dr_rus
Member
On quite the contrary, doing a rebrand of GCN to Polaris while shrinking the architecture from 28 to 14nm and getting the 2.5x perf/watt from this shrink sounds like something which they can afford. What I have my doubts they would be able to afford is to actually build a new architecture with Polaris which would be on the same level of difference to GCN as Pascal will be to Kepler - considering that this is the same timeframe of new architecture launches on NV's side. On the other hand GCN is a really nice architecture which doesn't need much updating even today but I hope that they'll support FL12_1 in Polaris at the very least.They can't afford that. They need to substantially outclass Nvidia's Pascal to regain marketshare and relevance within PC gaming.
It's obviously not. When's the last time they've actually done a rebrand on a high end gpu? 290x wasn't a rebrand. Fury wasn't a rebrand. Repositioning your former high end chip into the midrange isn't a bad thing for consumers, which is exactly what the 390/390x is.
Too bad that Polaris isn't a GPU.