oh boy 50% more powerful GPUs
it's not even a question how AMD will recover, because they won't
Oh boy 50% more powerful GPUs
Mason Cook
Other urls found in this thread:
polygon.com
twitter.com
Matthew Martinez
Who needs FP64 anyways?
Elijah Hall
it's useful for adult tasks, you don't need it for gaming
James Edwards
it's the same fucking node
this is bogus
Owen Long
Nvidia is known for their compulsive lying and their fanboys having more money than brains.
Jack Robinson
You mean AMD
Jonathan Cooper
it's for tensor cores in memelearning scenario
actual GPU tasks aren't even mentioned
John Jackson
>Nvidia
Uh oh
Connor Parker
For AI. Not gaming.
Luke Ross
Why don't you lying faggots fuck off?
Nvidia says FP32, not Tensor cores
Austin Scott
God damn I hate useless block diagrams
Jose Martin
>still gimping the consumer GPUs
Call when it's all FP16 divisible 'Tensor Cores'
Aiden Sanders
No asyc compute
Henry Nguyen
could you post a cropped image of the section where they say that?
Bentley King
You only read part of the site. You forgot to read the main crux. Its a deep learning architecture. The FPU32 performances alone don't do anything significant without major tensorcore changes.
This performance gain is through wholesale, not individual parts.
David Ross
nope, nvidia was literally forced to pay all 970 owners $30 for the 3.5GB thing.
polygon.com
Thomas Williams
But it did have 4GB, it's just the last .5GB was crippled as to not bottleneck the memory controllers. It doesn't make sense that Nvidia lost that case but AMD won their class action suit, where people accused AMD of lying about Bulldozer cores. yes, they were 8 real cores but performance was so gimped by two cores on each module it performed like a CPU using half the cores.
Dominic Jackson
The last .5GB wasn't usable.
Xavier Gutierrez
>Hey ill sell you a 4wheeled car!
>Good ill buy it. I really need 4 wheels.
>OOPS. the 4th wheel is actually made from wood. But you bought it now. And i didnt lie. it has 4 wheels
>wtf?
Daniel Baker
NO. IT. DIDN'T.
It had 3.5GB of GDDR5 vRAM and then a 0.5GB chip of DDR3. The lawsuit was successful because this was proven to be true and nvidia thus lied when the 970 boxes said (4GB GDDR5).
AMD got away with their 8-core claim because their processor actually did have 8 integer cores.
Christopher Harris
>May 10th
>muh big dies
>"data center GPU"
Why do you keep posting this? Consumer Volta isn't going to be out for almost a year.
Chase Ross
>It had 3.5GB of GDDR5 vRAM and then a 0.5GB chip of DDR3
wut, no it didn't lol, you're actual retardo.
It had 4GB GDDR5 but the last 512MiB was stuck behind a second crossbar which was not only slower (since it wasn't XOR'ed with the main memory, it wasn't actually 'slower' but isolated) but couldn't be accessed while the main memory was doing an IO op.
See pic related.
Kayden Reed
Thats just because of the tensor cores. Dont expect these bumps on customer hardware.
Alexander Lee
Whoa, then Fiji is 80% more "efficient" than Tahiti per CU on the same node.
W H O A.
Isaiah Rivera
It was usable but it could not be accessed at the same time as the 3.5GB and the link to it was slower. If you used all 4GB it may introduce stutter