Re: Top 3 conference!
- From: eugene@xxxxxxxxxxxx (Eugene Miya)
- Date: 23 May 2006 10:39:00 -0700
In article <1148183341.453118.189040@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
jidong <zhaijidong@xxxxxxxxx> wrote:
Which three conferences is the top 3 in parallel computing area!
I would say that if you have to ask: it's not likely you will be able to
attend. This is all incumbent on the meaning of "top."
Likely the top conferences don't really have papers, are invitation only,
and are small intimate gatherings where people who do the work can focus
on their problems without the curse of management looking over their
In the US, ICPP, ICDCS, ISCA, FTCS, and SC'xy would all have you believe
they are top. Likely in their beginnings, this was true.
In turn they spawned a series of many smaller conferences on
more focused topical areas like SPAA, ASPLOS, PPoPP, etc.
While well meaning and fine on the system development side
they suffer from distance from applications. So say the
people who run the older conferences like SC. And there are European
and Asian equivalents which struggle to maintain critical mass. None is
these is especially a bad conference (ask anyone whom has had a paper
rejected). A number of problems exist from old dead wood, concentration
on incrmental improvement over innovation, a prepondence of hardware
over software, and the complaints go on.
Independent of these conferences are a small number invitation only
meetings designed to cut the issue of peer review, deal with problems
like what it takes to run a parallel computing business. The most
visible and venable is "Salishan" after the resort the US DOE and
specifically LANL, LLNL, and SNL annually. About 100 people.
You will find very little of this meeting on the web (beside's the
resort's web site). That's the DOE community for you. The US DOD also
tends to have similar meetings mostly near the DC area. I think most of
those are somewhat wastes of money (called "dog and pony" shows).
Distrust any such meeting near govt. money centers. Salishan differs
in a hands off approach (that's yet another buzz phrase from the
perspective of money managers). You can guess similar meetings get held
in other countries. Salishan differs in that its select group of researchers
(and specifcally their students) attempt to mix a balance of people in
the private sector short on ideas but long on cash with people with
ideas and short on cash and infrastructure who have not had influence
in traditional lines of research/peer review, etc. Frankly, I think
it's not bad, but it was set up in a different era. It attempts to
redefine itself every few years. It's the most public of these good
meetings. Seymour Cray attended early on, but later he would send reps
(this gives a bad sign). I kind of liken it to pre-1942 fission.
We don't have critical mass.
What I will mention now will get more vague.
Other private funded venture meetings are held by people as consultants
like say... Esther Dyson (I'm more using her Release 1.0 Conference as
an example, she's more out of the parallel computing market) who hold
meetings which cost $Ks. There is less of that these days as money
especially in academia is tight. There is also a sense (you can see
some of this in the DARPA challenges) that academia is perceived as on
"welfare." This sits well with no one. Every one is looking for that
magic bullet, but it doesn't appear to exist.
So are there really top conferences in parallel computing? Perhaps not.
You have not heard much from Bell prizes I suspect. That should be a clue.
All this world has is the Top 500 list which is not saying much.
- Top 3 conference!
- From: jidong
- Top 3 conference!
- Prev by Date: Top 3 conference!
- Next by Date: MPI development position at NEC CCRLE, Germany
- Previous by thread: Top 3 conference!
- Next by thread: MPI development position at NEC CCRLE, Germany