If you have been watching the keynote speech of Nvidia CEO Jensen Huang this week, you probably will have noticed that there wasn’t any talk whatsoever about anything gaming hardware related. Many had hoped to hear at least something about Turing and Ampere.

Some new chatter on the web is that we might see an announcement or two in the late summer August timeframe, with maybe some announcements in July 2018? It all remains speculation though. The new information comes from SK Hynix who states that GDDR6 will reach mass production in three months. That means that the graphics industry is actively awaiting the mass volume production of GDDR6. History has learned us that once memory is available, things can move really fast.

GDDR6 will be able to achieve bandwidth running up towards 16Gb/. Though that is the high-ballpark number, I expect something in the 14Gb/s personally. It will be available in 8Gb and 16Gb densities. It gets sweeter as SK Hynix confirmed that this memory type will be used on (we expect all mainstream to high-end) Nvidia graphics cards. The SK Hynix info was derived from Steve Burke over at GN. One question remains, what would and could Nvidia actually release? We’ve heard two speculated codenames,  Ampere and Turing. It might be the case that Nvidia will split the gaming GPUs (Ampere) and then create a separated optimized line for anything blockchain. Much like what Quadro is to GeForce.

What do we know thus far?

All GPU names are still based on speculation, nothing was confirmed by Nvidia. The hearsay is kept alive by feeding small bits of info explaining or denying stuff. The cards would be released in March, then April, then all of the sudden it wouldn't. It was Ampere, all of the sudden there as Turing named. There even was a specific launch date mentioned, April 2018, the 12th, again based on nothing substantial. Neither Ampere or Turing ever surfaced on any of Nvidia's roadmaps, ever. To date, Nvidia has not talked about it. Logic dictates that Nvidia will release new stuff this year, but even that remains an assumption. We've seen it in the past so often, Nvidia lives and thrives on innovation, which gives them their reputation they have these days. So with that in mind, it would make much sense for them to release a GPU dedicated to mining (and here I'll mention Turing) and separate Gaming GPUs from cryptocurrency GPUs (Ampere). Nvidia could potentially optimize their drivers for dedicated mining based cards, getting better hash rates with such cards, while keeping cost down by lowering the amount of graphics memory, cooling and display outputs. Makes sense right? IMHO that could help settle down the price inflated gaming graphics card market, allowing gamers to purchase gamer graphics cards at more normalized prices where miners would get the best results with their own dedicated GPUs. 

Where did the Ampere name originate from?

This info and name is derived and traced back towards a German website called 3dcenter, and is nothing other than speculation based on a few observations. Nvidia would be readying a GPU series called Ampere, and not Volta. The name "Ampere" has been mentioned in the past a couple of times already, in fact, we wrote a few items about that GPU name popping up, constantly. It's mentioned that Nvidia halted production of the GP102 (e.g. 1080 Ti / Titan X) and likely GP104 GPU (e.g. 1070/1080), all originally released back in 2016. So, the question then arises, what is happening with Volta GPUs? Where are they?

So if it is Ampere, Why not Volta?

Currently, Volta is based on HBM2 memory which is not available in large enough mass-volume quantities, and it is expensive to purchase and implement as well. It also has an additional architecture layer, Tensor cores, that's expensive and will not benefiting gaming. So why produce high-end graphics cards, or better yet, GPUs with Tensor cores for gamers? Then there are the recent announcements on GDDR6 graphics memory. Earlier on we reported about GDDR6 closing in, really fast and actually already announced as available. Currently, it is already possible to fab graphics cards with the blazingly new fast GDDR6 memory. And the new info from SK Hynix is that mass production is available in three months. So my thesis here is that Volta will remain on track for the HPC / enterprise side of things including data-centers with its Tensor cores, and Ampere GPU series would be used for the consumer slash gaming parts. 

Ampere; a new architecture or Pascal refresh?

That we do not know. Considering it's not on any roadmap, we assume it is a Pascal refresh with GDDR6. Ampere as an architecture never made it into the long-term roadmaps from Nvidia, never ever, hence logic dictates here that Ampere would be based on refresh Pascal architecture, perhaps fabbed on a smaller node and optimized fabrication process. Basically, Nvidia would release say the GA104 and then we can speculate onwards, in a speculated chart, based on just assumptions and the rumors:

And BTW we named the new cards to say GeForce GTX 2080, but this very well could be named GTX 1180, of course. In closing we also need to realize, SH Hynix might be saying that they are ready with their mass production in three months, but that does not automatically mean Nvidia will out new cards, we assume that much. But if NV plans the Christmas season for something, hey that's their prerogative as well of course.

GeForce GTX 2070 and 2080 Could Launch Summer 2018