Whatever wumpus, your arse is just trolling. That demo runs perfectly fine on my machine AND on 3 of my friends machines. ALL of us have 9700 Pros. It also states in the readme that all the Radeon 8000/9000 series are supported. Check your facts before you talk trash.

As a matter of fact, that same readme suggest that NV card users install the 41.09 Dets (nvidia's latest) and ATi users have at lease the CATALYST 2.5s (which are close to 3 months old). What does THAT say about drivers....

I'll repeat this verbatim as it answers the same question you made in the other thread:

About the 50/50 profit split that you mentioned in the other thread, WaltC, do you happen to have a link to that conference? The reason I ask is because Beyond3D mentions that ATI said that the high-end only makes up 1% of the overall market.

"ATI point out that the 'Ultra High End' market, which ATI categorise by a suggested retail price of above $350, accounts for about 1% of the overall market. In comparison the 'Performance Mainstream' segment, categorised by a pricing of about $200, represents 8 times the number of sales and the sub $100 mainstream sector being 32 times the volume"

If you are willing to do some back tracking in the Rage3D archives, to hunt for the conference ATI held in which the R350 Product was *first* officially mentioned last year (the first time ATI actually admitted to the product) you'll find the quote. I remember it clearly and it's stuck with me ever since.

You can also look up developer quotes at ATI's mojo last year in which it was stated by an ATI developer during the conference that in the first 60 days ATI shipped over one million 9700P's--the term used was 9700P, not R300, so I think they were talking about finished card sales instead of VPUs.

The first problem I see is that what ATI said that I've quoted was a remark about profit split--not about unit shipments. The second problem is that all "$350 +" products are not alike, and as such the number you sell of them will differ wildly depending on demand. It was well publicized last year that ATI had to go hunting additional FAB capacity to meet demand for the 9700P. So obviously, you can have a mediocre to poor $350 product or you can have the performance-leading $350 product and differences in demand and sales between each will be dramatic. As will differences in profit. The ATI quote I reference here was made several months after R300s, specifically 9700P's ,started shipping, so it could well be the B3d info is based on much older information as it pertained to ATI's pre-9700P sales (which would certainly make sense to me.) I don't want to say that the information you references is flat wrong--just that it doesn't apparently reflect the state at ATI post R300.

[qo price cuts on the 9700'/q]
http://www.anandtech.com/guides/showdoc.html?i=1793&p=4

I just wonder if there is any allocation problem with ATI, because newegg has no 9700 in stock for sometime

wumpus - Funny, the full version runs great here.

Originally Posted by don27
Yes it sucks. It also means no price cuts on the 9700\'s, LOL!

2 0f 7

It's true only 1% make up the high-end market. However, the leader in speed performance usally gets the larger portion of the pie as it's lower end cards get bought from the consumer because of the hype surrounding that companies high end product

About the 50/50 profit split, WaltC, do you happen to have a link to that conference? The reason I ask is because Beyond3D mentions that ATI said that the high-end only makes up 1% of the overall market.

"ATI point out that the 'Ultra High End' market, which ATI categorise by a suggested retail price of above $350, accounts for about 1% of the overall market. In comparison the 'Performance Mainstream' segment, categorised by a pricing of about $200, represents 8 times the number of sales and the sub $100 mainstream sector being 32 times the volume"

That may be all well and true, AG #131, but the fact of the matter is, despite what the PR department said, everyone expected the FX to cream the R9700 based on what has happened before. Take the 8500 for example. It was clearly superior to the GF3, but nVidia fixed that right up with a driver release and ATI was playing second fiddle again. That's what everyone expected this time around, but it didn't happen. So get off your nVidia fanboy kick. Everyone, not just the fanboys, thought history would repeat itself.

Namarrgon, you were overreacting.
wumpus is bitching.
Nvidia should be punished for lying.

Driver issues are a dead one. My 9700pro is running all games current, past and demos fine. And i agree with JustAnEngineer regarding how much more gam designers are working with ATI's hardware. This seems to be the only area Nvidia freaks can rebuke the success of the 9700pro. (with great failure...)
Hey playing Unreal2003 at 1600x1200 with fsaa+ansioscopice flitering on is soooo sweet. And i don't ned to have a 350 watt power supply or a extra pci slot or a pair of head phones either......

