CCL:G: Parallelization efficiency of Gaussian03

Hello Agalya.

Well, I am working on  a parallel version of Gaussian03, this is on a cluster whith a Linux operative system. the cluster have various nodes, each one whith 4 processors, for my experience with the cluster I have seen when a job is carried out with more than tree nodes, the fail probability is increased.  so currently, I am working whith tree nodes maximum (12 procesors).  under this conditions Gaussian03 runs very good.

I hope that this small comment can help you in any thing .
if there are any writing mistake, please excuse me, my English is not good.


"Agalya Govindasamy agalya81]![gmail.com" <owner-chemistry+*+ccl.net> wrote:

Sent to CCL by: "Agalya Govindasamy" [agalya81|,|gmail.com]
Hello CCL members,

I have a query regarding the parallelization efficiency of Gaussian03 compared to other software packages such as VASP, MOLPRO, and NWChem. I would like to know which of these software have good efficiency in parallel processing, particularly I would appreciate if anyone could give me information on how well gaussian03 runs in parallel processors. I know that Gaussian03 has its own parallelizing package, Linda. But I would like to learn more about the parallelization efficiency of gaussian03.

Thanking you in advance,

Agalya







Never miss a thing. Make Yahoo your homepage.