PDA

View Full Version : ATI or Nvidia???


HapZungLam
01-07-2003, 04:04 PM
I personally got a G4 4200. i heard that there isn't a much different between 4200 and 4600.

Also, how much different between Radeon 9k and 9700?

Gentle Fury
01-07-2003, 04:52 PM
nvidia all the way....no contest

sedric
01-07-2003, 04:56 PM
ati seems to have the fastest cards across the board...regardless of price. i personally have the radeon 8500 128meg and am really happy with it. their drivers have gotten a LOT better in the last 8 months or so (IMHO).

As far as the higher end workstation video cards, im not too familiar so i cant offer an opinion.


hope this helps,

sedric

DeathCarrot
01-07-2003, 05:48 PM
the R9700 (non-pro) is quicker, its the 2nd quickest card out atm, after the 9700 pro. the ti4600 is comparable in terms of speed to the R9500 pro

there is quite a bit of difference between the 4600 and 4200, but bang for buck the 4200 is better.
im not sure about the differences between the 9k and 9700...
you can get a few benchies on a lot of cards at http://www6.tomshardware.com/graphic/20021218/index.html and decide which one's the best...

HapZungLam
01-07-2003, 06:06 PM
thx for the site. it is very obveious that 9700 is stronger. and AMD kills Intel

GregHess
01-07-2003, 06:15 PM
The 9700 pro is a great card...if your just playing games. However the instant you bring it into application situations, faults start appearing. There are known issues with a variety of apps. If I'm not mistaken, I've seen problems in both max, and in XSI with the card, I can't speak for maya or lw though.

In this regard, the real poll should be against a FireGL X1, vs a Quadro 980XL...as both of these cards are actually workstation designated.

wgreenlee1
01-07-2003, 07:05 PM
NVidia

HapZungLam
01-07-2003, 07:10 PM
i am going to get my card tobe soft quadro for sure. XSI will support it.

CgFX
01-08-2003, 06:47 PM
Originally posted by sedric
ati seems to have the fastest cards across the board...regardless of price... As far as the higher end workstation video cards, im not too familiar so i cant offer an opinion.


I think more often than not the NV cards are faster. R300 is a fast chip and all the varients built from that will be at or towards the top for gaming until GeForce FX ships this month.

For workstation, it is no contest. NV cards have been faster to date and I imagine a workstation NV30 will increase that lead.

elvis
01-08-2003, 10:04 PM
the top workstation cards are the ATi FireGL X1 and the nvidia quadro4 980XGL.

both are AGP8X cards based on the top of the line GPUs from both companies.

there are plenty of threads already about these two cards, with many a debate over which is better for what.

this is only my opinion, and entirely based on results from the web (as i don't have a firegl x1 to compare to my quadro4 cards, but that will change early march), but it seems with the latest driver release from both companies the ati is pulling ahead slightly in it's raw wireframe polygon throughput, while the nvidia still holds it's own, but does slightly better in heavily texture-shaded and complex lighting scenes. again, this is only from what i've heard, read and seen, so don't take it as gospel (hell, even i don't trust it until i have a chance to test it myself).

the release of both the r350 and nv30 chipsets from both companies is very close, which means there could well be workstation models of each within 6 months (indeed a few forum members have already confirmed the plans for nvidia to release a "workstation geforce FX", and i'm hoping ati follow suite with their r350 chip).

at any rate, chip performance is going up, and prices are coming down (slowly). exciting times for all. :)

but yes, looking at the gaming variants (geforce4ti and radeon 9X00) i'd go with the geforce4 for CAD/DCC, and the radeon for games. depends what you want most from your system.

CgFX
01-09-2003, 03:11 AM
We were told that ATI's FireGL X1 256 MB has been delayed so I don't believe any R350 product is coming soon.

elvis
01-09-2003, 03:52 AM
that's a real shame. if nvidia bring their geforceFX based workstation card out while ATi's top dog is still the X1, unless ATi change their pricing they're going to lose out big time.

