Mercury Research on graphics market share

[Dave Baumann]

Here's my question. Why the heck is Mercury research even lumping midrange/low end cards into the high end category?

Actually, taking a look between the two Analyst reports the definition appears to be "Second Generation DirectX9 greater than 100M Transistors". Second gen DX9 is arbitary, but would force the count to exclude 9800's; then of course, 6200's are currently just NV43's which are over 100M transistors.

In otherwords, it was a rediculous count.

 

[Bouncing Zabaglione Bros.]

I think it's pretty hilarious that the usual suspects are criticizing people who are questioning NV's ethics, especially given that two major financial analyst institutions have rubbished the way Nvidia spun the faulty figures. A third party company comes up with faulty figures and Nvidia issues misleading press releases to all and sundry. Hilarious!

How come no one questioned Nvidia's ethics in boasting about a record 3Dmark score that was based on an SLI based system with an unreleased motherboard, unreleased drivers and unreleased SLI hardware?

 

[jimmyjames123]

Actually, taking a look between the two Analyst reports the definition appears to be "Second Generation DirectX9 greater than 100M Transistors". Second gen DX9 is arbitary, but would force the count to exclude 9800's; then of course, 6200's are currently just NV43's which are over 100M transistors.

In otherwords, it was a rediculous count.

It would be helpful to see their count of how many 6200's, 6600's, and X700's were sold. I can't imagine that there were many 6200's sold since they just recently announced the product.

I also don't agree with that Analyst Report that they should have also included X600 and X300 cards. If this was the case, than what would have stopped them from including 5200 cards too? What would have been more appropriate would have been to exclude 6200 cards, but like I said above, how much quantity of these are available anyway?

 

[thatdude90210]

I thought the numbers were for chips that were shipped, not sold. So being just announced doesn't mean they haven't already shipped tons. I don't see anything wrong with them talking about it since it's from Mercury Research. As someone else said, when life hands you lemons...

What I fault Nvidia on is that they are so focused on the lunatic fringe, ie sm3.0/sli/duel-slot/2-molex, that they dropped the ball so badly... going from over 40-50% of graphic marketshare to 15% in just a couple of years is pathetic. And hovering at $450-500 million per quarter -- and even worst, minus $80 - 90 million per quarter once X-box production ends -- while your main competitor went from $350 million to nearly $600 million quarter is just as bad.

 

[jimmyjames123]

BZB, you are a one-trick pony

I think it's pretty hilarious that the usual suspects are criticizing people who are questioning NV's ethics, especially given that two major financial analyst institutions have rubbished the way Nvidia spun the faulty figures.

Apparently you don't read very well. Mercury Research put together the report. NV PR naturally showed off the best portions of the report. Any questionable report criteria is Mercury Research's responsibility. Your disdain for NVIDIA the company is making your arguments lose any objectivity and credibility.

By the way, can you tell me how many 6200's NV shipped, especially considering that the card has just barely been announced? Can you tell me which cards have shipped in more quantity: 6800+6600 vs X800+X700?

How come no one questioned Nvidia's ethics in boasting about a record 3Dmark score that was based on an SLI based system with an unreleased motherboard, unreleased drivers and unreleased SLI hardware?

This is a silly comment, considering that both ATI and NV boast about performance crowns using hardware reviews on new cards that are not even available to the public at time of review (like X800XT PE and 6800U)!

Also, when people actually do start to buy SLI setups, they will be able to easily achieve the 3dmark05 score that NV displayed. The 3dmark05 score that ATI displayed based on the ridiculously overclocked X800XT PE would be almost impossible for anyone else to achieve.

 

[Dave Baumann]

It would be helpful to see their count of how many 6200's, 6600's, and X700's were sold. I can't imagine that there were many 6200's sold since they just recently announced the product.

Mercury numbers are taken from a variety of sources - one of the primary being the volumes that the IHV's said they have shipped.

I also don't agree with that Analyst Report that they should have also included X600 and X300 cards. If this was the case, than what would have stopped them from including 5200 cards too? What would have been more appropriate would have been to exclude 6200 cards, but like I said above, how much quantity of these are available anyway?

Like I said, the split must have been "second generation DirectX9 with at least 100M Transitors"; 5200 would be excluded because it wasn't "second generation".

 

[jimmyjames123]

Like I said, the split must have been "second generation DirectX9 with at least 100M Transitors"; 5200 would be excluded because it wasn't "second generation".

Yeah, that criteria would fit the bill, although chategorizing based on manufacturer specified transistor count seems unbelievably shortsighted.

Maybe ATI will need to start counting transistors in a similar fashion to NV?

 

[Fodder]

How come no one questioned Nvidia's ethics in boasting about a record 3Dmark score that was based on an SLI based system with an unreleased motherboard, unreleased drivers and unreleased SLI hardware?

