[gmx-users] Performance of beowulf cluster

[Mirco Wahab]

On 05.08.2014 07:01, Abhishek Acharya wrote:
> I am planning on investing in a beowulf cluster with 6 nodes (48 cores) each with AMD Fx 8350 processor, 8 GB memory  connected by 1 Gigabit Ethernet switch. Although I plan to add more cores to this cluster later on, what is the max performance expected from the current specs for a 100,000 atom simulation box ? Also, is it better to invest in a  single 48 core server ? The cluster system can be set up at almost half the price of a 48 core server, but do we lose out on performance in the process?

6 AMD-8350 boxes, connected to *one* 1GB switch?

This system could be put to very good use if you are
able to perform 6 *independent simulations* on your
molecular system.

100,000 Atoms is a rather small system for "large scale"
parallelization. A 100K SPC box would have an edge length
of about 10nm?

If it's important for you to have parallel runs on single
molecular systems, you could consider a dual-socket-2011
system running  6-core i7 processors (i7/4930K or upcoming
Haswell-E 5930K) combined with quad-channel DDR3/4.
This would give you a 24x parallelization on a single
workstation.

What about modern (Nvidia) consumer graphics cards? These are
supported very well by Gromacs.

Regards

M.