You're right... I was thinking back to 2000's SGI. It's hard to track of when / where these companies are. SGI buys Cray, SGI spins Cray out to Tera, Rackable buys SGI, HP splits, HPE buys SGI...

It's all crazy.

But yes, Cray still has some impressive gear independently.

They are for sure not single image machines. Each node is running its own kernel, sees only its own resources and is scheduled by the system-wide batch scheduler node. Running across nodes requires MPI to tie things together (tenuously) at the application level.

If they are impressive, it's only because of physical size and aggregate resources available, not because there's the least bit of magic involved. It's the usual Intel+-NVidia CPUs/chipsets/etc. At the current time, it's mostly their legacy proprietary interconnect.

Could I have confused those with, oh the irony, some SGI machine? I remember someone had single-image boxes with up to a thousand CPUs. I assumed it was something derived from Cray's T3E.