Toeng.com
01-09-2003, 05:35 AM
Definitely nVidia :)
http://www.soltekusa.com/product/showproduct.php?productid=1041060567 looks good.

The Cross
01-09-2003, 05:50 AM
Personally i'd like to see ATI be the most advanced....Mainly cause they have a tendancy to cap their prices lower


I don't know too much about their new FireGL series, Just heard that it puts out and outstanding performance for 3D software in particular.

CgFX
01-09-2003, 06:18 AM
Originally posted by The Cross
Personally i'd like to see ATI be the most advanced....Mainly cause they have a tendancy to cap their prices lower

Slower requires lower. :-)

I don't know too much about their new FireGL series, Just heard that it puts out and outstanding performance for 3D software in particular.
The best measure of true 3D software performance "in particular" is SPECapc. If the new FireGL's are outstanding than the Quadro's must be spec-freakin-tacular as almost all the FireGL and Wildcat numbers have been pulled to avoid the obvious. (http://www.spec.org/gpc/apc.static/apcfaq.htm)

sedric
01-09-2003, 07:42 AM
i dunno, but just about every benckmark i read for any price range shows ATI ahead...the 9700 smokes the ti4600, (the 9500 pro is about the same as the ti4200 also) and in games with anti aliasing on, it isnt even close!

http://firingsquad.gamers.com/hardware/ati_radeon_9500_pro_overclock/page9.asp

there is even talk of the GEforce FX not really being up to the radeon 9700 standard at high resolutions with AA on because of ati's superior memory bandwidth(just rumors tho)

i think ati is the leader overall in the "game" performance...i cant really comment on which is better in 3d apps .

just my 2 cents....

sedric

loop29
01-09-2003, 09:13 AM
Good news from the softmodding section. Unwinder, the programmer of Rivatuner and Softquadro, is currently working on scripts for the Radeon 9x00 series of cards. Apparently scripts for turning your Radeon 9500 128 MB to a 9700 128 MB are coming. Rumours confirm that he is working on scripts for turning Radeon cards into the FireGL line of cards. As discussed before it seems that ATI is going the same direction with workstation cards and similarity to their gaming counterparts. That a Radeon 9500 can be turned into a Radeon 9700 was shown by hardware modification, the difference is enabling 256-bit wide bus on the 9500 cards which can be done while the memory layout is the same than Radeon 9700 cards. I heard that performance will be doubled in directx stuff like 3dmark.

Regards

Solesurvivor
01-09-2003, 12:47 PM
The best measure of true 3D software performance "in particular" is SPECapc

the best benchmark IMHO is real life application, like max, maya and so on..

but if you really want spec benchmarks, look at this one
benchies (http://www6.tomshardware.com/graphic/20021231/index.html)

nvidia isn't looking too good here , haha, good for them :)

CgFX
01-09-2003, 04:06 PM
Originally posted by Solesurvivor
the best benchmark IMHO is real life application, like max, maya and so on..

Agreed. And I think what you will find is that the Quadro's beat the new FireGL's pretty handily with customer specific benchmarks.


but if you really want spec benchmarks, look at this one
benchies (http://www6.tomshardware.com/graphic/20021231/index.html)

This is just Tom's Hardware quoting SPECviewperf. This has already been hashed out a bit on CGtalk. Viewperf (largely thanks to ATI) has become increasing irelevant thanks to all the "optimizations." i.e. cheating. If you look at the screen shots and evaluate what the cards are doing, it is clear that ATI is being very creative with their viewperf performance.

Proof is in the fact that the FireGL X1 has not carried that same relative performance over to SPECapc and gets thumped on applications that they show extraordinary leaps in performance over in viewperf. Unigraphics is the most obvious one.

Someday soon SPEC will need to do something to iliminate frame consistancy checks, image quality issues (AA lines being wrong), etc. or Viewperf will simply become unimportant when comparing performance.

GregHess
01-09-2003, 04:11 PM
Speaking of benchmarking....

The systems have started to arrive at cbow headquarters. Testing will begin quite shortly.

