Chuck writes:
I pointed it out because (a) the Intel i860 was yet
another stumble
by Intel to get away from the 8008 architecture (the 432 being only a
slightly earlier one--and one in a diametrically opposite direction)
and the application (cluster) was novel.
Personally I always thought the i860 was far from a stumble, because in
fact it was the highest MFlops/dollar and MFlops/watt for a considerable
length of time and used in a number of high-horsepower applications
in the real world.
Farms/clusters weren't awful novel, scientists had been assembling their
own for a decade before. Intel seeing that it could commercialize a massively
parallel computer, yes in some people's eyes that would make it legit
(when somehow it wasn't legit whenever anyone did it in the decades before...?)
Maybe by some measures i860 was a stumble because it didn't replace
x86 but by those measures, everything is a failure. Hardly seems like
a good definition of success. Aka "Where Are We Going? Planet Ten!
When Are we Going? Real Soon!"
Tim.