Paul G. Allen wrote:
On Mon, 2007-10-08 at 16:17 -0700, Andrew Lentvorski wrote:
Well, it served its purpose. It wiped out MIPS, SGI, HP, and DEC as
microprocessor competitors and left only IBM.
Actually, Compaq killed the Alpha, not Intel. DIGITAL was ready to
release a new, far faster Alpha when Compaq purchased them and killed
it.
Speaking as one of the Alpha designers, that's a little off.
There was no mythical "far faster" Alpha coming out. EV-8 was to be
multithreaded even then because the circuit designers and architects
were at the limit. The circuit designers and architects had already
wrung the fat out of gates and circuits. Gains were going to come with
increasing difficulty.
In addition, Compaq may have "officially" killed Alpha, but the death of
Alpha was the selling of the fab to Intel. Once Intel had the fab, they
weren't exactly motivated to help Alpha. So, Alpha had to change to the
IBM fabs. That cost time and money at a critical point.
However, the biggest failure was EV-6 (21264). A 300+MHz Alpha released
when Intel was still sitting at 66MHz would have provided momentum.
Instead, EV-6 got held up for two years due to ego wars between several
of the key folks. Yeah, EV-6 eventually came out at 500MHz, but by then
Intel had Pentium Pro's at 200MHz. Instead of a 6-fold (roughly)
performance improvement at about a 4-fold price bump, Alpha only had
about a factor of 2 for about a factor of 10 price difference.
Samsing continued to make processors for some time, but even those
may be no more.
Samsung is ... interesting. Samsung wants nothing less than the total
domination and subjugation of Intel. Making Alpha's was expected to be
a step along that path.
They didn't factor in the shortsightedness of American business.
-a
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
