IdiotInCharge
NVIDIA SHILL
- Joined
- Jun 13, 2003
- Messages
- 14,675
Indeed, Nvidia's architecture is very efficient, allowing them that power usage on a larger node. But so what. As a consumer, a 5700XT and 2070 super perform the same and use the same power.
The 'so what' is that we're talking about future products, where AMD is already reaping the 'gains' of a node shrink which has allowed them to 'catch up' to their competitor's product that is still made on a larger node.
And then Nvidia has a new architecture coming on that same smaller node.
However, AMD does still have architecture refinements. And their roadmaps are pretty aggressive right now. I'm betting that Nvidia still retains the top crown at the high end, with Ampere. But, I think that AMD will remain not too far behind, as they are now. And their R&D will close the gap sometime next year. I'm feeling like they may even squeeze out an Ampere competitor, before spring 2021. Even if it means another higher power usage part.
That's the thing -- the best indicator of AMD changing the course they've been on since they started making GPUs / bought ATI is Navi, and well, it's more of the same.
And it's still missing hardware ray tracing. So they not only have to make a huge leap in raster performance just to approach competitiveness, they also have to add ray tracing hardware into the mix.
The chances of AMD doing both, given their history?
I'd love to be optimistic, but I'm too rational for that.