Solesurvivor
01-09-2003, 05:10 PM
Ati is simply gaining in on Nvidia, maybe they can skip another cycle in Workstation cards this time, and produce an r350 Fire gl sooner the quadro FX... Anyway, i'm smelling Nvidia is making the same errors as 3Dfx, they thought they were allmighty...

CgFX
01-09-2003, 05:24 PM
Originally posted by Solesurvivor
Ati is simply gaining in on Nvidia
It is more that NVIDIA is slipping back to ATI. Don't forget that R200 and R250 shipped before nvidia. On one occasion nvidia released driver enhancements to spoil their launch and on the other the new nvidia product came soon enough and was enough faster that ATI's cycle lead didn't matter. The ArtX team was a good group to get though.
maybe they can skip another cycle in Workstation cards this time, and produce an r350 Fire gl sooner the quadro FX...
This is not going to happen. In fact, I think FireGL X1 is only now just starting to have any real availability and I expect Quadro FX (or whatever) to be availabile shortly.

Anyway, i'm smelling Nvidia is making the same errors as 3Dfx, they thought they were allmighty...
Which errors are those? I didn't hear of any 3dfx managment that went into nvidia and that was the main issue at 3dfx, not engineering talent. Oh, do you mean arrogance? Maybe. But don't forget it wasn't too long ago that ATI was the big man and NVIDIA was the small one. Now they are roughly the same size and although NVIDIA owns the desktop and workstation, ATI still owns mobile. I think NVIDIA will continue to steal market share from ATI in mobile and that ATI may get some back in desktop and workstation. It all depends on if ATI ever gets serious about workstation and if NV30/NV35 don't have good availability.

Solesurvivor
01-09-2003, 09:58 PM
quadro FX won't be available the next 3 / 4 months me thinks, and in august / september, expect the new cycle to be coming ...

Interesting times indeed, especially for the wallet ;)

CgFX
01-10-2003, 01:02 AM
Originally posted by Solesurvivor
quadro FX won't be available the next 3 / 4 months me thinks, and in august / september, expect the new cycle to be coming ...

Interesting times indeed, especially for the wallet ;)
So if Quadro FX comes out in January or February as opposed to June or July then you would say that ATI is in trouble?

Why aren't they in trouble now with the 980 XGL being faster than a FireGL X1 on all application benchmarks I have seen (as well as every other price segment match up favoring nvidia)?

loop29
01-10-2003, 08:38 AM
Do you have any links or articles where the 980 XGL beats the FireGL X1 in application benchmarks? Cause these tests seem to draw another picture: http://www.amazoninternational.com/html/benchmarks/graphicCards/atiFireGLX1/atiFireGLX1.asp
And I know about the facts about tweaking drivers for SpecView, but even in the 3dmarks test the firegl looked pretty well, maybe new hope for the hybrid gamer/3D apps community. :rolleyes:

regards

Solesurvivor
01-10-2003, 09:29 AM
according to these tests shown by loop29, which I saw already but forgot I saw them , (typically me ;) ) nvidia doesn't look too good... Not forgetting that Ati can ,and must, improve a lot with driver tweaking, i'd say Ati will come on top of Nvidia, X1 - 980gxl ... Not talking about quadro FX here...

But each card has it's own advantages: X1 will be bloody fast when only wireframe is shown, 980GXl will perform better with textured scenes...

And i've seen some benches on games : the X1 doesn't lag too far behind the 980GXl, like 10-15% or so, which is thought initially.

CgFX
01-10-2003, 01:02 PM
Good Lord, these are just the Viewperf 7 benchmarks reprinted again.

How many times, in how many threads do we have to go over these and the fact that ATI is tuning to these benchmarks but has yet to give application performance?

FGL X1 does not have application performance yet. They get thumped in SPECapc and as a result all of the X1 results have been excluded from being posted. Don't believe me? Run the tests yourself!

