hello LAM-MPI users,
after working two (2) full years in the industry, i have decided to take
my master degree at a local university here (obvious from the email sig below
ain't it). i am not required to think about my master's thesis yet as
i am still starting.
however, i feel that i *should* be scouting around for ideas. i would
like to do parallel computing projects again. however, i would like tofocus
on the tools instead of the computational science. (but, i have nothing against
the computation science. i just feel i would have more fun doing tools ;-) )
my particular strengths are mainly Linux/C programming. i do a bit
of open source software development mainly for the flow-tools project. so
i feel i can contribute more by building up tools. i was just wondering if
there are any nice things to stick into LAM-MPI that need to be built?
one idea i had in mind is reviving the old mpi-enabled make. i would probably
just hack functionality into the new GNU make to make it more sustainable.
.. OR ..
another idea would be build a job submission system. abiet, i am not sure
how to make it simpler or more usable than PBS or the Sungrid engine. or better
than Mosix for that matter.
.. OR ..
make some extensions to natively add functionality for some grid standard thing.
i have not done my reading here yet. but, is this a worthy direction to go or
is it just overhyped?
.. AND many others ..
now my problem. i do have some people to ask here but theses people are not
really into HPC tools development. just feel they are out of their element.
any advise? thanks.
-------------------------------------------------------
William Emmanuel S. Yu
Department of Information Systems and Computer Science
Ateneo de Manila University
email : wyu at ateneo dot edu
web : http://CNG.ateneo.net/cng/wyu/
phone : +63(2)4266001-4186
GPG : http://CNG.ateneo.net/cng/wyu/wyy.pgp
|
