ART of War, nV vs. ATi your thoughts

And with windows vista coming before cristmas they have 3 quarters (maybe even less) to slug it out with these cards until demand for them is sure to drop. Sure Nvidias dx10 design seems to be alot closer than ATIs which is what I would guess could buy them marketshare.
Is nvidia really interested in cutting prices to maybe sell some more cards (at less profit) at ATIs cost or will they want to maximize their profits and keep good margins? Or maybe ATI is willing to take a drop in ASP to gain/hold marketshare? Will this really have big effect on their marketshare? Speculations are that ATI has more experience at working with more unified architechtures and that their dx10 design will be superior to Nvidias, will this enable ATI to gain 60% ? Stay tuned to find out. :LOL:

Seriously though, my whole point is that they have half a year to press their advantage before Microsoft starts bombarding the world with Vista PR, how big difference do you think that it can possibly make to have what you percieve as an advantage? Maybe Nvidias yields are such that the actual costs of the cores are the same as ATIs competing product? You don't know, they wont say and we can only speculate. Neither is going away soon and with dx10 so close, future products will be much more important than this short period.

I must add at last that I think vista will be the slowest windows launch since 95a. But people are sheep anyway so they'll buy what's 'safe' for the future.

EDIT: Missed a word.
 
Last edited by a moderator:
maaoouud said:
And with windows vista coming before cristmas they have 3 quarters (maybe even less) to slug it out with these cards until demand for them is sure to drop. Sure Nvidias dx10 design seems to be alot closer than ATIs which is what I would guess could buy them marketshare.
Is nvidia really interested in cutting prices to maybe sell some more cards (at less profit) at ATIs cost or will they want to maximize their profits and keep good margins? Or maybe ATI is willing to take a drop in ASP to gain/hold marketshare? Will this really have big effect on their marketshare? Speculations are that ATI has more experience at working with more unified architechtures and that their dx10 design will be superior to Nvidias, will this enable ATI to gain 60% ? Stay tuned to find out. :LOL:

Seriously though, my whole point is that they have half a year to press their advantage before Microsoft starts bombarding the world with Vista PR, how big difference do you think that it can possibly make to have what you percieve as an advantage? Maybe Nvidias yields are such that the actual costs of the cores are the same as ATIs competing product? You don't know, they wont say and we can only speculate. Neither is going away soon and with dx10 so close, future products will be much more important than this short period.

I must add at last that I think vista will be the slowest windows launch since 95a. But people are sheep anyway so they'll buy what's 'safe' for the future.

EDIT: Missed a word.
Click to expand...

It will depend on the market conditions at the time when Vista comes out for ATi to gain marketshare at that time, for the time being there is no reason to think nV has bad yields, even if they did as you suggest to get to the same cost per chip would be nV has 40% to 45% chips that go to waste per wafer over ATi, thats alot, this would be for the 7900 and the r580. Then you have the the 7600 gt vs the x1800xt, nV will have to get something around 65% less chips per wafer. This is highly unlikely.

Now you mention dx10, nV has seen what ATi has both with xenos and the r580, its easier for them to speculate what to do to stay in competition. ATi has not seen what nV has this is a cause for guess some guess work on thier part.

Those yeild numbers are very unlikely again

http://www.digitimes.com/mobos/a20060316PR210.html
 
Last edited by a moderator:
_xxx_ said:
Not really the same class here, maybe 7600 vs. X1600?
Click to expand...

Opps ment x1800 gto :), pretty much the x1800 varients are up against the 7900 gt and 7600 gt yields. And when the rv570 comes out, it might be a breath of fresh air for ATi, but now it looks like it won't be on 80nm.
 
Last edited by a moderator:
Razor1 said:
Then you have the the 7600 gt vs the x1800xt, nV will have to get something around 65% less chips per wafer.
Click to expand...
As I stated earlier the x1800GTO is probably a temporary card that will cease to exist as soon as they have run out of old or defective cores. It's supposed to be replaced soon by RV560/570 anyway. Also it just occurred to me that this was a very quick reaction from them to counter the 7600 (and the older Nvidia cards still out there) while waiting for the new mainstrean/mainstream-performance solution. But you are quite right in arguing that Nvidia probably has a clear cost advantage here, it depends on how many poor dies ATI gets from the R580 batches - far from an optimal solution but it's still a quick counter.

Razor1 said:
Now you mention dx10, nV has seen what ATi has both with xenos and the r580, its easier for them to speculate what to do to stay in competition. ATi has not seen what nV has this is a cause for guess some guess work on thier part.
Click to expand...
Yes but Nvidias designs will be released sooner and then ATI would know their performance before they launch.
 
As I stated earlier the x1800GTO is probably a temporary card that will cease to exist as soon as they have run out of old or defective cores. It's supposed to be replaced soon by RV560/570 anyway. Also it just occurred to me that this was a very quick reaction from them to counter the 7600 (and the older Nvidia cards still out there) while waiting for the new mainstrean/mainstream-performance solution. But you are quite right in arguing that Nvidia probably has a clear cost advantage here, it depends on how many poor dies ATI gets from the R580 batches - far from an optimal solution but it's still a quick counter.
Click to expand...

I mentioned this aswell, but June/July is not soon enough ;) .

Yes but Nvidias designs will be released sooner and then ATI would know their performance before they launch.
Click to expand...

