio Max. Oh, and Unreal Tournament. No fill rates, no polygon counts, no NOTHIN. No mention of Linux, which is tearing into the market like crazy among top computer animation houses.

This is pathetic- they're just a bunch of guys who compile daily linkages to other cheeseball review sites. They have no industry background, no experience, no nothing...just a P4 3GHz and a (probably pirated) copy of 3D Studio Max.

Re:Leave reviewing to the big boys

[sl3xd]

Solidworks is certainly not a no-name program; I'm not even a mechanical engineer (its primary market), and I've known what solidworks is for quite a long time...

Re:Leave reviewing to the big boys

[afidel]

One no-name benchmark, and 3D Studio Max.

No name???
Guess you don't do much with CAD/CAM, Solidworks is one of the most featurefull CAD apps out there, its usefullness is second to only possibly CADIA. At my last job the physical design guys modified their AP encasement after running Solidworks simulations which pointed out non-optimal heatflow from the CPU to the case exterior. They built up the case from components whos exact thermal, electrical and other properties were in the materials database.

You are however correct that this was not a good test for workstation class machines.

Re:Leave reviewing to the big boys

[O2dude]

Folks, that's NOT a "workstation". A dual Xeon cHomPaq is a workstation.

If you're going to nitpick... I;d say that a SGI Tezro or a Sun Blade 2000 is a 'workstation'. They have 64bit non-Intel CPU's with massive L1/L2 caches, they have overengineered cases, they have system architectures designed for huge bandwidth, and their listprices are prohibitive to the point of extortion.

A dual Xeon or a single P4, be it from compaq/hp or boxx, is still only a boring PeeCee.

As whether it still makes sense to buy a true workstation or a boring PeeCee for tasks that used to be called 'high-end', well that's another discussion entirely.

Re:Leave reviewing to the big boys

[mnemonic_]

SCSI drives only have significant benefits when used in a server environment with numerous portions of the disk being accessed in rapid succession. When you're saving to a few files with in a 3d graphics creation environment, IDE is fine (which is why there has been a huge decline in SCSI usage across the board, especially with the advent of SATA and good SATA drives like the WD Raptor).

Someone else pointed out that SolidWorks is by no means a "no-name" benchmark. I'll add that fill rate and polycount benchmarks are near meaningless when you can do real benchmarks with real applications. They're about as useful as using clock speed to compare processors.

Might I ask whether you have any "industry background", or experience, or anything? Judging only by your ignorance of SolidWorks, I'd say "no".

Comment removed

[account_deleted]

Comment removed based on user account deletion

The nVidia "Hardware" tweaks consist of...

[pixel.jonah]

...a couple resistors and maybe hardware ID string. At least that's how it it was in the GF2 and GF3 cards.

A couple of years ago when I got my GF3 and was taking some Maya Classes, I learned that nice tidbit. The "consumer" drivers are D3D optmized while the "PRO" drivers are OGL optimized. and will switch routines depending on how the hardware is identified. So, at the time, soldering a couple of resistors would change a consumer card into a Quadro.

UNTIL, some tricky Russians patched the driver to allow you to run the Quadro OGL optimized routines on the GF3 with measurable increace in performance.

Disclaimer: While this is ofcourse unsupported, it is a nice way for students and other budget 3D modellers to get some more performance.

Re:The nVidia "Hardware" tweaks consist of...

[GoSpeedRacerGo]

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.

Re:The nVidia "Hardware" tweaks consist of...

[pixel.jonah]

Bummer...

It was a great example of "general purpose" hardware though.

Peny wars

[Anonymous Coward]

do we really need these video card peny wars on Slashdot, is this "stuff that matters" by any accounts?

Professional cards

[Ian-K]

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.

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 /. article on it), while the latter have been traditionally good with it.

It would have been very interesting if they'd included the VP990 Pro or the VP970 in the comparison...

