Thank goodness. They haven't innovated in decades and x86 is a mess. I can't wait until RISC-V does the same to ARM. It's super frustrating to see America, which used to be innovative become the complacent corrupt incumbent.
I guess you didn't read about their Xeon Phi bi-directional ring architecture that achieved dozens of terrabytes/sec of bandwidth, their 512b vector-processor ISA, or, heck, just for giggles, the industry's first FinFet (and high-k dielectric) on p1270, p1272, and p1274. But, please, do go on.
Intel has tons of cool stuff going on, but from the outside it looks like a big risk to adopt any of it. Optane is pretty cool, but who will be surprised if they EOL it tomorrow? AVX512 is obviously of certain utility, but 99% of computers lack it. I myself just got access to a CPU with this feature a week ago. They've got a sweet programmable 100gbps NIC that they only ship into the mobile base station market. Why? KNL is a complete joke that nobody wanted or wants. Features that everybody wants, like UMWAIT, only exist on obscure parts you can't get. Having a lot of cool hacks isn't making a great deal of visible success for Intel.
And it was such an outstanding success! Intel's customers were able to get so much use out of it - it'll always be remembered as a high point in supercomputing.
Seriously - it was a resounding failure, even with Intel spending big bucks to convince this class of customers it was the future. There was little uptake and vast dissatisfaction with achieved results. IMHO I expect Xe to follow the same (very shallow) trajectory in HPC as in gamer graphics.
There is one remaining large machine (DUG "Bubba") using Xeon Phi and that's only an artifact of software tolerance combined with a _really_ great deal received on Intel's remaining "Knights Whatever" parts inventory.
But whatever real innovation underlay Xeon Phi/KNx was performed two decades ago. Current Intel is only capable of SKUsmanship - minutely partitioning function for projected max margins.
Edit: Also too, remember that KNx was really an attempt to salvage Intel's failed Larrabee GPU effort.
> Also too, remember that KNx was really an attempt to salvage Intel's failed Larrabee GPU effort.
I do, and I don't think you know that the first "Knights" project was simply a rename of Larrabee. Same exact product. It was called Knights ... Landing I think? Then it became Xeon Phi.
Knights Ferry was the 1st. Unclear about the actual diffs between it and Larrabee silicon. I do know the 1st boards still had a video out port but it was disconnected.
Xeon Phi bi-directional ring architecture that achieved dozens of terrabytes/sec of bandwidth
...and yet, was it any better than a midrange GPU in compute?
Intel is the new IBM. IBM at the end of the 80’s was filled with smart people innovating piecemeal technologies that couldn’t be brought together to create compelling products. And the products that they were selling were expensive and technologically moribund.
Nope. That's were many people are wrong. Intel's main competitors are not in the US. They are in Taiwan and Korea. Intel's domination came not only because of a superior chip design, but due to the fact that only they could manufacture the superior chip design.
Now since that is not an issue anymore, you see Intel getting beaten everywhere. And that is only because Intel's actual competitors in Taiwan and Korea are have now surpassed Intel.