This whole thing is a joke. It was Nvidia's P.R. department that claimed they had the 9700pro KILLER. It was them! And all i see in this forum are Nvidia fan-boys saying..."you have to realize that both cards are from the same generation, the fx being late..."
If that's the case and we know the 5800 was late and still is, and will be. Being in the same generation, why the frig would Nvidia say they had the 9700pro KILLER then. They suckered you fan-boys into believing the hype. Hype with power. Hype with releaed dates. Hype with quanities. Hype with availabilities. And now it looks like the 5800 Ultra won't see the light of retail days.....
Remember, it was Nvidia who touted the 5800 Ultra as the 9700PRO killer.

never had a game NV tech couldnt run... how many ATi users can say that?... oh... sucks to be you...

Normally I ignore stuff like this but it's just too tempting...*snicker*

Me, me! I can say that. Since I replaced my GF4 Ti4600 with a 9700P last September, have had exactly the same number of games refuse to run that refused to run on the Ti4600....0. (None, nada, zippo.)

But, oh my, how much better the image quality is, and the performance far better, too.....*chuckle* I'm having a good time being me right now...;)

I just ordered a PowerColor Evil Commando 2 9700 Pro DDR board, I was always about NVidia until they pulled this FX bs.. ATi was never a company I trusted in the past, but times change, and I can afford the 9700 Pro. :p

NVidia's drivers aren't that great, either. In their insane quest to improve frame rate at the expense of everything else, they have released several drivers that were terrible. The detonator 28.32 drivers were especially bad, causing system crashes for thousands of users.

Most games work right out of the box on GeForce4Ti because that's the card that many developers used in their development machines, [intil last August[/i]. Since NVidia has had nothing to compete with Radeon 9700 Pro for six months, many game developers are working with the ATI hardware more than they have in the past. Problems are found and fixed before the game hits the shelves.

never had a game NV tech couldnt run... how many ATi users can say that?... oh... sucks to be you...

Originally Posted by wumpus
Hey, the brand new FilePlanet-exclusive Freelancer demo doesn\'t work on ATI 9700 cards. Check the readme. It says \"turn down hardware acceleration one notch\". Grrrrreat. I\'m already reading forum posts complaining about this.

Sigh-- the more things change, the more they stay the same. ATI [bs clearly ahead in the hardware rac/b], but does it matter when the drivers jack you at every available opportunity? Sure, the classic \"top 20\" games work, but opening a new game and installing it is a nail-biting game of \"will my new game run on the 9700?\" Been there, done that, sold the card on eBay.

C'mon. You guys just don't want to face the cruel truth.
Open up your eyes. It's too obvious.
If they don't do a radical redesigning on NV35, nVIDIA is really screwed up on this cycle.(or on more cycles)

MX200

Whoops, hits Post Anon.

Well, if it's true that RV350, which has already taped out, was manufactured on .13, then I'd say that ATI had very little problems, particularly since TSMC's .13 micron yields are hovering around 70%-80%.

[quote]While high-end gaming cards are good for bragging rights, the real money is made on the high-volume cards for the mass market. There has been surprisingly little said about NVidia's upcoming products for the casual gamer.[/quote]Walt beat me to ATi's 50/50 profit split. But he didn't mention that it seems nV's mainstream hardware (NV31/34) will be less capable than the NV30, and one of them may not even be DX9-compliant. The pressure is certainly on nVidia now.

Engineer, I'm not sure how you'd expect to a) find an FX Ultra for less than a six-month old card (which is produced with cheaper memory, cheaper boards, and on a higher-yield process), and b) live with a 70dBA fan.

The real fight will be R350 vs. NV35. Actually, it may be R400 vs. NV35, depending on how fast nVidia gets back on its feet.

The people expecting miracles from Nvidia's driver team because the GFFX is late are being overly optimistic. It's precisely BECAUSE it's late that we won't see such gains. Since the problems were with production rather than engineering they almost certainly had feature complete silicon for months, and thus plenty of time to work on drivers.

That was an interesting thread at B3D. Thanks for the reply, Walt.

AG #94/96/99/104,

I think you need to calm down a bit. In your frenzy to accuse nVidia of lying, you're making wild, unsubstantiated statements yourself. Please post some URLs to back up your "facts", rather than accusing everyone else of ignorance.

Personally, I think that if you put aside the 6 month delay on the NV30 for a moment and pretend that the R300 and NV30 were released at the same time, people would find the chips actually very similar, for the most part. NV30 is undeniably more flexible, but this isn't relevant to most people right now (it is to my industry, however). Both are clearly very fast - R300 is able to compete comfortably with NV30, despite the latter's 50% clock advantage; there's no clear winner here yet. Really, the biggest problem with NV30 is that it requires the dustbuster to get that clockspeed, and thus to compete.

However, the reality is NV30 *was* delayed for 6 months by the decision to go with a 0.13 um design. This is directly responsible for the other NV30 problems.

