JoshMST said:
Dear God Dave, who coached you with that bit of constructive criticism? Are you saying that 90 nm Low-K wafers are less expensive than 110 nm? Heh, well maybe I took "dinner plate" too far. Still, damn big die though.
Again, the die size itself isn't siginficantly different from G70, so I'm not really sure why these type of comments come now when the die size doesn't really appear to be particularly out of the ordinary from previous die growth - R580's die size is inline with previous trends, G71 is currently the the abnormality, and unless we see some further high end parts go in the same direction as G71 then R580 isn't really out of the ordinary.
However, even a quick look at two figures would tell you that there is something seriously wrong with the calcuations. Your costings for a high end wafer was at $20K (while I don't have anything definate, thats I figure that I have had floating around in my head for a while now as well, so I'll bite that) and then there is 164 cores per wafer for R580. This puts the cost per core at around $121 and applying a 35% margin (IIRC this is what they said for desktop without integrated, and there is likely to be a premium for parts like R580 over that, but we don't know how quantifyable that is), which puts the selling price per core at around $164 - thats for all chips on a wafer to come out at the profit margins they are stating. That's not unreasonable, roadmaps from last year suggested that R580's die cost was ~$200, so allowing for some yeild drop off we are probably in sane areas. However, thats not supported by a yeild as low as 36 cores per wafer as the selling price would need to be significantly greater than that.
If we do consider that the fully workable yeild is that low then that would also probably suggest that they actually have significantly more chip inventory at fewer quads, and yet it has been now 3 months since the release of R580 and no products with these configurations have yet been released - by contrast, 7800 GT was announced a few weeks after GTX. ATI said recently that they sold 50,000 X1900's, which would equate to ~1400 wafers, yeilding at 36 full cores, so they have this many wafers worth of potential product just lying around after the first 5 weeks?
The fact that the X1900 "GTO" isn't here yet, and the rumours that there is a specific RV570 chip to fill this spot suggests that the yeild drop-off for this configuration perhaps isn't as great as may have been thought, and perhaps they have a redundancy mechanism that isn't the same as seen with other configurations.
Also, the comparisons to the R520 products are a little out of place, given that the lifecycles are at opposing ends - or, alternatively, why not have compared NV42 to RV530?
