Re: MPI and Java

[rossini@blindglobe.net (A.J. Rossini)]

>>>>> "e" == exa  <Eray> writes:

    e> Well, MPI is about high performance. A '99 paper I read about
    e> Java MPI showed it to be 20 times slower than MPI in
    e> C/C++/FORTRAN so I don't think that's a good bet :)

    >> And we can safely discount the gains by Java bytecompiler
    >> technology, right :-)?
    >> 

    e> What technology! :)

Benchmarks are highly specific to the task.  Using statistical
simulations as benchmarks, we've seen comparable Java and C
performance (which I would've discounted, except it was done by one of
our better programmers; for reference, I still find Java annoying as
hell, compared to say, python...).

best,
-tony

-- 
A.J. Rossini				Rsrch. Asst. Prof. of Biostatistics
UW Biostat/Center for AIDS Research	rossini@u.washington.edu	
FHCRC/SCHARP/HIV Vaccine Trials Net	rossini@scharp.org
-------- (friday is unknown) --------
FHCRC: M--W : 206-667-7025 (fax=4812)|Voicemail is pretty sketchy
CFAR:  ??   : 206-731-3647 (fax=3694)|Email is far better than phone
UW:    Th   : 206-543-1044 (fax=3286)|Change last 4 digits of phone to FAX