Trian

Re:ati vs nvidia

[iantri]

I have to agree that they have crappy drivers. I bought an ATI TV Wonder VE with Remote Wonder and gave up and installed the opensource generic drivers (http://btwincap.sourceforge.net) since ATI's drivers and software froze and generally misbehaved so much.

The scary thing is; it's better than an older card I had.. built-in hardware MPEG decoding and so you needed to use ATI's player software for that.. if you breathed the wrong way it crashed..

Re: Not a lot longer than under Windoze

[Dodger73]

I downloaded the drivers for my 9700Pro, installed them (RedHat 8), configured and restarted the X server, and that was it. Took me about 10 minutes total. Another 5 minutes to add a device block to xfree68.conf, and dual monitor was running too. The only real problem I have is, that I can't seem to manage to run a different resolution on each of my monitors - otherwise the drivers are working just fine.

Differences between Pro and Non-pro cards?

[porp]

I was just wondering what exactly are the differences, excluding price, between these workstation cards and the common desktop cards that you can pick up at Best Buy (radeon/geforce3/4/fx)? More precisely, can these things run games and other desktop tasks like video encoding, and if, say, one had much cash, would buying these cards be a better idea than purchasing a desktop card?

porp

Re:ati vs nvidia

[13Echo]

Not so...

nVidia's recent Linux driver sets have been utter trash this year. Why don't you explain the "best in the business" stuff to those that have had continuous system lock-ups on Gnome 2 desktops because nVidia treats its "Linux customers" like test subjects for its Windows driver base.

You may not realize it, but ATi's drivers are more stable than nVidia's on Linux. Shoot, even the lowly PowerVR is writing Linux drivers that are far more reliable than nVidia's. The fact that nVidia updates their drivers frequently is nice, but they can't seem to refrain from breaking something new each and every time that they do it. Don't believe me? Check out all of the Gnome 2 message boards out there.

The fact is, you're going to pay for it in some form of stability problem when you start running wonky drivers with non-standard features and rendering code. The nVidia drivers don't draw off of the DRI mechanism either, which is unfortunate. Their proprietary rendering mechanism is likely the cause of many of the strange instability problems. We see this problem every day, on a support forum, in which some user claims that they are having problems with desktop hangs and rendering anomalies. Nine times out of ten, it's a problem with the closed-source nVidia binaries. Asking someone to test the "NV" driver instead of the "NVIDIA" driver almost always corrects the problem.

I'm not trying to flame nVidia. I like their products... But until they straighten out their driver problems on Linux (their Windows drivers are quite good), I'm not buying one of their cards.

ATi isn't much better, however. Their commerical drivers haven't been updated since November, though the FireGL and Gatos drivers are good, alternate choices in many cases, though they sometimes lack features. ATi, however, has often been pretty generous with providing documentation about some of its products for some features of the chipsets. And, though ATi's commercial driver options have been limited of late, what they do have is typically quite solid on Linux. I'm holding out for a unified driver update for XF86 4.3 though, before I even consider one of their cards.

Not enough competition

[archen]

Other benchmarks include a leaf blower and a flaming jar of gasoline. The leaf blower actually did quite well in the noise tests against the Nvidia card but lost out due to the fact that it consumes less power than the video card. Ati unfortunately did not fare as well, and lagged behind with the noise factor. An ATI spokesman recommended that the card be coupled with a "cheap ass Athlon CPU fan" which would develop a good rattle to help the card become more obnoxious.

The jar of flaming gasoline also did pretty good in the heat department against both cards, but unfortunately had to be refilled which was considered a drawback. Aside from that the life like animation the fire produced only ran at one frame rate, but was always consistent. Unfortunately the jar of gas lost out big time in the cost arena, but it seems that can be compensated for by tossing $1 bills into the flames at various intervals to get the costs up higher.

Some wondered why we didn't benchmark a toaster as well, or instead of a jar of gasoline - but as we pointed out before, a toaster is far to practical to compare in a contest of flushing money down the toilet.

