They can't because they need four stacks for that increase in bandwidth which enables the extra performance.If they can bring out an 8GB version for £100 less.
They've probably already bought a lot of HBM2 memory.
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
They can't because they need four stacks for that increase in bandwidth which enables the extra performance.If they can bring out an 8GB version for £100 less.
They've probably already bought a lot of HBM2 memory.
They can't because they need four stacks for that increase in bandwidth which enables the extra performance.
I don't know enough about it to say, it's just that with Vega 64 they used 2*4GB and yet Vega is bandwidth starved. There must be reason why they didn't use 4*2GB.But why can't they use 4x2? Or doesn't it come in that size? Speaking of which, how do yields work with RAM? If you get a 4gb module with defects, can it be used as a 2gb module like they do with cut down CPUs?
Given the time (or lack of) AMD has been working on this and how quickly they transitioned to TSMC I suspect this is more an equivalent of a straight up design shrink than something designed from the ground up to take advantage of the node - they seem to have kind of jumped on 7nm in a hurry.
There are plenty of Turing compute cards too using GDDR6 check out the new Quadro RTX cards.
It is a nice looking card.
I like it, it's minimalist!It looks like the time spent on the cooler design was about 4 minutes: "A grey box, oh and make sure there's holes for the fans".