Thats in gtx1080 territory. So what will vega 64 be in relation to the 1080? I seriously hope the figures are true, it will certainly kick the competition into high gear. If so, I can foresee Nvidia rushing out Volta earlier than expected.
Thats in gtx1080 territory. So what will vega 64 be in relation to the 1080? I seriously hope the figures are true, it will certainly kick the competition into high gear. If so, I can foresee Nvidia rushing out Volta earlier than expected.
More than likely, price cuts, if true, and that is a big if.Thats in gtx1080 territory. So what will vega 64 be in relation to the 1080? I seriously hope the figures are true, it will certainly kick the competition into high gear. If so, I can foresee Nvidia rushing out Volta earlier than expected.
More than likely, price cuts, if true, and that is a big if.
Jarred Walton • a day ago
I don't think I can discuss the full specifics of any performance figures yet -- gotta wait for the actual launch. And, perhaps!, performance will be better than we were led to believe at the event.
I skimmed a few places for 1070 benchmarks on similar equipment and those in that leak to compare against seem a bit on the lower side.
PC Gamer Vega article author -
http://www.samsung.com/semiconductor/foundry/process-technology/14nm/I should have added. "In Theory".
Remember Intel complaining that it's competitors node names were bullshit? This is what it is all about.
The intention of Node Names were that full Nodes were to be double transistor density so ~.7 linear scaling.
But marketing departments have got too involved in the naming process, and thus we get a lot of fudging of supposed full node advances.
Do you have link where Samsung claims that. I would love to see in writing where they admit that 2 full Marketing nodes is only equal to one real node.
http://www.samsung.com/semiconductor/foundry/process-technology/14nm/
^Literally first link when I googled "Samsung 14nm lpp"
“Not all 10nm technologies are the same,” said Mark Bohr, a senior fellow and director of process architecture and integration at Intel. “It’s now becoming clear that what other companies call a ‘10nm’ technology will not be as dense as Intel’s 10nm technology. We expect that what others call ‘7nm’ will be close to Intel’s 10nm technology for density.”
It wasn’t always like that. Traditionally, chipmakers scaled the key transistor specs by 0.7X at each node. This, in turn, roughly doubles the transistor density at each node.
Intel continues to follow this formula. At 16nm/14nm, though, others deviated from the equation from a density standpoint. For example, foundry vendors introduced finFETs at 16nm/14nm, but it incorporated a 20nm interconnect scheme.
Technically, the foundries didn’t introduce finFETs at a full node (14nm), but rather at a half node. TSMC, for one, calls it 16nm.
Still, foundries found a way to provide value to their customers at 16nm/14nm. “Foundries are less fixed on sticking to a 0.7X pitch shrink per node and more on providing their customers with some combination of power, performance, area and cost benefit at a half-node cadence,” said Mike Chudzik, senior director of strategic planning at Applied Materials
That's pathetic. Intel is doing better than that on one node, than Samsung did on two.
Which justifies Intels complaints about their competitors process size claims.
Here is what proper node advancement looks like:
Intel has more than doubled density at each node step. Though 22->14 is 1.5 nodes, but it gives the proper 2.5X scaling (If I did my calcs correct). The big winner is 14nm-10nm, 1 node and bigger than 2.5X jump in transistors.
https://semiengineering.com/10nm-versus-7nm/
Lets see here.That's pathetic. Intel is doing better than that on one node, than Samsung did on two.
Which justifies Intels complaints about their competitors process size claims.
Here is what proper node advancement looks like:
Intel has more than doubled density at each node step. Though 22->14 is 1.5 nodes, but it gives the proper 2.5X scaling (If I did my calcs correct). The big winner is 14nm-10nm, 1 node and bigger than 2.5X jump in transistors.
https://semiengineering.com/10nm-versus-7nm/
See above to rethink conclusion.Interesting
Fury hit 14.9 MTr/mm^2 on 28nm which was equal to Intels 22nm (they increased density by 22% from Tahiti)
Vega on the other hand is at 25.8 MTr/mm^2 which is quite a bit lower than Intels 14nm. (So far AMD has only increased density by about 4-5% from Polaris on the current node)
So Intels 14nm is a good 45% more dense.
Lets see here.
Year 2016
14nm
Core i7-6950X
246mm^2
3,200 million transistors
Ta Da:
13 MTr/mm^2
See above to rethink conclusion.
Lets see here.
Year 2016
14nm
Core i7-6950X
246mm^2
3,200 million transistors
Ta Da:
13 MTr/mm^2
See above to rethink conclusion.
thoughts?