Re:Not enough competition

[paradesign]

lol, wheres my mod points when i need 'em.

+1 funny i add.

Only 1 benchmark matters.

[Raven42rac]

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.

NVidia epoxy, how I trust thee, just like ATI

[SlashCrunchPop]

My late NVidia TNT2 was dynamite, it got so hot the heatsink fell off, which resulted in a permanent visual effect that I can best describe as psychedelic shadow. Now, why would I trust NVidia to use better epoxy on their latest GPU, which dissipates 9 times more heat than my late TNT2?

And why would I trust ATI to get fully functional drivers out in less than 2764 attempts or before my unborn children start school?

Choosing between NVidia and ATI is like choosing between enduring hot needles and the Chinese water torture. In the end you die and someone makes a buck.

Re:NVidia epoxy, how I trust thee, just like ATI

[GoSpeedRacerGo]

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.

Re:NVidia epoxy, how I trust thee, just like ATI

[SlashCrunchPop]

True, but AMD could also leave cooling up to whoever buys their CPUs to assemble systems for the retail market. But do they do that? No, quite the opposite. Why? Because those who integrate your product into a substandard system will damage your reputation as well when their systems start dying because of their neglicence and substandard production.

It's also quite laughable that a TNT2 cannot last a minute without a heatsink without getting damaged beyond repair.

Re:NVidia epoxy, how I trust thee, just like ATI

[SlashCrunchPop]

It was epoxy, I could not believe my hands. Anyway, read my other reply to another guy's similar comments regarding board making.

I agree with you on the other points, remember Day of the Tentacle? Indiana Jones? Monkey Island? Wing Commander? Ah, those were the days... When a Diamond Stealth with a mere 1 MB of video memory did the job and did it well.

Re:NVidia epoxy, how I trust thee, just like ATI

[SlashCrunchPop]

You're comparing apples to pears, but just for argument's sake: just so you know, an Intel PIII CPU can survive without a heatsink without sustaining permanent damage! I guess now you're going to tell me you meant AMD Athlons because somebody told you they burn out within a couple of seconds.

Re: Not a lot longer than under Windoze

[13Echo]

You may be able to try the "xrandr" command along with the "-d" flag, that lets you resize a particular display (e.g. 0 or 1, in most instances. Check your XF86Config file). If it works, then you can add it to your Gnome "sessions" startup options (assuming you use Gnome on RedHat).

"me too"

[Large Green Mallard]

I'm a systems person at a university department.. over the last 2 years, with all the various machines at work and "friends", I've probably been in contact with about 600 -700 machines.

Of these, I've had to replace 3 video cards. One early AGP Matrox, an AGP TNT 2 M64, and a PCI S3 VirgeDX. All of them more than 3 years old. And the TNT2 was a maybe, but after fitzing with Windows for 2 hours, and plonking in another TNT2 and having it work perfectly, replacing it was an easier option all round.

OT, but if you want to know what dies a lot, it's hard drives, mice, monitors and power supplies... about 30 mice per year and about 15 each of hard drives, monitors and PSUs.

Benchmarks are fine... but what about accuracy

[DrJohnnie]

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?

Re:ati vs nvidia

[DataPath]

I'll be a little less wary of ATI cards when people stop saying that the drivers have been getting better over the past few years and start saying "they're great!" I don't mean person - I mean people.

In all fairness, I've been hearing it more and more "works great for me! no problems!" (windows-side, of course)

My statement, though, was referring to the software developer aspect. The API's change very little from year to year. The hardware that implements the API does.

And that I should find a more recent statement to cite, well, Carmack doesn't make a habit of shouting praises about the hardware he develops for. Seriously - just TRY and find a statement from ANY respected developer about the quality of nVidia's or ATI's API.

