[Beowulf] SC13 wrapup, please post your own
prentice.bisbal at rutgers.edu
Mon Nov 25 15:03:18 PST 2013
Okay, here's my attempt at summarizing my SC13 experience.
1. The show was actually kind of boring this year. I think this was
because there were now big new technologies or products to talk about.
Also, there's usually a lot of activity at the NOAA and OLCF booths
every years, and neither of those booths were there this year, which
could contribute to that. Neither were there last year, but there were
several industry announcements to make up for that. The DoE had one
booth for all the DoE labs, and there really wasn't much activity there
this year. Someone mentioned that the exihibitor floor was smaller than
usual. In hindsight, I have to agree.
2. What I think was the biggest announcement of the show went unnoticed
by most people: IBM and NVIDIA are teaming up to put a Power8 processor
and NVIDIA GPU together. Thinks of 'Project Denver' but with a Power8
processor instead of an ARM. I think this is a first move towards an
architecture for an Exascale system. Blue Genes and GPU clusters have
been taking turns at the top of the Green500 list. ARM processors much
more energy efficient than either, but don't have the floating-point
performance. If these Power8 processors are as power-efficient as the
processors in the Blue Gene products, then this (I think) is the best
chance at reaching the 20 MW limit cited for an exascale system. I don't
know much about the Power8 processor, so please don't skewer me if I'm
completely wrong about it being very energy-efficient.
3. Once again, NVidia's schooled everyone else on how to have a busy
booth that gets everyone's attention. Best location (on the corner of
two busy walkways, as usual), large screen facing passers-by, and great
audio making it easy to hear the presentations without being annoyingly
loud. And they're green scarves are the best give-away of the conference
: their practical, have an air of exclusivity which makes everyone want
to wear theirs, turning many attendees into walking advertisements for
NVidia. Why more vendors don't take lessons from NVidia is beyond me.
Seriously. When it comes to marketing at SC13, Nvidia makes everyone
else look like a bunch of chumps.
4. I went to a BoF on ROI on HPC investment. All the presentations in
the BoF frustrated me. Not because they were poorly done, but because
they tried to measure the value of a cluster by number of papers
published that used that HPC resource. I think that's a crappy, crappy
metric, but haven't been able to come up with a better one myself yet. I
was very vocal with my comments and criticisms of the presentations, so
if any of the presenters are reading this now, I apologize for
hi-jacking your BoF. Getting good ROI on a cluster is close to my heart,
but is also difficult to quantify and measure. I hope I can be part of
the discussion next year.
5. There were a LOT of people at the D-Wave quantum computer presentation.
6. The NSA had the balls to have a booth there, looking to recruit
people. My own politics aside, I though they were crazy. I know they're
there every year, but scientists are known for liking openness, after
the Edward Snowden revelations and the fact that the US govt. banned
chinese scientists from a NASA conference (which led to boycotts of the
conference), I would to keep a low-profile around scientists if I was
part of the US Security/Defence complex.
7. The cover band 'London Calling' played the IBM Platform
Computing/Intel party again. Despite calling themselves 'London Calling'
they still do not play any Clash songs. They are a good cover band, but
it's starting to get boring seeing the same band play the same set year
More information about the Beowulf