[Beowulf] Re: vectors vs. loops

Jim Lux James.P.Lux at jpl.nasa.gov
Thu May 5 10:18:11 PDT 2005


>
>with the chickens the same money would buy.  Not to mention the risk of
>mad oxen disease...

would that not be "Mad Ox Disease"? <grin>  But Avian flu is much more 
contagious than Max Ox, so it might wipe out your flock. Perhaps 5%-10% 
would survive, being immune... so you'd still have some compute capacity, 
while your entire ox is dead.  Maybe there IS a lesson here?  genetic 
diversity in clusters is a good thing! (manifesting itself as different 
distros, different OSes, different CPU architectures).


>The list tends to be tolerant as well of discussions of big iron, of
>"oxen", of vector processors, of storage arrays, of networks, of shared
>memory designs, of message passing software, of all sorts of HPC issues
>even though the list address is NOT "hpc at hpc.org".  It has come to be
>more or less synonymous with HPC because of the enormous success of
>cluster computing economically to the extent that it now dominates HPC,
>but the two aren't really the same thing.  It tends to be LESS tolerant
>the further out there one goes away from the COTS part (especially if a
>point is made obnoxiously), though, as the bulk of us are here because
>of that mix of hardware addiction, infinite needs, and finite budgets,
>and appreciate that this list is ABOUT COTS clusters if it is about any
>single unifying factor.  It has been known to (gently) ridicule even
>COTS clusters that were built for the wrong reasons or with a silly
>design process (usually as a showcase piece for some major vendor in
>cahoots with some perhaps understandably greedy University department
>somewhere, a hammer in search of a nail and a top 10 slot on the Top 500
>list).

We also are very aware of Santayana's aphorism about history and 
repetition.  Knowing what went before is very useful. If I hadn't thrown 
away all those old copies of Datamation from the 60s and 70s, I could 
probably make a good living writing columns and articles for the cluster 
trade press by just copying and changing a few names.  "Today, Framostat 
Corporation introduced a New Batch Scheduler for OS/360, including dynamic 
job queue priority"  or "Justify the ROI on your new system with improved 
job cost accounting systems. The new system accounts for CPU seconds, 
KiloCoreSecs, and I/O channel bandwidth, allowing efficient distribution of 
total system costs among your users"

All clever and elegant architectures start to gradually get more complex 
and feature laden, until someone comes along with something radically new 
that fixes the problem.  But even technical elegance isn't it's own 
reward.  Nobody will argue that Intel's x86 CPU architectures are more 
elegant and clean than Motorola's 68K, but Moto's not making PC processors 
anymore, and Intel is. (A pox upon the segment register, and all it's ilk)

I suspect that when Dave Cutler came to Microsoft to build Win NT, he had a 
vision of a nice clean multitasking operating system like RSX-11M or 
VAX/VMS or TOPS-10/20.  But no, it had to retain compatibility with the 
CP/M heritage of MS-DOS and Windows, so that got grafted on, then this, 
that, and the other thing got rolled in (sometimes for decent technical 
reasons, sometimes as an anticompetitive measure).

And nobody will claim that, underneath it all, *nix is any less freighted 
with the baggage of days gone by than WinNT, or, for that matter, whatever 
IBM's MVS is now called.  They're all hideously complex, mind bendingly 
arcane at anything more than a superficial level, and so radically 
different in philosophy, that moving back and forth between them (which I 
have to do on occasion) is tedious and slow.  Why can't we all just program 
little Rabbit Semiconductor superZ80's running  uCOS?
(http://www.rabbitsemiconductor.com/)





James Lux, P.E.
Spacecraft Radio Frequency Subsystems Group
Flight Communications Systems Section
Jet Propulsion Laboratory, Mail Stop 161-213
4800 Oak Grove Drive
Pasadena CA 91109
tel: (818)354-2075
fax: (818)393-6875





More information about the Beowulf mailing list