First, the driver disparity. ATi have had 6 months to optimise their drivers, and this has doubtless helped. No-one can deny that both nVidia and ATi have historically produced enormous speed increases with later drivers, and this is never more true than with new architectures. nVidia are at a real disadvantage here, for a change. Has anyone compared the known GFFX results to early 9700 PRO results? I suspect the picture would be more in nVidia's favour.

More importantly, I'm looking forward to seeing how the FX's performance compares to the 9700 PRO once their driver team has had the time to work their magic. If they can get the sort of 25+% gains we've seen in the past, it'll put the NV30 well ahead of R300, unquestionably, maybe even enough to allow the non-Ultra, non-dustbuster GeForce FX 5800 to compete with R300, or the Ultra to compete with R350. Time will tell.

Second, the gamer backlash. People are clearly disappointed with the NV30 because it wasn't as good as they expected - that is, they expected it to be a lot *faster* than R300, thanks to both nVidia hype and the extra 6 months of waiting. Instead of considering it to be an equivalent chip that was simply delayed, they expected more from it, and are seizing on everything they can as evidence of nVidia's screwups.

This is unsurprising. For those with memories longer than a few months, the exact same thing happened for ATi's past two flagship releases. The Radeon looked clearly superior on paper, but was 6 months later than the GF2 and was let down by immature drivers. Everyone and their dog took their turn at saying how hopeless ATi were; it was not a good time to be an ATi fan. A year later, the Radeon 8500 looked even better on paper, with lots of innovative hardware features, but again nVidia's older, less fancy GF3 was able to pull past once again simply through more mature, optimised drivers, and the nVidia camp had another field day. In both cases ATi eventually delivered the promise of those chips, in speed and features, with mature & fully optimised drivers, but too late - no-one cared anymore, and many ATi users had already switched to nVidia.

It's just the same today, only now it's nVidia who have released the fancier chip 6 months later, but are struggling to compete. They managed to snatch a bare lead in many benchmarks, but they need a crazy-ass cooling system to get the clockspeed they need to make up for their immature drivers. I predict that in 6 months or less, the GeForce FX 5800 with optimised drivers will no longer need the dustbuster to compete with the 9700 PRO - and no-one will care, because the 9800 PRO will be out.

There's nothing new under the sun, except that now it's nVidia who are behind. The *only* relevant question is whether nVidia can take the lead once more, and that, to me, is entirely dependant on whether ATi have the same sort of trouble with their 0.13 um process that nVidia did.

Consumers don't need 32 bit textures.....I mean 8 rendering pipelines.

Here's but one of a couple of threads on the matter at B3D:

http://www.beyond3d.com/forum/viewtopic.php?t=4252&start=0

My mistake - the Beyond3D interview was not with David Kirk, but with senior product manager Geoff Ballew.

While high-end gaming cards are good for bragging rights, the real money is made on the high-volume cards for the mass market.

First, this is an oft-repeated, very common misconception. At a conference last year, in fact the very conference in which ATI first officially admitted the existence of R350, ATI was asked to define the company's revenue split between the "high end" (Radeon 9700-ish) and the "middle-low end" (everything else) and responded with "...It's about a 50-50 split in terms of profit." Many of the "high-volume" products are very cheap and several of them have to be sold for every one of the high-end cards to equal the profit. The idea that the high end means nothing and is something any of these companies can flippantly dismiss is at best an apology and at worst just plain false.

There's so much money in this market segment that you'll see even small companies like Matrox fighting to have a presence in it (not counting the $400,000,000 or so nVidia claims to have invested in nV30.) Then of course there are all the peripheral benefits that go along with the mindshare winning the "performance crown" garners a company--reverberations from it affect sales of the entire line, OEM contracts--you name it. Much, much more than "bragging rights" involved here....*chuckle*

There has been surprisingly little said about NVidia's upcoming products for the casual gamer. The sheer volume of cards sold in this niche drives what the game developers can do, too. That's why folks like Carmack spoke out derisively when GeForce4MX was introduced without the capabilities that GeForce3 had. While NVidia helped gaming tremendously with the shaders in GeForce3 and GeForce4Ti, they hurt the industry as a whole by introducing a "new" product that actually went backward a generation in capability.

Carmack's objection, I believe, had much more to do with the naming conventions nVidia was using. He wasn't too thrilled with nVidia calling a GF2 a GF4, as I recall, was about the extent of it. He said pretty pointedly it would have been fine had nVidia left off the "4" in the product designation (since the architecture didn't deserve it.)

