AMD-ATI Conference Call Transcript [LIVE]

That's the part I don't get. If this was a merger why would Nvidia's management have any more pull than AMD's beyond the fact that they seem a bit more aggressive. AMD certainly has more of a global presence and influence no? Why would AMD relinquish control of their fabs and tech to a bunch of snotty IHV brats?

Though I fear what Nvidia's management would do with control of AMD tech.

 

Who gets what --titles, power, board seats-- is usually determined by relative size tho. Once in a great while, not --sometimes a little company gets bought specifically to get a couple of key execs who then get a much higher position than their company size would "entitle" them to; see Ray Ozzie going to Microsoft for an example of that one.

I really do think that was code for "We couldn't make a deal for Jen-Hsun's participation go-forward that would please both him and us".

 

It's not about nVidia having more control, it's rather that it would have been a merger instead of a buyout, which would have placed the leaders of nVidia on par with the leaders of AMD, something which AMD would obviously prefer not to happen (who doesn't want to stay in control?).

 

Totally agreed, that's the way I read it too.

JHH ain't really known for his sharing of authority, nVidia is HIS baby and he ain't never gonna let anyone else forget it.

I can't really say I blame him either given how he started it off and built it, but his attitude ain't conducive to a merger.

 

Digitimes interview with AMD: http://www.digitimes.com/systems/a20060725VL204.html

Some interesting tidbits:

Anybody else hearing the old Reese's peanut butter cups commercial? "Two great tastes that taste great together!" What'll it be today, sir? AMD? ATI? or AMD/ATI?

The opportunity to drag the low-end up some might be being undervalued here a bit.

 

Nothing points to faster better GPU's in this merger.

I would like to see ATI chips made on AMD's fabs. AMD already said they aren't going to be made on their fabs...maybe in the future they will, but if they were really serious about making faster GPU's they would be made on ATI's fabs.

AMD bringing up the CPU/GPU in one chip idea. This can only be a low to midrange performance solution. Effort on such a chip will only drain resources from R&D of higher end GPU design. Also, there is no bottleneck of CPU to GPU. I dont see why it would improve performance. Look at the super fast memory they put on video cards these days. Memory thats like 2 or 3 generations faster than DDR2. And its dedicated video memory. A CPU/GPU solution that shared the same system memory == slower performance.

AMD/ATI merger means slower GPU progress anyway you look at it. The future of GPU's are not incorporation into CPU's. I could see them being incorporated into the motherboard however and off the PCI-e. So you would have a slot for your CPU and a slot for your GPU on the motherboard. That would make sense. Not sure about the memory though. DDR2 memory is too slow for todays high end GPU's.

 

I don't think it is actually. Integrated is not even low-end today. It is completely useless for any kind of reasonable gaming experience. Even if this AMD/ATi combination brings integrated up to the level of low-end discrete it won't make our high-end gaming experiences any better since developers today don't seem to target low-end or even lower-mid-range hardware.

 

pc.watch.impress.co.jp posted an article on AMD/ATI integration
It fuses the processor of AMD and ATI to one
(English translation here)
The article is about Torrenza and there is one diagram show how the two will be integrated on 'Vector processor core to CPU to integration'

on the last diagram, there are 4 configurations on the integration but it seems indicating that at the end, the vector processor would end up on the same die as a cpu

 

are those diagrams there own interpretation? Interesting stuff.

 

I dunno. When I hear EPIC bagging all over Intel's crappy graphics as a threat to the gaming industry, I tend to think the ISV's seriously disagree. Healthy ISV's equal healthier gaming.

 

This is alot to suck in. But despite Intels or AMD efforts. I have a hard time believing that this technology can ever replace a fully dedicated piece of hardware to graphics. Maybe I am not thinking in the right terms. But I just cant see it right now.

 

That's because its not suppose to replace a dedicated card. A CPU/GPU on die is about the low end and making the low end faster at a cheap price. Its not about replacing dedicated graphics cards.

 

Thats pretty much what I thought, But you'd be surprised across the vast amount of forums I visit the bizarre ideas that are being presented. ;p

 

Yeah there are a lot of crazy stuff going around right now. People just need to sit back and watch. In the end this is a good thing and the stuff coming from AMD and ATi about the merger has all been good and shows great forward thinking.

 

Re my question about the possibility about lower ATI-based card prices :

Is this a business rule?

I'm sorry but I don't understand why a 3D card can or should cost $499, regardless of your comments.

 

Where was a jusitifying the cards price? I was just simply pointing out that it probably will not come down in price and probably not get any fab space, at least soon anyway.

Also, consider things like R&D and also the fact that many of the high end cards have issues with binning, etc and you'll find quickly that the margins on high end cards are actually very thin.

 

I think you've got that turned around. R&D and binning costs are not high-end specific. Once yields are good the price/cost ratio on high-end parts is significantly higher than the lower-end. This is why chipset/integrated/mobile are all low margin markets.

 

There is a much higher margin on CPUs than graphics cards so by that respect graphics cards are relatively inexpensive.

 

McD's make more profit on a 'large' drink or fries than on a 'regular', which is why they want you to go large. The same applies pretty much throughout the business world. However there is still an advantage to the graphics card buyer in that while the manufacturer might make more profit on a high-end part (necessary due to smaller volumes of production, it is after all one unit) the customer often receives significantly better performance from their 'large' unit than they would from a 'regular'.

 

Look no further than ArtX & ATI -- I wouldn't think ATI flipped over the Flipper, but over Dave-O and his crack engineer team. (First R300, then CEO position.)
</Captain Obvious>