Re:ati vs nvidia

[DataPath]

maybe ATI's drivers ARE better on linux. I don't know - I haven't gotten them working yet. Not a criticism of the drivers, just that there's an extreme shortage of documentation on the internet about how to install them. Of course, there are a number of options, Gatos, ATI, the DRI driver that ships with the kernel. I've only tried ATI's and the kernel DRI driver and not had much success.

Re:ati vs nvidia

[Uncle Gropey]

Yeah I bought an ATI card (FireGL 8600) only to later find out that it wont work on win95, win98, or winME. I need to boot into one of those for certain games and apps sometimes but like... it ain't gonna happen with this card in there. With that said, it's a decent performer for it's price but sheesh... at least put a warning sticker on the box or something guys!

for you cheap bastards...

[Alpha_Nerd]

There's no need to actually buy a FireGl X1 as you can easily soft mod a Radeon 9700 Pro into one. instructions can be found here.

Quick trivia question

[Sycraft-fu]

What is the difference between current (as in P4 architecture) Xeons and normal P4s?

Answer to rhetorical question: Xeons do multi-processor and do not currently run on the 800mhz bus.

That's it. Just receantly Intel did release one with double (1MB) the cache, but any Xeon slower than a 3.06 is a 512k, just like the normal P4.

A normal P4 is fine for a workstation, in fact Intel notes it as being a workstation chip. Given the higher memory bus it can even be faster for some tasks. Xeons are basically only if you need MP or if you have some app that will benifit form the larger cache (and then only if you buy the fastest model).

Also I fail to see how SCSI is relivant to a test of a graphics card.

P.S. Solidworks is not no-name in the CAD community.

Re:ati vs nvidia

[Methuseus]

The fan stopped working on my Radeon 9000 Pro because I accidentally bumped the connector for the fan's power. All I got was artifacting, no other errors. When I got the fan working again, it's prefect. Still not the best performance, but mangnatudes better than any comparable card from nVidia when I bought it (now the 5200 FX is out, so I dunno, nobody seems to compare the 2 cards).

Ati's Linux drivers...

[Kynde]

Could someone with recent Ati 3D products say a few words about the current state of their linux drivers? Any glitches? Stability? Do they provide OpenGL headers for their libraries? Tv-out functional? etc...

Personally I'm still a content Nvidia user, solely due to their drivers, they even run smoothly on top of 2.6.0-test3 (with the minion.de patch), _but_ I'm seriously thinking about Ati the next time around, which is around the corner as Doom3 comes out, also for linux as you all know.

Re:ati vs nvidia

[Methuseus]

I personally feel ATI's drivers are currently better than nVidia's. If you asked me 6 months ago, I'd tell you they were about the same to me. Go far enough before that, and nVidia's drivers were better.

Also, keep in mind that most of what I say about drivers being better is the control panels they each have. I used to hate ATI's control panel, but now I prefer it even to nVidia's old one. the current one nVidia has limits your settings potential, and I like the option, even if I don't use it all the time.

As far as driver problems, I've only had problems twice. Once with an old ATI XPert 128 that sometimes didn't like D3D engines, and the other with nVidia's GeForce 2 MX 200 that didn't like OpenGL all the time. That is the extent of my problems with drivers themselves. The control panels I already talked about.

Re:ati vs nvidia

[680x0]

Nine times out of ten, it's a problem with the closed-source nVidia binaries.

ATi isn't much better, however.

Ok, so which card should a discerning Linux user get? I'm shopping for a new card. I want good 3D performance, as I'm teaching myself OpenGL programming. Matrox? Can I get a decent card for something in the $100 range?

Re:ati vs nvidia

[Methuseus]

Umm, nvidia actually says the same thing about laptop drivers: get them from the manufacturer!

My friend's laptop was hosed repeatedly when he tried loading the nvidia website drivers on it. You should always, with any integrated component, get the drivers from the manufacturer of the biggest part (the mobo for desktop integrated video/LAN/USB or the laptop for a laptop anything).

Boy, I hate to say this, but...