I don't believe that we know enough yet to say that GeForceFX is a failure. After 10 months of hype, it still hasn't made it to the consumer. Let's see how it does in the marketplace. If GeForceFX 5800 Ultra were to show up in-stock at my favorite on-line vendor priced between Radeon 9700 and Radeon 9700 Pro, I would probably buy it.

Hmmm....very interesting sentiment. You don't believe we "know enough" about it to make a judgement on it, yet you say you "might buy it" if it showed up at the right price. I think you've got some rose-colored glasses on there, Mr. Engineer....;) Why would you even think of buying *anything* which you don't know enough about yet to evaluate (regardless of price)...? *chuckle*

What's being bandied about now is the very interesting premise that the nv30 is not even an 8x1 architecture card with a real, physical 8x1 architecture such as the R300's. Rather, that it's a 4x2 architecture with some limited 8-pixel-per-clock capability--which is diametrically different from what nVidia's been advertising and promoting about the architecture. Indeed, that's the subject of this thread, right? People currently testing the chip report that its behavior is that of a 4x2 rather than an 8x1 architecture. If true, I'd say that gives the disgruntled plenty to be disgruntled about...;) (And I don't just mean people like Fuad--but responsible people at sites like B3D are reaching the same conclusions--based on their own testing.) So you better hang on as the chip *might be* an even bigger failure than it was first assumed...It's too early to say for sure, but the evidence is certainly stacking in that direction at this time. Of course, the fact that nVidia isn't saying much about any of this, at least directly, isn't helping anyone to "know enough" about nv30, is it?

Myself, I see this as a shocking development, but am content to wait out the proof one way or the other...

[quote]Calling around to find out about using DDR II memory with a 256bit interface, we learned that it would be possible to get some - though in limited quantities because there is no demand for it and it would be extremely expensive.[/quote]
Once again Fuad shows us why he shouldn't be writing about graphics technology.

How many DDR I memory chips have you seen with 256+ pins, Fuad? RAM chips do not have a 128 bit interface or a 256 bit interface, they have 32 bit interfaces, and are paralleled to give 128 bits or 256 bits. There's no reason you can't make a 256 bit DDR II setup, it just requires 8 chips and more layout (and more pins on the gfx chip). nVidia didn't go with 256 bit DDR II because they didn't feel it was necessary, not because they couldn't find thumping great DDR II chips with hundreds of pins.

The Inq (and Fuad in particular) are often clever at ferreting out tidbits, but as soon as they start drawing their own conclusions from those tidbits they go all awry. Worse, they rarely make any distinction between reporting and speculation.

AFAIK, the only evidence we have that NV30 has only 4 pipelines (as opposed to being limited to rendering 4 colour+z pixels per clock, which isn't necessarily the same thing) is from a single sentance from the Inq that claims an unnamed nVidia marketing guy told them so. They don't even quote him, they paraphrase it.

OTOH, the Beyond3D interview with David Kirk (who ought to know) states quite clearly that "We have 8 pipelines and they can each apply one texture per clock".

Who do you believe?

I find it strange that folks are so intent on bashing NVidia for their disappointment with GeForceFX 5800 Ultra when comparing it to Radeon 9700 Pro while ATI still receives little praise for their accomplishment.

Why are people so bitter about GeForceFX?

Is it because they are upset with themselves for believing the NVidia hype machine in the past ten months? Is it because they really wanted NV30 to be something that it is not? Is it because they bought NVDA a year ago at over $65 per share (now at $12.86)? (ATI has done even worse.) For the really agitated gerbils posting here--What is your reason for being so upset?

While high-end gaming cards are good for bragging rights, the real money is made on the high-volume cards for the mass market. There has been surprisingly little said about NVidia's upcoming products for the casual gamer. The sheer volume of cards sold in this niche drives what the game developers can do, too. That's why folks like Carmack spoke out derisively when GeForce4MX was introduced without the capabilities that GeForce3 had. While NVidia helped gaming tremendously with the shaders in GeForce3 and GeForce4Ti, they hurt the industry as a whole by introducing a "new" product that actually went backward a generation in capability.

I don't believe that we know enough yet to say that GeForceFX is a failure. After 10 months of hype, it [btil/b] hasn't made it to the consumer. Let's see how it does in the marketplace. If GeForceFX 5800 Ultra were to show up in-stock at my favorite on-line vendor priced between Radeon 9700 and Radeon 9700 Pro, I would probably buy it.

I feel like Nvidia is on a verge of a huge back lash from developers to users to review sites. NVNews is an example where NVidia has clearly alienated, strung along, BSed and really treated the people who promoted the most their cards with utter comtempt. If Nvidia keeps this up they will make RamBus look like angels. My two cents worth.