16 Jul
2006
16 Jul
'06
9:15 p.m.
Just a few fixups on Jay's report... (and a few things to add/clarify)
On Sunday 16 July 2006 22:42, Jay West wrote:
Friday night a large group of us met for some awesome
Indian food at a
buffet, then got a tour of the Purdue datacenter. The IBM pSeries 1600
supercomputer cluster was the high point of my own walk through (even
though it runs AIX ;) ). I have to admit, I think most of the million+
dollar pricetag is for that cute gold stripe down the front ;)
(The "gold" -- really copper -- stripe was on our two p690 16x POWER4
Regattas. The "cluster 1600" or SP still has nice sexy black racks,
though. :)
Watching the
Adic tape library robot come screaming down the racks and stopping about 8
inches from your face on the other side of the glass was rather titilating
as well. I was rather disappointed in that I couldn't find a single
"emergency pull" button to hit *sigh*. Although... there was this rabbit
trap on the wall with a wooden stick holding up the.... (Pat, someone
should write on the front of that panel "shhhhh be vewy vewy quiet... I'm
hunting wabbits").
... holding up a 400A 208V 3phase contactor, which *should* be fixed but no
one is willing to test...
After the tour we attempted to go see Pat's own
personal
datacenter in downtown Lafayette, but the police who cordoned off a couple
block radius around his datacenter had other ideas. They weren't talking,
but we were able to come back the next day. I did get pictures of some of
the datacenter (Purdue's) which I'll post if Pat says it's ok.
Saturday it took quite a few hours to get everyones stuff from the loading
docs to the exhibit room. There were around five flattrucks and many people
working them continuously for a few hours. I liked all the displays, but I
think my own favorite was the "Geek Museum" due to the gamer in me. I did
Mike Lee did a wonderful job with his exhibit, and probably could have used
another table or three to spread everything out on.
manage to take a lot of pictures of the exhibit hall
during the show. I
know there was some concern about if there would be any speakers lined up
for the show. Let me tell you - it was worth anyones plane ticket price and
then some to see/hear the two speakers (George Goble and Mike Marsh) talk
about their creation of the dual processor Vax 11/780. That was pure,
unadulterated Classic Computer heaven. I think they went slightly over the
time allowed, and I was hoping they'd still keep right on going. I am
hoping that Pat can make that tape available for purchase - if so, I will
be buying a copy and I'd strongly encourage others to as well. It was chock
I'll see how the recording came out. George sat somewhat off-camera and
off-mic for most of it, but it might still come out ok. In retrospect, I
should have set them up like a "panel" like we did for the 2nd set of
speakers..
full of exactly the type of fascinating tidbits that
we all love, like why
they had to set the drives to seek when idle to avoid a headcrash. Or how
to tell of an impending "drives gone critical" (no relation to the late
night video advertisement) by the strobe pattern effect of the top surface.
There was also some fascinating discussion of some of the problems they had
with certain instructions failing when they resided in the last few bytes
of a page. For the next "speaker", several of us (myself, Tom Uban, and Dan
Cohoe) were recruited to sit on a discussion panel with Jack Rubin just a
couple hours before it was to start. It was hard to tell for sure from my
view, but I think this actually came off fairly good - mainly due to great
points by Tom, Dan, and Jack rather than myself :) After the show we spent
many hours carting off everyones gear to the loading doc (you should see
the pictures of it all stacked up waiting for loading). Then a large group
of us went to the same Italian restraunt as last year and it was just as
good again (by the way, George Goble went with us and was still fascinating
to listen to). Good food and beer was enjoyed :)
Nit, actually Mike Marsh was the one that trailed us to the restraunt, to
my "data center" and back for the loading fun later...
After leaving the restraunt we headed to see Pats own
"slightly crowded"
datacenter. Let's just say that it's a good thing his building doesn't have
a basement. Anything other than cement slab over bedrock would surely give
way at the plethora of mainframe gear. Pat, that microwave looks sorely out
I actually chose the place *because of* the cement slab floor, which is level
with the sidewalk outside.
of place. Actually, it looks very scared.... Of course
I enjoyed this part
It happened to be in one of my 10x10 storage units that I cleared out into the
place.
of the evening as at this point I consumated a trade
with pat that allowed
me to put some new (to me) DEC gear on my trailer. Yup, my trailer was
present. As was a fire extinguisher (J/K).
It's a shame we couldn't get you some IBM SP gear on that trailer. ;)
My favorite part of the trip - assisting around five
other guys with
loading Dan C's newest acquisition in to his minivan. An IBM RS/6000. And
not just "any" RS/6000. It was an SP machine. Think Cray (as to appearance
{and size/weight}). Yes, I did say minivan. I don't think I've ever laughed
For the record, I'd estimate the weight of a "fully configured" rack to be
around 2000lbs, 1800lbs if you take the steel plates out of the bottom of the
rack. Fortunately, we weren't putting *exactly* that into Dan's minivan.
quite so hard as seeing the precise way that the
forklift moving forward
and van moving backward was choreographed to get the machine, base
pedestal, and two of the four processing nodes in that minivan. Along with
a lawn mower (don't ask). I don't recall what time we started loading this
machine, I am guessing around 10:30 or so. I do know that it took until
around 2am. I have lots of pictures of this operation, including a side
view of the minivan's wheel well (or lack thereof). I belive I did hear a
leafspring cry out. Thank god CarFax doesn't track that kind of information
or Dan may have trouble selling the vehicle some day ;)
With only 2 nodes, he probably got *only* about 1150lbs of IBM gear. ;)
The next morning a few various groups met at separate
places for breakfast
or to complete personal trades. I myself left around 10am to head back to
St. Louis. Oh, by the way.... folks looking for DEC gear that didn't show
up to VCF/M missed out - Paul Anderson was there with a large amount of
great gear for sale. I was like a kid in a candy store with the M module
card inventory, printsets, documentation, tapes, and even a nice RK05
drive.
I'd agree. I had a module that Paul couldn't find in time to bring to the
show with him, but he had a lot great looking things that I would have
considered buying if I had a use for them.
Overall, I think VCF/M v2.0 was a great success. That
being said, I humbly
submit that we can still grow it further and be better still next year. I
want to publicly thank (and commend) Pat for pulling the event together and
making it happen. He deserves a round of applause and thanks! I (and I know
for sure others) are already really looking forward to VCF/M v3.0.
As always, it looks a lot better now than it did the week before (when I was
working my arse off to get things prepared for the show)..
Jay West
PS - Pat, I really did like the new VCF/Midwest v2.0 T-shirt. Don't forget
to send me an "L". Where do I sent the money?
I made up a cafepress store to sell shirts, anyone can order them from here:
http://www.cafepress.com/vcfmw2/
Pat
--
Purdue University ITAP/RCAC --- http://www.rcac.purdue.edu/
The Computer Refuge --- http://computer-refuge.org