cray

But generic code is just always going to be cray-cray (See: Haskell lenses, C++ Boost, Scala collections).

Now we've all got a Cray in our pocket yet the industry is breeding faster horses.

Probably has something to do with never working in a datacenter, and never doing anything super cray with storage.

Funny reading back through the paper: The Cray T3E supercomputer is currently ranked the twenty-third fastest computer in the world.

Shockingly, nobody had ever asked Cray for something like this, and they really came through.

I remember the Cray X-MP in the early 80s supported up to 16MB and frequently came with 4 to start.

/scratch on a Cray is (last I used one) is a Lustre filesystem.

Old Cray sustained bandwidth is still higher than recent Core iX cpus, making them 'faster' for actual workload, even though their peak bandwidth is lower.

Unless I'm mistaken, a Cray supercomputer generally had a dedicated and consistent power supply, i.e.

I didn't realize Cray was the only vendor who went the no-local-storage route.

The comment that i think you're referencing in that article is talking about a cray x1, which was only 2 years old when the comparison was done.

Recently learned that reading this article about Cray computers.

They really serve no purpose other than "look how crazy this stuff is" - deep learning is cray-cray.

Both IBM and Cray had nearly all the features implemented per "Data enhancements and object orientation" on page 6.

Asked what were the secrets of his success, Cray said "Well, we have elves here, and they help me".

> "I think every engineer in the world would like to have their own Cray" Interesting, most of us have machines that are roughly analogous to a mid-80's Cray machine.

John Rollwagen, a colleague for many years, tells the story of a French scientist who visited Cray's home in Chippewa Falls.

The Cray’s speed does not come from just one processor, however.

But they still smoke a Cray-1, and yet we use them as dumb terminals.