• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Setting up large folding farm

H

hacktor2

Limp Gawd
Joined
Oct 18, 2004
Messages
263
At my university, we have a computer research lab that consists of various clusters. Recently they recieved a 100 node cluster which is currently being built. Being involved in the build process, another student and I are looking to start some folding@home. I can't speak for the other students knowledge regarding folding@home, but I am a novice. I have complete permission to do this FYI.

What recommendations would you make regarding setup of the f@h client? Install a client on the master image, and deploy the image to the rest. Thus, having a client on each node?

or

Set the cluster up and have the client has a distributed process? If thats possible...

Node specs: (from memory... will update later)
dual 2.6 ghz xeon
8 gigs ram
4 x 80 hdd

Any and all recommendations will be greatly appreciated!
Thanks,
Hacktor
 
I'm not very familiar with this type of system, but if you can get a VM on each of the nodes running the linux client you will pull some impressive PPD and have very good stability. You may not need the VM part, but I figured that would allow you to keep it independent of whatever else is running on thoes nodes.

Notfreds has a wonderful linux VM that many of us run. You will just be running quite a few.

If you do it this way: "Install a client on the master image, and deploy the image to the rest. Thus, having a client on each node." it would work wonderfuly. (to my knowlege)
 
Last edited:
Someone on the FoldingForums mentioned they would compile a linux build for a cluster setup.
Although the linux image on each node would prob work the easiest.
 
The F@H workunits aren't designed to be run on clusters, so you probably wouldn't see any performance advantage from that approach compared to just running a single instance on each individual node. The easiest and most straightforward method would be to run an instance of the Linux SMP client on each one, as has already been suggested. Also, this:
avatardelta said:
Make sure your permission is in writing so it doesn't backfire on you.
Click to expand...
(enlarged for emphasis)
 
You must log in or register to reply here.
Back
Top