Isn't it obvious that if Dell or Fujitsu-Siemens is running SPECviewperf and submitting those results that they are also running SPECapc to possibly submit those? It speaks volumes that their are zero X1 SPECapc submissions. It is because they loose and SPECapc is application specific performance of the workstation in question.

I have worked with a couple internal company specific benchmarks and SPECapc and in all these the X1 is currently dragging behind. Sure, there will come a day when they make progress on this but not before Quadro5 comes out. Then you will have ATI's good hardware with their sub-par software going against NVIDIA's better hardware with great software.

Did you even read the article?? The entire eight paragraph conclusion (which I won't repost here) slams ATI and the X1 for their behavior, questionable "optimizations" with Viewperf, and sub-par application performance.

elvis
01-10-2003, 11:31 PM
i'm really starting to be put off the idea of using X1's in our upcoming workstation renewals in march. it's a shame i can't borrow a couple of machine from dell to test before i make the purchase.

Solesurvivor
01-11-2003, 12:06 AM
yes, i did read the article, and if you are some aware of vid cards performance, you should know there's a problem with the placer chipset from intel concerning agp, so that could be one of the reasons why i didn't perform that good on the dual xeons...

Do we know this is a 'final' product here, cuz they are speaking of exclusive review, never been put online before, benches of 9500 sample weren't real good either at first, because it was just a test sample

As for their behaviour, the only thing I could come up with is the fact that this was an test sample and in that case, it somewhat justifies the behaviour, if this wasn't a test sample, shame on Ati...

Solesurvivor
01-11-2003, 12:11 AM
as I thougt, this is a pre-production sample, the same that Tom's Hardware has gotten, i quote 'Even this pre-production sample with its pre-release driver made its rival's flagships look old by comparison, allowing it to come away with first place. '

this puts all of your arguments aside, until you can test or show a benchmark, on the final sample of the X1.

Solesurvivor
01-11-2003, 12:18 AM
and here you have your o so honourable apc test with an E1 card in it, that's based on a 9500 radeon core which is a lot slower than the X1, and as you can see for yourself, the E1 isn't lagging to far behind the 980gxl. In fact, one time it's even ahead of the 980gxl, haha, what a surprise, in wireframe mode


apc bench (http://www.specbench.org/gpc/apc.data/apc_proe2001_summary.html)

CgFX
01-11-2003, 07:27 AM
Originally posted by Solesurvivor
yes, i did read the article, and if you are some aware of vid cards performance, you should know there's a problem with the placer chipset from intel concerning agp

Uh, the problem isn't intel, the problem is R300's broken AGP 8x. This has been shown over and over with many different chipsets and at various AGP plug-fest sessions. Even ATI's website speaks to the issues they are having with AGP 8x compatibility and stability.

BTW, all of my X1 benchmark info was based on production boards provided to us or bought in the retail channel. Not pre-production.

CgFX
01-11-2003, 07:31 AM
Originally posted by Solesurvivor
and here you have your o so honourable apc test with an E1 card in it, that's based on a 9500 radeon core which is a lot slower than the X1, and as you can see for yourself, the E1 isn't lagging to far behind the 980gxl. In fact, one time it's even ahead of the 980gxl, haha, what a surprise, in wireframe mode


apc bench (http://www.specbench.org/gpc/apc.data/apc_proe2001_summary.html)

The R300 (E1/X1) has a very fast AA lines (wireframe) engine. I sure hope that is all you are doing and that application stability and AGP 8x are not important to you. If so, enjoy your great new wireframe card... :-)... in between lockups in Maya. :-)

I hope you don't think that someone ran the E1 SPECapc's but nobody has just gotten the chance to run the X1. The fact is, they are about the same (again, run them for yourself) and the E1 is a lot cheaper.

Solesurvivor
01-11-2003, 08:29 AM
i'm not the kind of person who has enough money to buy an X1, hell, i can't even find a reseller here in belgium :annoyed:

again, show me your benchmarks and I will believe you, because you are nothing but an nvidiaboy to me, (hence your name).

