Nvidia didn't release as much information on the GTX 1070, but we know it will have 8GB GDDR5 (not GDDR5X) and offer 6.5 TFLOPs (vs. 9 TFLOPs for the GTX 1080). It will have a $379 MSRP and officially launch on June 10.
Weird comment. Like a grandma driving a Ferrari with a broken gearbox: "I am perfectly fine driving at moderate speed, maybe I am not pushing the car hard enough. I am perfectly fine that the mechanic deactivated the malfunctioning higher gears".
You may not notice it nowadays because several games received patches and newer nVidia drivers limit the memory consumption to just below 3.5 GB. Otherwise you would experience a major slowdown. Also a major problem for CUDA.
If games truly don't need access to the last .5GB (and they seem to not in most cases if there's patches there) and all the parent does is game (which is what I do on my 970) then why would it even matter if the last .5GB is slow, or that the card wasn't being "pushed hard enough"? It's a gaming card, and using it for gaming seems to be working just fine, in my anecdotal experience at least. Gaming benchmarks have also been great for the card.
Just because not everyone writes CUDA applications that need access to all 4GB does not mean that people are using it wrong, or that they should mind that the last .5GB is slow. If it works for others' use then that's great for them.
I see you point, but I use a 970 and just finished Dark Souls 3 on max settings without a single hiccup.
No doubt that with future games the 970 will not perform as well as it should, but I will have a different card by then. What they did is a shame and shouldn't happen again, but I haven't noticed any real-world ramifications yet.
GTX 970: It was revealed that the card was designed to access its memory as a 3.5 GB section, plus a 0.5 GB one, access to the latter being 7 times slower than the first one. -- https://en.wikipedia.org/wiki/GeForce_900_series#False_adver...