When were you when Nvidia died?
>No HMB2 memory.
high bandwidth memes
Forgot pic.
hardware async memes
>>53558110
Buy a Titan if you want HBM2, stupid goy. You're not poor... are you?
>>53558067
Did you miss the Polaris news?
>>53558157
Still HBM, just HBM1.
>>53558110
Sweet, when can I order them?
I've been on a 680 for a while now.
GameBrokes meme
>>53558146
Fucking jew
>>53558067
HBM1 is overhyped
>>53558196
GDDR5 8Gbps@512 bits bandwidth = HBM1 bandwidth
>>53558067
>No HMB2 memory.
2017 for AMD
2016 for Nvidia
>>53558608
so we are not getting hbm2 at all in 2016.
kinda sucks actually since i wanted my low power SFF monster gpu
>>53558574
6 stack HBM1 bandwidth = 50% more bandwidth at significantly lower power consumption
>>53559061
Yeah but it's limited to 4 GB (I think)
>>53558608
There's not hbm2 at all this year, shill.
The value of HBM is honestly vastly overstated by AMD and AMD fanboys.
Anyone who's worked with GPUs before can tell you the biggest bottleneck is CPU->GPU->CPU round trip data transfers. While HBM significantly increases data transfer speeds within the GPU space, it does absolutely nothing to help with the largest bottleneck that GPUs suffer from today.
>>53559103
It's not.
>>53559061
>>53559369
4GB is the limit
>>53559539
No, it's not.
https://news.samsung.com/global/samsung-begins-mass-producing-worlds-fastest-dram-based-on-newest-high-bandwidth-memory-hbm-interface
>January 19, 2016
AMD HAS NO GPUS AND NO HBM2
HAPPY WAITING FOR 2017, AMDPOORFAGS
>>53559796
It's almost like you think you'll get HBM2 this year, nvidiot.
>>53559539
HELLO NVIDIA DF
>>53558067
If amd can afford to use hbm2 it cant even be any good. I bet nvidia will skip it entirely.
http://wccftech.com/sk-hynix-hbm2-mass-production-q3-2016/
>AMD'S PARTNER SK HYNIX CAN'T EVEN START MASS PRODUCTION UNTIL Q3
>SAMSUNG ALREADY STARTED MASS PRODUCTION IN JANUARY 2016
IT'S OVER, AMD IS FINISHED & BANKRUPT
HAPPY WAITING 2017, AMDPOORFAGS
The amount of times I've seen "wait for Pascal" or "wait for Polaris" on this board.
Have fun waiting 6 months after the release of GP204 for the bullshit Titan Q model and 9 months after for the 1080ti.
980ti already kills every game on the market, literally upgrading to play no games.
hbm video cards have smaller form factors, which is why we have gems like the nano.
>>53559369
>>53559763
You can't physically put 6 or 8 die, 4 is the max = 4GB
>>53559838
Nvidia has the same problem, it's why they skip the HBM1...
>>53559763
Skylake it's the limit
>>53559369
It is. Don't you think AMD would have released that dual-GPU monster of theirs with something more than 4GB per GPU?
Go be a fucking retard somewhere else.
>>53558067
amd only puts hbm in their highest tier card too, will you shills shut the fuck up about it already.
