At this point it should be pretty obvious that the first chip released WONT be the fully fledged chip. It will be the x60 chip "upgraded" to an x80 chip again. They might even come up with even a new way to market it. Nobody can act like there was no prior notice this time.
Makes sense from a business point of view:
1) 980Ti is faster than Fury X so why release a 70-80% faster Pascal GP100 chip for $650-699 and cannibalize massive profits on the 980TI when you are under no pressure to do that? NV can just utilize the same strategy of
marketing flagships aka $499 680 2GB/$579 GTX680 4GB, $650 780 and $550 980. There is no way they are going back to selling those cards for $250-300 considering how much profits they have realized with this new strategy.
2) There is likely huge pent-up demand for Tesla cards with DP. NV would be able to sell these cards for $5000+ to these customers vs. maximum $1000-1500 on the PC as a Titan Y? Even less incentive to release a 90-100% unlocked GP100 for $699 right off the bat.
3) Going back to point #1, since GTX680 and on, customers as a whole group have shown by voting with their wallets that they no longer care if the chip is mid-range or true flagship -- they will pay flagship prices for more performance regardless on where the chip lies in the NV architectural designation, even if it's just 15-25% more rather than the flagship historical 50-100% more. NV has heard this loud and clear. NV now has the data to completely abandon the decade(s) old way of launching graphics cards and move forward with bifurcating a generation, or maybe even splitting it into 3 parts like they did with 680->780->780Ti :awe:. Why would any business deviate from a strategy that's been SO successful and when consumers do not show resentment?
I think it is still possible for us to see a $650-700 GP100 Pascal but if it does launch in April 2016 as a consumer graphics card, I would bet it'll be something like a cut down successor aka in-line with a GTX780, not the full fledged flagship / direct replacement of the 980Ti/Titan X the Pascal generation. I actually foresee a mid-range successor to 980 at $550-600 and would be pleasantly surprised if NV goes back to the good old days. ^_^
On the whole though looks like Pascal is moving along very nicely - since the chip already tapped out with such a massive die and 1TB/sec HBM2, it looks like from their engineering end the fabs, things are smooth. Sounds like they have the execution down which means the only thing left is figuring out the order of releasing low-end to high-end cards to maximize profits over the next 2 years and ramping up yields. HBM2 and 16nm should provide
massive gains in performance just by looking at GTX580->780 as a point of reference of a full node shrink + new architecture.
Mental note: as of October 2015, the cheapest GTX980 is $480 US. It will be interesting to see how much that level of performance costs by October 2016 with Pascal. I am expecting big improvements in the $450-550 price level by end of 2016. I expect much lower improvements in the $200 segment though where cards like the R9 390 are creeping down to $260-265 and are at least 50% faster. We might see another generation where the best cards to buy are $250+ and everything below is overpriced and under-powered offerings like right now.
Facebook
Twitter