Believe me, I have nvidia myself, and they were playing the same game as ati concerning 'improved perforance' in some benchmarks. But then I'm talking about gamer cards, not the pro's. Benchmarks like 3dmark2001 were heavily in favour of Nvidia. So Ati is now doing the same thing with SpecviewPerf. Who's the bad guy? It's all marketing baby

CgFX
01-11-2003, 07:08 PM
Originally posted by Solesurvivor
again, show me your benchmarks and I will believe you, because you are nothing but an nvidiaboy to me, (hence your name).

Sorry, not going to happen. Funny how companies can be about that. The beauty of SPEC is that you can download and run them yourself. The X1's SPECapc numbers are just about the same as the posted E1's (well, the partially posted E1's) so you can see the information you are asking for w.r.t. SPECapc.

Believe me, I have nvidia myself, and they were playing the same game as ati concerning 'improved perforance' in some benchmarks. But then I'm talking about gamer cards, not the pro's. Benchmarks like 3dmark2001 were heavily in favour of Nvidia. So Ati is now doing the same thing with SpecviewPerf. Who's the bad guy? It's all marketing baby
I have no doubt that NVIDIA probably tunes. That is not the point I was trying to make. Viewperf 7 has been offered up by many (including you) as prove that the FGL X1 is the fastest workstation card on the market. This is not currently the case. Frankly, it has the potential to be (until Quadro5 comes out) but ATI's software mediocrity is holding it back. I felt it was important to point out that in application performance (app benchmariks and customer specific benchmarks) the 980 XGL is noticeably faster than the FGL X1. A year old solution beating a solution that is just now starting to show up. Not real great news for ATI and when Quadro5 or Quadro FX ships, it will be tough for X1.

Take care,

nVidiaBoy :-)

elvis
01-11-2003, 09:11 PM
Originally posted by Solesurvivor
as I thougt, this is a pre-production sample, the same that Tom's Hardware has gotten, i quote 'Even this pre-production sample with its pre-release driver made its rival's flagships look old by comparison, allowing it to come away with first place. '

this puts all of your arguments aside, until you can test or show a benchmark, on the final sample of the X1.

i don't want to pick a fight here, but anything on tom's hardware guide is not what i'd call conclusive in any way shape or form. he's lost all cred with me, and most others in these forums a long time ago.

CgFX
01-11-2003, 11:40 PM
Originally posted by elvis
i don't want to pick a fight here, but anything on tom's hardware guide is not what i'd call conclusive in any way shape or form. he's lost all cred with me, and most others in these forums a long time ago.

Amen to that. Tom and Omid have only one agenda, to pick fights with every vendor (ATI and NV included) in order to create a sensationalized ruckus in order to get page hit$.

They have nowhere near the review integrity of someone like Amand at amandtech.

dvornik
01-12-2003, 12:57 AM
I can't believe anyone can say Nvidia's drivers are better that ATI. You guys argue about a couple of percent difference that a certain chipset or RAM type would cause. And somehow completely disregard the ability of Nvidia pro cards to support 3ds max5 - could be a 70% difference in benchmarks.

Their benchmarks are much worse than anyone's except Matrox. Maxtreme (that officially supports max5) doesn't work. I like SpecAPC for max, but you have to realize you can't test it with Maxtreme. As far as 3ds max 5 goes Nvidia completely failed and there's no reason to expect them to improve in the future.

Buying a professional Nvidia card for 3ds max is a complete waste of money. Might as well buy a 4200.

Xilica
01-12-2003, 12:59 AM
nVidia hands down

dvornik
01-12-2003, 01:02 AM
I'll repost the NO-Maxtreme graph. Nvidia uses OGL, other cards - "special drivers".

http://home.earthlink.net/~dvornik/cgt/tech/quadrrsux.png

And while at it I'll repost my own Quadro4 750 SpecAPC max5 results, with and without maxtreme:

Graphic mean 3.17 (was 6.56 with Maxtreme)
CPU mean 2.53 (was 4.62)
Overall geometric 2.93 (was 5.8)

