From: Gadi Oron (gadi_at_proteologics.com)
Date: Mon Dec 22 2003 - 11:30:55 CST
Hi again Alex,
<quote who="Alexandre Vakhrouchev">
> Hi again!
>
> Please tell me the reason why do you need using
> Linux+Win platforms for your calculations?
> As for me I tested some my programs on C++
> compiling in Windows and Linux. For example
> one test prog was completed under Windows within 50 seconds,
> while running under Linux it was finished under 23 seconds... and
> that is not the _limit_
Well, I have access to 11 CPUS under linux and 40 under Windows desktops
which are nearly 100% idle. So it is clear that I may gain a lot of CPU
time running NAMD on those.
> If you don't want to install Linux on your Windows hosts,
> you can use Clustermatic distribution for loading your slave
> hosts from CD.
This is an interesting option, but it has some issues:
- These are desktops, and it would be hard for me to run arround inserting
the CDs rebooting etc. I might do it through PXE, bu still, it is a
hassle.
- I could use them only during WE/nights
- NAMD (unlike PVM) does not manage disconnecting/reconnecting hosts, so
the simulation will crash whenever a client stops (correct me if I am wrong)
> Avoid using Windows (don't trust them). I'm jocking (partly))))
I am a great believer of Linux/Unix, but well, the others are not allways
like me :-))
Best regards.
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:37:15 CST