[Info-vax] Poulson info from Dave Cantor

[John Wallace]

On Nov 27, 4:10 am, JF Mezei <jfmezei.spam... at vaxination.ca> wrote:
> Neil Rieck wrote:
> > So today, Intel and HP are in a codependent relationship.
>
> I am not so sure. I have a *feeling* that Intel is just a contractor,
> building IA64 things for HP and  Intel is perfectly happy with the huge
> sums of money HP is sending over to continue funding this chip.
>
> > Like a
> > thoroughbred horse, this Itanium thing is capable of really high
> > performance but one metaphoric accident in the market place may cause
> > Itanium to be put down
>
> I do not see the IA64 as a thouroughbred horse. Seems to me like Intel
> started on the wrong foot with it and it is mired in molasses that makes
> it very hard to move fast and keep up with the rest of the market.
>
> If Intel was able to turn the 8086 toy controller into a high
> performance 64 bit chip, then it should have been able to produce that
> IA64 with unmatched performance since it was meant to be a "clean"
> design unhindered by the toy controller era of the 8086.
>
> It looks to me like the whole philosophy and architecture of IA64 makes
> it very expensive and tedious to upgrade , and this matches what Digital
> was saying back in 1999 about it.

What were DIGITAL saying about Alpha vs IA64 performance in 1999?
Well, for example, there was a rather anonymous looking 32-page
whitepaper comparing Alpha and IA64's opportunities for exploiting
parallelism (out of order execution and simultaneous multithreading
for Alpha, compile-time stuff for EPIC). Google doesn't seem to find
it just now and obviously Compaq and HP websites don't have it but
(for now at least) there is still a copy at:
http://www.cs.trinity.edu/~mlewis/CSCI3294-F01/Papers/alpha_ia64.pdf

Historical interest only, obviously.