Indiana University has purchased a new Cray supercomputer called the Big Red 200. Assembly is currently in progress, and whilst it could be completed on a much shorter timeframe, the university has decided to postpone the GPU installation until summer in order to get access to Nvidia's next-generation hardware, splitting the assembly up into two phases.
The exact details surrounding the next-generation hardware are scarce. Nevertheless, Brad Wheeler told The Next Platform that although the original plan was to fit the Big Red 200 with Nvidia Tesla V100 GPUs, a last minute opportunity gave them the chance to postpone the GPU installation and opt for the next-generation GPUs -- presumably the Tesla GPUs stemming from the upcoming Ampere microarchitecture.
This alternate plan is because the new GPUs are expected to offer 70-75% improvement in performance over the current-generation hardware, as noted by The Next Platform, the Volta-based V100 GPUs that would have been installed otherwise. For the IU Big Red 200 supercomputer, it means that the performance will still jump from the original predicted 5.9 petaflops to 8 petaflops with fewer GPUs.
The Big Red 200 succeeds the Big Red 2 installed in 2013. The new supercomputer is a Cray Shasta machine and is being built using 1,344 AMD Epyc 7742 processors. With 64 cores per CPU, the supercomputer will have a total of 86,000 cores and 172,000 threads to play with. This makes it one of the smaller supercomputers based on the Cray Shasta platform, as the UK Research and Innovation's Archer 2 system packs a significantly bigger punch. The GPU count is expected to be 256 units. Nevertheless, it will be the first Cray Shasta supercomputer in operation.
Exactly when the GPUs will be available or announced remains a mystery, but Indiana University has indicated that 256 Tensor Core GPUs will be installed in the fall. Given that timeframe and Nvidia announcing GPUs at GTC conferences in the past , it's likely we'll be see details of the Ampere architecture revealed at GTC during Nvidia's keynote on March 23.
Granted, this must be compute tasks which doesn't translate to gaming, but with that much of an improvement, the 50% rumor in gaming looks very attainable.
They are sometimes used for other things. Even parts of "computer things" that are ostensibly for "games".
A decade ago:
https://phys.org/news/2010-12-air-playstation-3s-supercomputer.html
Going by product names alone, the RTX 20-series cards were around "50% faster" than their similarly-named GTX 10-series predecessors at launch. However, in reality, the cards names were simply shifted up to the next higher tier to disguise what were actually far-smaller performance gains at any given price level.
It was similar with AMD's RX 5000 series cards, though at least there they added an extra digit to help differentiate the two naming conventions (but oddly stuck with the same first number) An RX 5700 may be twice as fast as an RX 570, but it also launched for double the price. Due to the move to the the 7nm manufacturing process though, the graphics chip of a 5700 or 5700 XT isn't actually much larger than that used by an RX 570 or 580.
Comparing cards using the full processors, an RX 580 at 232mm to a 5700 XT at 251mm, AMD saw around an 85% performance uplift relative to the size of the graphics chip. The launch price of the card was around 75% higher than that of the 8GB RX 580 launched over 2-years prior though, so not much was really gained in terms of price to performance.
It sounds like Nvidia will be moving to a new process node soon as well, so there could very well be large performance gains for chips of a given size, though that doesn't necessarily mean there will be a substantial increase in performance at a given price level.
I'd wager their pricing was more dictated by 7 nm supply constraints than anything else.