And Greg, if you benchmark the cards please take note of the newest maxtreme problems:
http://www.cgtalk.com/showthread.php?s=&threadid=33610

GregHess
01-12-2003, 01:34 AM
Aye Dvornik, I will.

elvis
01-12-2003, 05:57 AM
Originally posted by dvornik
I can't believe anyone can say Nvidia's drivers are better that ATI.
our 3d guys are currently running quadro4900xgl's in 3dsmax5 with the latest maxtreme drivers, and have yet to comment on any instability or graphical problems. overall viewport rendering performance is much better than bog standard openGL. the existing issues we had with older maxtreme drivers in both max 4.2 and 5 are now gone.

on the whole, nvidia drivers are much more stable than ati drivers for their high-end cards. sit in the rage3d forums and have a look how many people are getting strange results with catalyst drivers.

john carmack (id software and openGL guru) himself calls nvidia's drivers his "golden sample" for accuracy and correctness in their code. he's even said if his software gives funny results with ati, he'll ring ati and tell them to fix their drivers. if his software gives funny results with nvidia, he'll fix his software. :)

dvornik
01-12-2003, 06:45 AM
Well, any tips on how your 3d guys made Maxtreme work? It doesn't work on any machines here. Boxx's dual 2200+ with 750 XGL (several configurations, including factory), custom single 1900+ with 750 XGL, various Dells with Quadro and Quadro2 Pro.

Performance is better in FPS with maxtreme, if you can use it . I wasn't able to use it, period.

elvis
01-12-2003, 07:50 AM
without being able to sit down at your machines, i can only offer the standard advice i always do: clean format and reinstall with latest service packs, drivers and software.

make sure all your hardware is running the latest firmware and bios flashes.

we've got dell precision 530's, dual 2.2GHz Xeon's, 1.5GB PC800 ECC RDRAM, windows 2000 pro SP3.

the latest intel inf drivers and application accellerators (from memory builds 1012 and 2.3 respectively).

quadro4 900xgl with nvidia 41.09 detonators (from nvidia's website), and maxtreme 4.00.26 drivers on 3d s max 4.26 and 5.

on the same machine is a variety of software including half a dozen adobe packages (photoshop, indesign, premiere, etc), ms office 2000 suite, mcafee anti-virus, autodesk ADT 3.3, bentley microstation V8, etc, and still rock solid stability.

scenes are typically quite complex architectural scenes of entire city blocks, sporting stadia, freeways etc, and well over several hundred thousand polys (sometimes into the millions).

maxtreme 4.00.25 had a few problems in max 4.2 with things randomly missing onscreen, so we ditched them. i didn't even bother trying the 4.00.26 release until i got back from christmas holidays, and the last 3 or so weeks using them has proven most successful. we've also migrated to max 5 in that time, but there are old projects still running max 4.26 for client and legacy reasons. all is well and happy on the viewport front. definitely just as stable as openGL, but now a lot faster and prettier.

loop29
01-12-2003, 03:31 PM
No offense dvornik, but what is the odd area behind the quadro results. The bars of Wildcat and ATI´s going from light red to dark red in continuity while the quadro bars does not. I didn´t imply that you´re cheating in some ways here, but it could be misconceived :hmm:


regards

CgFX
01-12-2003, 08:13 PM
Originally posted by loop29
No offense dvornik, but what is the odd area behind the quadro results. The bars of Wildcat and ATI´s going from light red to dark red in continuity while the quadro bars does not. I didn´t imply that you´re cheating in some ways here, but it could be misconceived :hmm:

regards

It was a previous chart that he modified.

here ya go.... :-)

Interesting that a 2 year old Quadro DCC is faster than the ATI and 3dlabs boards.

dvornik
01-12-2003, 08:18 PM
The original test from digit-life (http://www.digit-life.com/articles2/profcards/profcards-11-2002-3dsmax5.html) was ran with Maxtreme. I substituted the Maxtreme results with their OGL results from the same article. You can run benchmarks with Maxtreme and they are fast but you can't use Maxtreme in production due to a number of problems.

