[gmx-users] GROMACS in parallel on Dual CPU Linux machine

[Pratul K. Agarwal]

A lot of people have responded to my queries. Here is a 
summarized response:

1. The motherboard is ASUS A7M266-D and the system runs 
   well without any problems. (I too have had some 
   experience with Tyan Tiger MP S2460 -- they kept dropping 
   like dead flies).

2. These boxes are right next to cooling vent and have 
   plenty of ventilation (the nearby temp. is around 
   65F/19C); there is an extra system fan in addition to 
   power supply fan. I am not entirely convinced that 
   this is heating issue. A suggestion has been made that 
   this might be a communication issue. Well since the 
   communication is through the loopback device (or is it 
   not for the same host?), I am not clear what other 
   hardware might be facing heating problems. 
3. I have been able to run GAMESS, Gaussian98, AMBER (sander 
   v7) in parallel without any problems (numbers do match
   single CPU runs).
4. I found out that I have Registered Non ECC memory. I am 
   wondering if this might have anything to do with it.

*Most Importantly* I have not been able to reproduce the
errors yesterday. The systems were rebooted before these
tests. On the other hand when I noticed these errors
previously there were other jobs running on the same
machine. I will keep an eye on the issue and report anything
new.

Thanks to all who responded.