brain is exploding. Lets hope it has display port 1.3

Yes, my heart start to beat harder only when Intel launch new Xtreme and NVIDIA new 250W power consumption cards.
For everything else is interesting but not so much...
From other side I like this ASUS 980Ti Matrix... He is huge and overclocked properly.
But to be honest if I need to choose 800 euro for Maxwell crippled GM200 or NVIDIA Pascal GP100 not crippled for 1000 euro.
I choose NVIDIA Pascal. Both is not possible for me when distance between launch is so close, lower than 8 months.
Very important and good thing is when NVIDIA make same chip for professional and gaming purpose, Tesla and GeForce. 
Than we can expect best cards with less problems. Off course they will cut half of video memory, but processor is important. 

kougar
I wish CPUs could get this exciting again... very much looking forward to Pascal and Volta. 

Maybe next year with Amd zen? Bwahahaha couldn't even type that with a straight face. I hope for the best but expect the worst with Amd.

seth89
Sweet, looks like we may want SLI for 2016.

Fixed (why need when you can want :P)

Pascal sounds great, but NVlink sounds like it might be a MOBO nightmare. I already own a MOBO I would like to use for next years GPU... sigh

I found this:
Pascal and Volta GPUs Now Supported By Latest GeForce 358.66 Drivers
http://wccftech.com/nvidi...ds-support-vulkan-api/

Sounds like there will be a PCIE consumer version.

I would like to imagine gaming at the level to actually use/need NVlink, but I'm just hoping to see gaming utilize the best of what I hear about DX12. I think that could be the next huge up grade.

aka_STEVE_b
I'm interested to see what both companies do with the next-gen of HBM(2) of memory modules ....
and why systems haven't started getting the stacked memory cubes yet either  ?? 

You mean system memory like this?
Micron Will Unveil the Hybrid Memory Cube 3.0 Specification in 2016 – Significant Gains Expected Over HMC 2.0
Read more: http://wccftech.com/micron-hybrid-memory-cube-3-0-specification/#ixzz3rtTYZu7p
http://wccftech.com/micron-hybrid-memory-cube-3-0-specification/

candle_86
but can it play Crysis at 4k

You already know the answer is no.

z999z3mystorys
...also more than triple the bandwidth of current memory is nice, I just wonder how much that will help the processor when it's never waiting on memory.

Just find a scenario where a reliance on memory bandwidth is created: a GPU-bound resolution and setting.  The only example I know where a real-world triple in bandwidth had an impact is with Intel's Iris Pro IGP.

Intel's HD 5100 has 40-eu at 1100MHz, but depended on dual-channel DDR3-1600 system RAM at 25.6GB/s for bandwidth.
Intel's HD 5200 has 40-eu at 1150MHz, but has an eDRAM cache running at 75.8GB/s for bandwidth.

The effect is a 3Dmark11 Performance preset score of 1151 on HD 5100 and 1940 on HD 5200--  70% increase with a negligible change in the GPU spec.

Therefore, hypothetically, if nVidia remade GTX980 Ti with HBM2.0, the card would be faster while using less power.  Although the caveat is a hard pill to swallow as the graphics details would have to be so high (fully GPU-bound) that the frame rates wouldn't be optimal, i.e. force 4K surround on a single card.  I mean, seriously, who would push the graphics details on an IGP just to make it scale so high?

But of course, I don't think nVidia isn't using HBM for the bandwidth (that is just marketing), I think they are using it for the power savings, i.e. an excuse to use larger GPU dies since the GDDRx takes up such a share of space and power on the PCB.  Pascal Titans won't be 250W+ cards unless the GPU itself is larger than 600m2 despite the shrink from 28nm to account for the need for power saving HBM.

The hype is strong in this article. Cynicism in me will wait till I see real-world data on the new architecture from those that don't have a vested interest in the product line.

candle_86

Scarlet-Tech

candle_86
but can it play Crysis at 4k

You already know the answer is no.

i can dream cant i

I would say it's entirely possible, Here is an older article that shows SLI 780s running it at 4K 60fps. http://www.anandtech.com/...deon-r9-290x-review/13

Some news say Pascal in April. Again GeForce TITAN. 
I think before summer everything will be finished and we will see maybe full GP100 maybe not.
I hope they will repeat same scenario as with TITAN X, because with Kepler they cut some CUDA cores.
 
http://www.guru3d.com/news-story/nvidia-big-pascal-gp100-to-debut-in-april-mid-range-gtx-1080-in-june.html
 
It will be interesting to see how new cards behave in DX12.