I think on architectural scenes maxtreme problems are less significant since most problems are on subobject level, such as vertices not being displayed. Some problems would still be noticeable - wrong color on spline objects (that are used for helpers in animation) in certain viewports. On our Boxx machines (I tried 3 different fresh installs with different drivers) Max 5 with maxtreme is unable to open any of the sample files - it crashes on file open. I'm rebuilding images for all our labs before the next semester so I'm dealing with these issues on a daily basis now - I was unable to put Maxtreme on any of our machines.


Here's (http://www.cgtalk.com/showthread.php?s=&threadid=32081&highlight=maxtreme) my original post on the subject.

Quadro DCC was originally 44.4. My point is WITHOUT Maxtreme you are getting an overpriced Geforce, not a pro card.

CgFX
01-12-2003, 08:29 PM
Originally posted by dvornik
The original test from digit-life (http://www.digit-life.com/articles2/profcards/profcards-11-2002-3dsmax5.html) was ran with Maxtreme. I substituted the Maxtreme results with their OGL results from the same article. You can run benchmarks with Maxtreme and they are fast but you can't use Maxtreme in production due to a number of problems.


Have you contacted NVIDIA (or PNY, Dell, HP, IBM, etc.) about these issues? Have you confirmed that there are valid bugs here?

It would seem that some people are using max 5 with maxtreme and NVIDIA just fine. I watched the max 5 launch and NVIDIA was Discreet's launch partner for the software. I also went to one of the launch road shows and nvidia was the demo platform used by the Discreet artist (she was very pumped about the Cg plug-in).

I just wonder if you should make sure that this is not unique to your environment or something that is known and fixed before you run around jamming hot pokers into people and posting _really_ misleading graphs extracted from someone else's article.

dvornik
01-12-2003, 08:34 PM
I've been dealing with this ever since max 5 came out. Nvidia doesn't have tech support in case you haven't noticed. My emails to Nvidia were ignored. Until couple of weeks ago Maxtreme did not officially support max 5 so I was told by Nvidia fans to wait for the next release. 4.00.26 is certified for max 5. It's still too buggy for most people to use.

There's nothing misleading about my graph. It clearly states that Nvidia cards are using OGL.

Here's my thread on new maxtreme if you think I'm the only guy who has problems with it:
http://www.cgtalk.com/showthread.php?s=&threadid=33610

Serge Cashman
PC Support
Pratt Manhattan

CgFX
01-12-2003, 08:43 PM
Originally posted by dvornik
I've been dealing with this ever since max 5 came out. Nvidia doesn't have tech support in case you haven't noticed. My emails to Nvidia were ignored.

Actually, NVIDIA does have a very extensive developer and customer support program. For smaller outfits most of the issues get passed through PNY or the workstation vendor you are using. Either way the bugs still get submitted and tracked and resolved. For larger customers there is direct support available and I have seen first hand how impressive the response of that group can be.

Who did you send email to at NVIDIA?

Are you using PNY boards or boards from a workstation vendor? Have you submitted a support call or email request to the company that provided your boards?

One of the main benefits to Quadro in my book is application certification and support. It is there and I have seen how effective it can be if used and used properly.

loop29
01-12-2003, 08:55 PM
What kinda benchmarks they used there for the results, I only know the specapc benchmark suite for 3dsmax4.2? I will have a look how it is behaving on my system if I get those demo´s.

regards

dvornik
01-12-2003, 09:17 PM
I think they used maxbench. It's the scenes on the second CD in the Benchmarks folder, at least they are in max 5. It's the test discreet uses for benchmarks on it's site. digit-life (I admit they are not the most accurate reviewers out there) are using the older version from 3 or 4 I believe.

I posted my SpecAPC results in a previous post and as you can see OGL is about twice lower than Maxtreme.

CgFX I swear the fact that Maxtreme doesn't work is not my fault.

