[Bioclusters] Re: How to build a biocluster Part 2 (hardware)

[Ron Chen]

> ->Interface card: Yeah, Myrinet would be great.
> Back to Earth: we are thinking either 100Base-T or
> Gigabit Ethernet (most probably the first one.

It depends which kind of BLAST you want to run --
-> 1 large parallel, or
-> many small serial

In practice, small serial jobs provide a better
throughput than just the large 1. This is mainly due
to the communication time needed between processes
for the large, parallel job.

A fast network card will help in both cases, since
data will travel faster from the file server to the
execution nodes. However, if you go with the "many
small serial jobs" model, then your network is mainly
for sending data to/from the execution nodes, but not
for job execution.

-Ron

__________________________________________________
Do You Yahoo!?
Yahoo! Health - your guide to health and wellness
http://health.yahoo.com