Originally posted by: thilan29
Originally posted by: Idontcare
As a process technologist, considering that NV was able to produce
this and sell it for <$300 is a testament to their DFM capabilities. You'll be hard-pressed, and I say this with confidence, to find ANY integrated circuit even remotely close to that diesize being sold for less than four figures, and for very good reasons.
They were sort of forced into selling it that low though with the competition they had. And I've read that they were losing money on those chips once they cut prices...although that was never proved beyond a doubt but I could see them taking the hit to keep their marketshare.
thilan that doesn't really negate my point in regards to making diesize comparisons as the basis for cost estimations alone, if that isn't self-evident then I'll concede I failed to properly make my point to begin with and I'll happily go back to the drawing board and try again if think you would find value in my doing so. I don't mind making the effort.
Originally posted by: SlowSpyder
For making an x2 card, wouldn't it be possible that building that part could take consderably less time though? No doubt that building the GPU can take years, but I would think (maybe incorrectly) that once they have the GPU building an x2 part could be done somewhat quickly. In Nvidia's case, I would think that building the sandwich card could be done quickly enough, the real work would be in the boards and cooler since the GPU is already there.
If I'm wrong on this, just tell me to sit down and shuddup.
Possible to take considerable less time? Of course. I'd be a fool to try and argue against that.
I'm not trying to make this an open and shut fact-laden argument. It is very much just like that Charlie article about NV yields being 2%, it wasn't impossible for that to be true, but in the scheme of things and how businesses generally operate the relevance of the data (if the data are even true) just doesn't make sense with how reality operates.
Can an X2 product be whipped up, expedited at lightening speed? You bet, 100% absolutely it can be done. At what cost? And to what gain?
Decision makers aren't as rash and quick to change programs as they are portrayed to be by the time their actions get filtered thru sensationalistic headlines in the media to reach our ears.
I am by no means trying to imply that an X2-type product requires anywhere near the development pipeline as a single-GPU derived product, but it isn't practical from any business sense to put the resources on rushing a project from non-existence to market in say less than 6 months time. 9-12 months is an economically viable timeline, still crisis mode and the R&D efficiency for your budget is totally blown, but you aren't going to risk ruination of the company putting a team on that kind of pace.
For example, I can tell you with 99% confidence that right now whatever GT400-derived X2 type product is going to come to the market is already locked in stone on Nvidia's roadmap for both budgetary reasons and project milestone cadence. They
could change the specs or plans, but doing so would come with markedly elevated risk to their market intersection timeline.
GT500-based product is still early enough in development that changes are manageable without significantly elevating risk or budget to counter the elevated risk.
I have seen, been a part of, crisis mode style management before and it is so not pretty and actually quite ineffective when it comes to maximizing shareholder value. Doesn't mean it doesn't happen, just means when we reach for our Occam's razor to try and pare down the probable scenarios to explain certain market events, based on experience I personally would be comfortable with ruling out the "sli on a stick" sku as being purely reactionary to AMD's X2 product strategy.
I don't have neatly compiled and presentable data to offer you in this situation to compell you to believe my opinion on this, but we do go back a ways and so I did want to take the time to just try and round-out your perspective on more of the backstory that can go on with these things. If we were in a bar having a few brewskies together right now there is no doubt we'd have scribbled all over a few napkins with some distribution graphs and timeline milestones by now
:laugh:
Facebook
Twitter