Solesurvivor
01-12-2003, 09:22 PM
my contribution to all this is, I used the Maxtreme drivers as well, and hell, even Direct3d works better than that, if this specialized driver support from nvidia, i'd take anything from ATI.. Okay, almost everything :)

CgFX
01-12-2003, 09:26 PM
Originally posted by dvornik
CgFX I swear the fact that Maxtreme doesn't work is not my fault.

I believe that but what about my questions?

Who did you send email to at NVIDIA?

Who did you submit a support call or request on this issue with? PNY if they are PNY boards? Dell, HP, or IBM if they are name brand workstations?

If you did submit a support call, did they give you a bug ID? Are you able to track the status of a fix if you have indeed found one or more existing bugs?

dvornik
01-12-2003, 10:09 PM
I don't know anybody personally at Nvidia. Nvidia doesn't have support contact information and doesn't support their products directly. I was only able to find generic email addresses such as info@. Any specific contact information would be appreciated.

No I haven't called Dell or Boxx support about this issue, since until recently Maxtreme didn't officially support max 5. Bug ID? From Dell? I haven't dealt with Boxx support yet, maybe it makes sense to call them.

Max 5 was out since July, I believe Nvidia had enough time to write functional drivers for it. Maxtreme is mentioned in their promotional materials for Quadro and is used for their benchmarks.

elvis
01-12-2003, 10:21 PM
i've said it before and i'll say it again: maxtreme 4.00.26 works great in max5, no stability problems, no visual problems. these drivers are more than "functional".

i've posted my recommended install methods in a few threads, so i won't go over it again. the point is there's a difference between a company with driver problems and not being able to get the drivers working yourself.

you'd probably be wise to sit down and really nut this problem out for the benefit of yourself and the people you support. if you can't get it working, then you'd be wise to ring everyone and anyone you can that has anything to do with the system (dell, boxx, nvidia, pny, etc). no-one will do anything to fix it if they don't know there are issues. you'd be surprised what a bit of bitching from customers can achieve! :)

dvornik
01-12-2003, 10:48 PM
Once again, I'm not the only person that can't make it work properly. Thanks for clarifying to me how to install Windows.

The latest Maxtreme on Dell's site is 4.00.21. I'm sure they will be able to offer some meaningful support, like they always do. Boxx has 4.00.22 that they call 6.02 for some reason.

I believe if it has problems on several radically different machines (fresh install of win2k SP3) with several drivers it's safe to assume that there might be other issues here besides my computer illiteracy.

David Watters
01-13-2003, 12:29 AM
Hello all.

Originally posted by dvornik
I don't know anybody personally at Nvidia. Nvidia doesn't have support contact information and doesn't support their products directly.

Actually, we can provide direct support for our workstation products in certain situations.

I was only able to find generic email addresses such as info@. Any specific contact information would be appreciated.

The volume of email traffic that goes into generic addresses like that is pretty high. Sometimes you can get lucky but the best way to get support is to go through the proper channels which depends on where you bought your Quadro boards.

I would be happy to help with this if you want to contact me directly at dwatters@nvidia.com.

No I haven't called Dell or Boxx support about this issue, since until recently Maxtreme didn't officially support max 5. Bug ID? From Dell? I haven't dealt with Boxx support yet, maybe it makes sense to call them.
If you do have Boxx systems they are great about providing support ang getting bug related information into NVIDIA and then tracking the progress. I strongly recommend placing a support call with them if you are having troubles and they will be very quick to get you taken care of.

Dell is the same way but as you suggest below they may lag behind a bit on driver revisions based on their volumes and rigid qualification process.

Maxtreme is mentioned in their promotional materials for Quadro and is used for their benchmarks.
I am checking to see if you are running into know issues or not.

dvornik
01-13-2003, 12:38 AM
Thanks, David. I'll email you the detailed description of the Maxtreme issues we have been unable to resolve. I plan on consulting with Boxx support as well.

It's very encouraging to see Nvidia on this board.

Serge Cashman,
PC Support
Pratt Manhattan
scashman@rand.pratt.edu

CGTalk Moderation
01-14-2006, 02:00 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.