20 Jan
2014
20 Jan
'14
4:18 p.m.
On 1/20/14 7:40 AM, Ken Seefried wrote:
Yup. I know more than one FPS that ended up getting
tossed because no one
kept the software (and in one case, because the software was for a host
that no one had any more). Unfortunate...they're really interesting
devices.
Yup, it was the poor man's supercomputer. Take a vector processor, and glue
it onto a minicomputer (later onto a supermini). That's how first generation
GE CAT scanners worked. DG Eclipse and an FPS AP-120 array processor doing
convolution. Analogic and several other companies made them in the 80's. Not
much documentation or software around for any of them. I supported a stripped
down AP-120 called the FPS-100 on the 11/44 I supported at the Medical College
of Wisconsin before I moved to California. The FPS-100 was popular with oil
exploration people and MCoW picked one up cheap from a broker in Chicago when
they got stuck with a bunch after the early 80's oil glut.
There were a few supermini companies that had intergrated vector units by the
mid 80's, then the whole thing went away as microprocessors and MPP kicked in.
They were (are) also power hogs. A modern desktop CPU can blow the doors off 'em.