Because it was barely better (1/5) than a single X800 score.

 

[Mark]

How come no one questioned ATI's ethics in boasting about a record 3dmark05 score that was based on a ridiculously overclocked video card?

it wasn't so much the score they were boasting about as it was the ridiculously overclocked X800 video card...

 

[ChrisRay]

I really doubt the 6200 has accounted for that much, When this report was made. I have no certaintly on the 6600 either, But it is definately more available than the 6200.

 

[WaltC]

I think it's pretty hilarious that the usual suspects are criticizing NV's ethics. A 3rd party company comes up with some figures, and NV PR shows off the positive portion of these figures just as any other public company would do, and this is unethical? Hilarious. How come no one questioned ATI's ethics in boasting about a record 3dmark05 score that was based on a ridiculously overclocked video card?

It's pretty hilarious that the "usual suspects" as usual are almost totally blind to anything except the blind acceptance of whatever "positive portion" spin nV wishes to dish out in terms of PR propaganda. As I pointed out, nV made very similar misrepresentations in '03. It's not really "hilarious" though--it's just pathetic, if anything.

 

[PatrickL]

One thing seems true about all that: the next R520 will show if ATI is able to produce 2 competitives architectures in a row.
If ATI is not very well positioned in the high end part next generation, they will lose market share in discrete graphic card segment. But i wonder if that fight is a fight ATI want to have and if the other activities will not be more important and profitable for them ?

 

[Pete]

I'm sure it'll say as much about how the Xbox2 and Revolution contracts affected them, too, although perhaps MS and nV stumbled thru most of the learning process involved in a console GPU already.

I have to say, though, that I'd probably be slightly disappointed in R520 if it were SM3.0 but didn't offer FP blending and buffers (perhaps redundant, as one may require the other).

 

[assen]

I think you are placing way too much faith in the Xbox2 being a key decider in the desktop/oem graphic market share. It would definately affect profits...

If the Xbox shows anything, it will definitely affect profits negatively for ATI.

 

[Dave Baumann]

Not sure what you are getting at there assen, however a.) Xbox is very profitable for NVIDIA (post arbitration), b.) ATI has a very different deal with MS than NVIDIA did.

 

[thegrommit]

If the Xbox shows anything, it will definitely affect profits negatively for ATI.

During an earlier conference call, ATI stated that console parts will improve their average profit margins

 

[kemosabe]

I think you are placing way too much faith in the Xbox2 being a key decider in the desktop/oem graphic market share. It would definately affect profits...

If the Xbox shows anything, it will definitely affect profits negatively for ATI.

You are completely clueless. ATI is not manufacturing the Xbox2 GPU. TSMC is fabbing the hardware for MSFT and ATI will get a royalty payment on each unit sold. This means pure profit (i.e. 100% margins) on this business segment, just like it is now with GameCube.

 

[duncan36]

One thing seems true about all that: the next R520 will show if ATI is able to produce 2 competitives architectures in a row.
If ATI is not very well positioned in the high end part next generation, they will lose market share in discrete graphic card segment. But i wonder if that fight is a fight ATI want to have and if the other activities will not be more important and profitable for them ?

I dont think ATi would be showing experimental tech at an investors conference, its already done and working as far as I know.

I dont get your point about if this is a fight ATi wants? They're releasing their next gen part wayyyy before Nvidia and ATi is the GPU market leader. Anyway is it a fight if your competition doesnt have a competing card anywhere close to being ready when the R520 comes out?

 

[Bjorn]

Anyway is it a fight if your competition doesnt have a competing card anywhere close to being ready when the R520 comes out?

And you have proof of this i guess ?

Ati has to add full SM3.0 this time around. Nvidia could add some minor tweaks , a few more pipelines and be done. If they'll remain competitive with that solution is another thing though. But isn't the R520 supposed to be yet another tweak of the R300 architecture ?

 

[Richthofen]

Here's my question. Why the heck is Mercury research even lumping midrange/low end cards into the high end category?

Actually, taking a look between the two Analyst reports the definition appears to be "Second Generation DirectX9 greater than 100M Transistors". Second gen DX9 is arbitary, but would force the count to exclude 9800's; then of course, 6200's are currently just NV43's which are over 100M transistors.

In otherwords, it was a rediculous count.

According to the conference call yesterday the mercury report has been finished in september.
At that time there was no significant shipement of NV43 parts.
According to the management the report did not contain midrange parts from the new generation. I guess the high end numbers are mostly 6800U, 6800GT, 6800nonU and 6800LE.
It's no secret that Nvidia is better positioned in that segment and of course was able to deliver a littlebit more than ATI. Both still have delivery problems but ATIs seemed to be a little bit bigger.