>10 times Maxwell
>1080
http://wccftech.com/nvidia-pascal-geforce-gtx-1080-gp104-gpu/
>>53434660
I very much hope AMD is competitive this gen. if only so Nvidia can't charge meme prices for things
>>53434711
nvidia probably won't raise prices for a generation or two, and they already gave us the massively undercut 970/980 pricing this gen. i'd be more worried about what happens once AMD goes under, rather than what happens in a time where their competition is token at best
>>53434711
>AMD
>not competive
Buying the first generation of a new architecture...
Let's see how this claim of 10x performance holds
Samsung memory is good news
>>53434957
These "ooh it's 100x faster than X" claims never hold true.
It might theoretically be 10x the performance in one thing, but I doubt it's going to show in real life.
They'll most likely get something like 50% increase in performance or something.
This year is going to be pretty damn interesting though.
So many improvements going to hit the market, a lot of competing new tech out there.
>The GP100 consumer variant will feature up to 16 GB HBM2 memory while the HPC variants will ship with up to 32 GB HBM2 memory.
I fucking doubt it
>>53434660
tesla, fermi, kepler, maxwell and pascal must be rolling in their graves for how much nvidia has disgraced their names
>>53434711
If anything, AMD's Polaris will be better at 4K due to HBM, while Nvidia's Pascal will run on lower memory bandwidth with it's GDDR5.
Mark my words, you will see memes of Nvidia GTX 1080 being a 1080p only card.
>>53436637
>AMD
>better
>at anything except burning houses down
>>53436651
meanwhile nvidia is the one actually burning down houses with their shield tablets
>>53436637
If they do the same thing they did last generation with everything being bumped down one level and the top level being taken over by the new fury X, then it's very reasonable that would happen. GDDR5X even going up against HBM1 is still a fucking laughfest in both performance and power draw.
>>53436637
>Nvidia's Pascal will run on lower memory bandwidth with it's GDDR5.
I thought Pascal was to be equipped with HBM2 modules. Maybe not all models but AMD isn't going to use HBM on all modules with Polaris either.
>>53436786
only the full chips. the ti and titan will offer hbm and their professional cards.
1080 and 1070 will use gddr5x which wont be shipping till may. which makes this article bogus about a may launch.
https://www.micron.com/about/blogs/2016/february/gddr5x-has-arrived
>There have been a variety of rumors in the industry about GDDR5X availability timeframes, so I’d like to clear that up from the beginning: Micron’s GDDR5X program is in full swing and first components have already completed manufacturing. We plan to hit mass production this summer.
>We plan to hit mass production this summer.
>>53437255
>only the full chips.
Thank you for telling me the same thing when I wrote:
>t Pascal was to be equipped with HBM2 modules. Maybe not all models
>>53434660
>http://wccftech.com/nvidia-pascal-geforce-gtx-1080-gp104-gpu/
can't wait for their 1080ti
what would be better, selling my 980ti now to get a 1080ti or get a second 980ti?
>10%
THEY ALWAYS LEAVE OUT THE FACT THAT NVIDIA HAS ONLY SAID THAT ABOUT HBM2, NVLINK, FP64 USAGE. LET ALONE ONLY ON THE FULL, BIG CHIP.
BTW NVLINK IS PROFESSIONAL MARKET ONLY FOR THINGS LIKE TESLA
>>53437279
Get a second 980Ti in the meantime if you want more power, then sell them both for 1080Ti SLI when the time comes.
>>53437270
ok?
i was going off what you stated but ok
>>53437279
you do know the ti isn't slated for 2017 right?
>>53437301
>i was going off what you stated but ok
I take it you do not run into very many rhetorical questions?
>>53437301
>you do know the ti isn't slated for 2017 right?
I still want a 1:1 traded upgrade, I don't want to go from a 980ti to a 1080 when I know there will be a 1080ti some time later
>>53436536
Why would you doubt it? Samshit & SK Hynix already said they'll manufacture 8 stack HBM2 later this year
4 stack HBM2 x 4 will give you 16GB capacity
