It's been a long time coming but NVIDIA is ready to announce their new consumer graphics card next week. Looking back in the past, they’ll start with the GeForce GTX or what I now believe will be called the GeForce RTX 2080 series. In this weekend write-up I wanted to have a peek at what we’re bound to expect as really, the actual GPU got announced this week on Siggraph already. There’s a number of things we need to talk you through. First of names and codenames. We’ve seen and heard it all for the past months. One guy at a forum yells Ampere, and all of the sudden everybody gossiping something called Ampere. Meanwhile, with GDDR6 memory on the rise and horizon, a viable alternative would have been a 12nm Pascal refresh with that new snazzy memory. It would be a cheaper route to pursue as currently the competition really does not have an answer, cheaper to fab however is not always better and in this case, I was already quite certain that Pascal (the current generational architecture from NVIDIA) is not compatible with GDDR6 in the memory interface.

Back in May I visited NVIDIA for a small event and talking to them it already became apparent that a Pascal refresh was not likely. Literally, the comment from the NVIDIA representative was ‘well, you know that if we release something, you know we’re going to do it right’. Ergo, that comment pretty much placed the theory of a Pascal respin down the drain. But hey, you never know right? But there also was this theory. Another route would have been Volta with GDDR6. And that seemed the more viable and logical solution. However, the dynamic changed, big time this week. However, I now need to go back towards February this year where I posted a news-item then, Reuters mentioned a new codename that was not present on any of NVIDIAs roadmaps, Turing. And that was a pivotal point really, but also added more confusion. See, Turing is a name more related and better fitted to AI products (the Turing test for artificial intelligence).