Re: The Seymour Cray Era of Supercomputers

On 2025-05-21, John Levine <johnl@taugh.com> wrote:

Says here late 1964. It was a huge embarassment to IBM. I imagine a large part
of that was that it blew up IBM's longstanding belief that you had to make a
computer really complicated to make it fast, viz. STRETCH and 360/91. 
>
https://mncomputinghistory.com/control-data-corporation/
>
IBM sort of came around to that with the 360/44, which implemented a scientific
subset of the 360's instruction set and ran nearly as fast as a /65. It was
intended for process control so they added priority interrupts and some real
time I/O.

I think it was early 1970 that I visited Haldor Topsoe (chemical
engineering co in Denmark) which had recently installed a 360/44. I was
disappointed to learn that it was not program compatible with other 360
machines, so it had to use a tailored OS; AFAIR a modified DOS system.
And I think also it had a different floating point format.