[rescue] SMP on intel wasteful?

[Bjorn Ramqvist]

Francisco Javier Mesa-Martinez wrote:
> 
> >   To recap, this was when you suggested that, since a few dozen labs and
> > colleges have built big Linux clusters, rooms full of cheap PeeCees
> > will obviously replace all of the powerful computers in the world.
> 
> Well about 70 of the top500 computers belong to that category. And since I
> am in an academic setting, you would be surprised at how many more are
> coming online in the next half a year or so. Heck even Cray had to phase
> out the T3s with a cheapo cluster of PCs.

Is Cray phasing out their T3-line?
To me, that sounds like a big pile of bull-####, but it makes sense in
one way; it's old.
Oh yes, put faster processors in it, but that would defeat the overall
machine in the end. That machine (or those machines) were designed for a
specific purpose with specific processors and all that. No way they're
getting scalable juice out of that by just putting some anabolics here
and speed up there.

> The truth is that discussions like these are moot. Vector is hardly used
> anymore, and when it is usually specialized SIMD approaches are used in
> that field. Granted each approach has its niche, clusters are better for
> non coupled data problems, whereas vector/SIMD machines are better for
> data parallel problems. Once you have specified your problem, then by al
> means chose the RIGHT tool to solve it, And let's stop this pissing
> contest, please.

What does the T3 have in common with vector machines? Nothing.
It was designed by a company that makes both vector and parallel
machines. Other than that?
And, from what I know, you don't do parallel computations on a vector
machine, or the other way around. That kind of defeats its purpose...

I sure know Mr. McGuire would enlighten a few folks about this.

<snip>

> really hard to justify the order of magnitude extra cost for a recent
> workstation, when there is not even an order of magnitude of performance
> advantage over their PC counterparts, even in some cases there is a
> performance drop (case in point the SGI fuel... granted it is an
> specialized machine, and it is very good at certain tasks...)

Just as I was saying when I refered to the SGI Fuel and our problem to
justify its cost; it depends entirely on the applications. Period. If
the application companies would ever pay attention to real hardware,
there shouldn't be this kind of situation we're in. Right at the root we
have the salesdroids in their suits, manipulating big old-time companies
with its new management-suits leading it, waving with creditcards in
front of their noses.
I mean, can someone please enlighten me WHY SDRC droped support for the
Alpha platform? Yes. Suit-troubles and Mickeysoft getting greedy when
they realized that they already had SDRC down on their knees, riding the
Windoze-wave and porting their application to Windoze.

> These debates will always exist. In the 80s was the attack of the "killer
> micros", where workstation manufacturers were starting to beat the Mini
> people at their own game (heck, there are still people that swear how
> great their Pr1mes were...). And in the 70s it was the minis that were
> taking over the mainframes. And I assume pretty soon it will be the PDAs
> taking over the PCs. And so on and so forth. I guess things get smaller,
> and get cheaper... not necessarily better. And as always the best
> technology doesn't really win the "war"....

I'd agree with you by that.
These things are common everywhere, not just in the computing-industry.
Not to mention VHS vs V2000 vs $video_brand, the CDROM vs
$other_technologies, DVD-R vs DVD+R vs DVD-RAM vs $DVD_format.
Blame it on the suits.

-- 
Vnliga hlsningar/Best Regards
Bjrn Ramqvist, Hgglunds Vehicle AB