[Nunar]

Jeez, I'm too drunk to do anything but reply, but the link has a comma after the www. I can't point-and-click to the video hardware review that I so desparately need.

"Please don't hate me just because gin is my friend." - Me

Re:ati vs nvidia

[Methuseus]

did you look for hardware/software requirements on the box? it will usually say "For win NT/2000/XP" on the box in plain sight.

That's why the board interface is secret

[Animats]

Yes, the consumer and "pro" boards are almost identical. In some earlier NVidia boards, as others have pointed out, you could change a jumper and make a "consumer" board into a "pro" board. And all the jumper did was change the board type, which the driver read.

This is why NVidia won't provide the hardware interface specs for their boards - the consumer/pro distinction would disappear. That's why the Linux drivers are closed source (and drivers can't be written for some other OSs.)

I'm suprised NVidia still bothers. I expected that they'd give up on the distinction when ELSA tanked. NVidia bought into ELSA, which made "pro boards" with NVidia chips. ELSA became the sole vendor selling "pro" boards based on NVidia technology. ELSA then went bankrupt.

But instead of abandoning the market, NVidia still services it, although in a very limited way. There used to be quite a number of "pro graphics" board makers, but they've been eaten by the low end. So it's a minor cash cow for the graphics chip makers.

It's getting to be one of those carefully maintained artificial product niche things, like overpriced SCSI gear.

Re:That's why the board interface is secret

[Sevn]

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.

Re:That's why the board interface is secret

[Kuad]

Actually, you can make a faster array for cheaper using IDE by using many, many more drives in a RAID 10 array. Even EMC sells these now. You're probably going to see SCSI's dominance of large arrays go away in the near future. This will leave SCSI with the server internal disk market and not a lot else.

Re:That's why the board interface is secret

[Animats]

There's profit in making plug-in drives expensive. Look what happened to Device Bay, the standard for plug-in FireWire drives and drive bays. Adaptec hated it, because it would have killed their profitable SCSI board market.

Re:That's why the board interface is secret

[Animats]

NVidia can't provide hardware interface specs for their boards because they don't own the rights to do so.

Huh? NVidia develops their own IC designs, and it's about all they do. NVidia chips are actually fabricated by Taiwan Semiconductor Manufacturing Company. NVidia is also starting to oursource to an IBM fab in upstate New York.

NVidia has bought licenses on some intellectual property. They license some Artisan RAM interface designs, and some old SGI patents. But neither of those limits what they can disclose at the bus level.

A much more robust review/test here ...

[Lightman_73]

HotHardware, where they test pro graphics card with games... cool...

Now, for the ones who want a quite better review of the FireGL X1, QuadroFX 2000, FireGL Z1, compared to 6 others pro boards (including 3DLabs Wildcat VP970), Tom's Hardware has a nice one, dated March, 21st (so not only HW has an all but complete review, it is much late, too) :

Tom's Hardware FireGL X1 vs QuadroFX 2000 Review

Have fun...

Selective Memory?

[imsabbel]

Outperforming the new cards?
With 19fps in quake2 timedemo in 640x480?
The card SUCKED ASS n gaming.
It was only better than the OTHER professional 3D cards at that time, which didnt even run games because of lacking d3d drivers or abysmal fillrate at that time.

Re:Selective Memory?

[Ian-K]

I vaguely remember some Powerslide / Q1 / don't remember games. IIRC my mates were using a savage 4 (?), some voodoo banshees, stuff like that.

I don't recall the d3d performance being superb either, but did you try to use opengl?

Trian

Photorealistic games = better gameplay?

[Channard]

What interests me is if we keep getting better and better cards like this, will we one day get games which look so good so as to be indistinguishable from reality (albeit still on a screen). I certainly hope so because when/if this happens, games companies will have nowhere to go with graphics and will actually have to give more focus to making games more enjoyable. Fun to play instead of just flash, whereas the onus these days tends to be on graphics that take advantage of graphics card feature x.

pessimism

[mnemonic_]

On Slashdot, it's cool to be an irrational pessimist.

Re:pessimism

[EmagGeek]

On slashdot, it's hard to piss away karma, even when you want to :)

image shearing aka "tearing"

[mnemonic_]

The image "shearing" that you mention is caused by vsync being disabled. You can enable it in the control panel (assuming you're using Windows).

I second the parent post (NT)

[mnemonic_]

NT = No Text

Re:ati vs nvidia

[13Echo]

Well, there's no reason that you shouldn't buy nVidia, but you may be trying a few different versions of the driver in order to find one that doesn't hard-lock the machine. I guess that it's not as much of a common problem if you aren't using Gnome2.

ATi cards are generally pretty well supported, in most cases. With the exception of the newest cards, you should be able to find a driver set that works. As noted in a post below, ATi was holding a closed beta test for XF86 4.3 drivers. So, provided you can wait for a driver release, we will more than likely recieve some quality XF86 4.3 drivers in the next month or two, but I can't guarantee that. Otherwise, the 4.2 drivers are supposed to be pretty good, but they don't seem to work on 4.3 machines. The Gatos project is working on R300 drivers as well.

Matrox has parhelia drivers, but I couldn't tell you how well they work. Most Matrox users have been really impressed with the drivers for their older cards though.

Have you written a letter to ATI?

[renehollan]

I mean a real, honest-to- letter, sent via post, stating your beef?

Disclaimer: I work for ATI, though this is strictly a personal observation, and not even specific to ATI. You know the drill: I gotta shutup when it comes to what we do.

I have seen far more crap from the inside of all the companies that I have ever worked for. The consumer sees very little of this, and in a perfect world, none at all.

But, the bottom line is this: just which crap gets cleaned up and what stupid policies are corrected depends on what feedback is received. Griping in slashdot "feedback not" as Yoda might muse.

A lot more goes on "on the inside", some of which you might drool over. Your job as consumer is to push a company into the direction you would like to see it move.

If you think a company has poor quality control, say so. There are probably people inside who agree, and need "ammunition" to change things.

From experience...

[blueorder]

Our SolidWorks station that I use on a daily basis has an ATI FireGL X1 in it and I can say that this card can perform.

Our assemblies can get pretty big component wise and the FireGL keeps chugging along.

In the end, the buying decision between ATI and NVIDIA workstation vcards came down to price. After shelling out the huge amount or money for SW (I'm in a small company), justifying the purchase of a $1300 card was near impossible. Now, paying $400 for the FireGL, with great performance, made my managers smile (for a microsecond, at least)...

Few observations..

[clifgriffin]

The reviewers didn't seem to be quite the experts in the field they were commenting on. The article was short and used very few factors. I'd like to see a Tom's Hardware review. Even though ATI seems to have the edge right now with consumer video cards, I must say I'll probably stay with Nvidia. My past experiences with ATI's software and drivers has been bad. Besides that I can't afford to have the latest and greatest so it doesn't matter what I buy. ;-)

Re:ati vs nvidia

[mwood]

I love my Matrox Millennium II, old though it is. Next time I buy a non-bargain-basement card it will be a Matrox.

Those cheap ATI 7000 cards that were on deep discount everywhere a while back are good for office type WinXP boxes, though. They beat the pants off the SiS 530 onboard display hardware they replaced.

Re:ati vs nvidia

[680x0]

not as much of a common problem if you aren't using Gnome2.

Actually, I'm running RedHat 9, which I think uses Gnome 2, doesn't it? It does use XFree86 4.3.

I previously had an All-In-Wonder Rage 128, with which I used the Gatos drivers. It gave me pretty good 3D performance, but the docs seemed to suggest the DRI drivers were incompatable with the drivers I'd need to use to run Video4Linux and watch TV.

Guess I just need to keep reading Linux-specific reviews when I can find them. Thanks for the info.