Yes that is true but this also puts ATi in a reactive position again, ATi will have to have a quick back up plan if nV in all likly hood will have a plan of thier own if thier chip is out gunned, and if ATi is late by 1 or 2 month as being speculated thus far, this won't be good since OEM's will be renegotating deals around this time, if I remember correctly June/July is the time this starts.
 
jb said:
Oh good greif, again having ATI screw up with the r520 delays was a lucky break for NV. Just like NV screwing up with the NV30 opened the door wide open for ATIs R300 parts. Neither was planned for be ceritanly helped the other IHV.
Click to expand...

To err is human, after all. But, I see very little "luck" involved. The failure of nV3x (I promised never to speak of it again as you have phrased it above, so I hope I am permitted to use the "x" nomenclature...;)), was a three-pronged failure:

(1) It was an architectural design failure to begin with (which is why nV came out swinging in 'late '02 and for all of '03 with its various and ill-received "Who needs DX9 when you've got DX8?" PR and marketing campaigns, which attempted to defend nV3x while castigating both Microsoft and R3x0.)

(2) It relied primarily on an unproven and unperfected manufacturing process (.13) that was hoped to provide nV3x with enough clockspeed and yield to be superior to all comers. Which leads to the third cardinal error surrounding nV3x...

(3) The conception widely held within nV at the time was that in the 3d gpu markets it had eliminated all high-end performance competition with the absorption of 3dfx.

As late as early August '02, while I wasn't of course guilty of either of the first two errors above, I was myself guilty of believing in #3 just as much as nV apparently did at the time. That all changed, of course, when I bought a 9700P in September of '02--just as I'm sure it changed for nV when nV got its hands on a few 9700P's at about the same time.

These three factors simply did not involve "luck," imo, but rather were representative of errors in judgment on all counts. Simply, they were mistakes, which all people and companies alike inevitably make from time to time. The only difference between individuals and corporations in this regard is how long it takes them to learn from their mistakes in order to stop making them. During the '02-'03 period, I felt it took nV an unusually long period of time to adjust its attitude in response to its own mistakes and the successes of its competitor. However, contrasted with the length of time it took Intel to finally come to understand the degree of competition that AMD represented, for instance, nV's adjustment period was very brief by comparison.

To compound these mistakes 100-fold, of course, was the fact that far from being dead and gone in the high-end 3d-gpu market arena, ATi had been at work for a couple of years on R300 prior to bringing it to market. More specifically, ATi had been hard at work reinventing itself almost from the ground up (with the acquisition of ArtX and its personnel, for instance), and this was not something that nV ever gave the appearance of recognizing as an important event. Indeed, what ATi was about during the post ArtX-acquisition, pre R300-shipping period was a very well-kept secret. At least, it certainly seems so in retrospect. But perhaps it was simply because few were paying attention to ATi at the time.

So quite apart from luck, we have nVidia resting on its laurels and counting its chickens long before they hatched, and we have ATi quietly working very hard to reinvent itself and to produce the significantly new approach to high-end 3d-gpu technology that came to be known as R300. This combination of both events and attitudes created the outcome, no luck involved on either side. And of course, the role that Microsoft played in its decisions as to DX9 was not a matter of luck, either, but rather one of conscious deliberation and decision, as was nV's decision to initially reject the path that DX9 was slated to follow.

In short, while I can think of quite a few adjectives I might use to describe the nV3x-R300 years, "luck" would not be among them...;)

As to what "happened" to ATi with R4x0, it boils down, imo, to a simple, and again, all too human, lapse in judgment: ATi overestimated the speed at which PCI-E would usurp AGP in the high-end 3d-enthusiast markets, and was too hasty in shifting the bulk of its manufacturing to PCI-E at the expense of AGP support. This was a mistake that nV did not make.

The interesting thing about all of this competition is that it reveals that judgment relative to when, where, why, and how to release a given technology into the market can be as fundamentally important to the success of a technology as the merits of the technology itself. Corporate judgment at the top is often crucial and decisive. I just don't see simple "luck" as playing much of a role in the very complex business of technology competition.
 
Last edited by a moderator:
I woke up violently, convulsing in pain for an instant. I gasped for air as I felt the sheets on my clammy skin. That horrible feeling in the pit of my stomach was back. I already knew what was going on. I didn't want to believe it. I watched the shadow of the ceiling fan in the late afternoon light as I lit a cigarrette. The smoke filled my lungs; it didn't make me feel better like it usually did. I stood up, and I walked to my desk. I had to make sure.

And there it was, the sight like a ghost from the past.

WaltC posted about the NV30.
 
Nice summation by WaltC there. I guess niehter company will be caught with their pants down again, at least in regards to the competition. Fumbles will still occur of course.
 
You must log in or register to reply here.

Similar threads

F
Dedicated gaming GPUs bottom of the priority list?
2
Replies
34
Views
3K
Seanspeed
S
iroboto
Square Enix Previous Director of Business Holdings at SE discusses the Games Industry
Replies
3
Views
749
eastmen
E
S
TSMC warns of excessive chip inventory
2 3 4
Replies
74
Views
12K
eastmen
E
D
Video Game Console Cooling - A Hot Topic (pun kind of intended).
Replies
14
Views
3K
Shifty Geezer
Shifty Geezer
S
Value of business buy-backs *spawn
Replies
11
Views
994
eastmen
E
Back
Top