Microsoft Releases Computational Cluster Technical Preview Toolkit

Gerry Creager gerry at cs.tamu.edu
Wed Feb 14 00:00:37 PST 2001


I feel compelled to weigh in on this one.

First, a rant:  PLEASE turn off the HTML messaging.  Thanks.

Now:  FWIW, I know of few who dislike the "commercial" MPI tools.  A
number of us use them religiously.  Others of us have other favorites. 
There are the folks who roll their own, for a variety of reasons.  If
they get what they want, that's what they use.

As for the Microsoft contribution, while the workbench is pretty classy,
from my view, MS never quite got their classes right.  Admittedly, I
abandoned Visual C and C++ a couple of years ago, throwing up my hands
in disgust when I couldn't cleanly mix real ANSI standard classes with
MS, and I had requirements to use the "real" thing, but geez, that's the
point:  If you take a standard, modify it to suit your taste and then
tell the world that this is the unilateral new standard, I'm not likely
to follow along.

While Win2K is a marked improvement over NT4 in stability (NT4 was not
ever a clustering solution for me for that reason alone) some of the
neat user interface stuff I'd still have to face on the head is likely
to keep me away.  Call me a dinosaur:  I still do a lot with the command
line, and having a gui I can't really eliminate because someone decided
to limit my options to their preference isn't where I'm likely to go.

I share some of the concerns about the "evaluation period," as well: 
How long, how restrictive, and who to bitch to when it doesn't work as
expected.

As already has been stated:  This is a linux-centric group.  If we
wanted to run MS OS, we'd do so: few of us are incapable of making that
as a conscious decision.  For several very pragmatic reasons, I'll
likely stay with Linux for a while... and got toward FreeBSD when I find
the requiremetn to change.  I'm not likely to go back to Redmond anytime
soon.

I seem to recall a member of the MS Technology "team" wandering thru
this list a couple of years ago, seeking to debate the "wisdom" of us
working with an upstart, unproven OS, when he could provide us with
"proven" NT technology.  Tell me this:  How does this announcement and
your endorsement differ from that experience?

Respectfully,
Gerry Creager
-- 
Gerry Creager                        |      Never ascribe to Malice that
AATLT			             |      which can adequately be
Texas A&M University                 |      explained by Stupidity.
979.458.4020  (Phone)                |      -- Lazarus Long
979.847.8578  (Fax)




More information about the Beowulf mailing list