Re: "geekability" of IA64 (was: Alpha dead ..... etc.)
> Hi list(s),
Also, hi lists. To those reading both. please forgive my "crosspost".
> It has already slightly appeared in the discussions about Alpha's "dead" is
> - the ideologic aspect.
> Alpha, especially combined with Linux or *BSD, is for many a symbol of
> freedom and individuality. (not only for youths in puberty struggling to
> escape "the establishment" ;-)
>
> There are many alternatives to mainstream (intel - PC), like Sun, PPC, but
> Alpha seems most common to me. It has a high geek-capability (nice
> expression, isn't it), because there's much used equipment available.
> Relatively little money for a multia takes you to "the Alpha-experience" and
> for a little more you get a "big" LX-class machine or a miata. Since the
> ports of Linux and the free BSDs to Alpha, a large (geek-)community has grown,
> which uses Alpha because of its technological superiority, but I'm sure,
> also for ideological reasons.
> [... lots of stuff deleted...]
When I first had contact with UNIX in university, I wanted to have my own UNIX system
ever since then. Running on a low budged ( as most students do ), theobvious way was to
get a 286 PC with Xenix on it, which took you quite close to UNIX at the time given. That
"solution" did quite cost *a lot*, but allowed you to do all sorts of things, that you could
never do with a student's account on the university systems.
When I left university, i got my first "UNIX-based" job on the experience earned by learning
with my 286PC, which was replaced by 386/486 systems, as they appeared, running Sys VR4
at that time.
Since I always spend most of my money on computers, I eventually bought a SparcStation 10,
which was offered to me at an affordable price in '92. To me, the decision for that system
was made on two arguments:
1. Geek-factor. The SS10 was a nice desktop-MP and SunOS 4.1.3 was among the best
Unices at it's time ( IMHO, of course ).
2. SunOS/SPARC knowledge opened lots of good job-opportunities, so that the "investment"
could pay off.
I don't want to bore you with my life, but it did turn out to be a quite reasonable decision.
In 96, I got somewhat tired ( and sick ) of SPARC and Solaris and wanted to turn my focus
on Linux, of which I still believe today, that having in-depth Linux-knowledge does not
only have the obvious high-geek-factor, but also will pay off ( and does already today for me ).
At that time, the only decision left, was which platform to choose. PC's were quite fast, but
mostly crap in quality. I bought a PC164 for the same arguments, that I bought the SS10.
It was a 64bit-machine, had 64bit Linux on it, had a very high geek-factor and was a
system of excellent design ( IMHO ) compared to what was available.
Again, it paid off mostly due to the fact, that the Alpha taught me the subtle pitfalls of
porting/developing software for a 64bit platform. I was convinced, that the Alpha platform
would succeed, because of obvious technical reasons. At the time, i got my PC164 up
and running, I realized, that i had the One And Only True Unix System (TM) in front of
me:
1. Fast 64bit CPU, excellent FP performance.
2. Large memory ( 128Megs was BIG at that time )
3. SCSI Bus.
4. Complete system software in source
5. Excellent Docs ( I'll miss the Digital Techdocs )
6. Serial console and a terminal ( were talking geek here ;-)
That setup was replaced by a UP2000, a pc with multihead-X being managed by the UP2000
and a bunch of other systems, I use for various tasks. I still consider that system to have
the ultimate geek-factor from my perspective, but given the announcement of Alphas
death, the argument of learning-for-money doesn't quite pay anymore. As a consultant, I
heavily influence purchase decisions and get paid for that. Unluckily, from a business
perspective, there no real point anymore for recommending Alpha.
In consequence, I will get rid of my Alpha-equipment over the next months and switch over
to IA-64. I do not really like it, although I think, IA-64 does have some strong points and
some geek-factor. The sad thing is, that I am still *very* pleased with my Alphas ( there are
10 sitting in my room ), especially with my UP2000. But I do have to fill my fridge and therefore
follow economical constraints.
Of course, Compaq did the wrong thing in killing Alpha, and it will come back to them in a
few years. Of course, Intel did the *right* thing, because they killed the only architecture
with enough performance *and* a roadmap into the future, that really could threaten IA-64.
Unluckily, Intel will succeed with IA-64. They won, period.
I already managed to get hold of my first linux-based dual Itanium, which I expect to
receive in the next months.
I can only hope, it has a geek-factor about as high as the Alpha, which would make me
anjoy my work as much as i enjoyed it on my Alphas. But one thing is sure:
I'll miss the Alpha architecture.....
Just my $.02
Regards
Kirk
--
Thomas Weyergraf kirk@colinet.de
My Favorite IA64 Opcode-guess ( see arch/ia64/lib/memset.S )
"br.ret.spnt.few" - got back from getting beer, did not spend a lot.
Reply to: