So I've finally started folding...a few questions

[holli4pirating]

I've finally started folding on server (listed in my sig).

I went through the config guide on hardfolding, and at the moment I'm running 4 instances (1 per virtual core), each set with 128 megs of RAM and running at 50% core. I might up the ram and core depending on how much it impacts the server's normal functions (it's a fileserver for about 15 people). So with that out of the way, on to the questions.

On the four instance I just started, they have different total steps (one is 2mil, the other are 4 mil). What causes that/does it matter? They should all have exactly the same configs.

How often does it update the number of steps completed? They've each been running for about 30 mins and there has been no change on the screen.

I tried searching the Hardfolding members list for myself, and even though I've created a profile, I'm not getting anything when searching for myself. Is something wrong, or does it just take a while to update?

Would I be better off running just two instances (one per physical core), or should I keep it at 4 (one per virtual core)?

If I need to reboot or shutdown, how should I close each instance of folding so as not to lose results? Is there a pause or halt command?

I think that's about all for now, but I might have more questions as time passes. Thanks for any help or insight that you might be able to provide.

 

[ICE_9]

I've finally started folding on server (listed in my sig).

I went through the config guide on hardfolding, and at the moment I'm running 4 instances (1 per virtual core), each set with 128 megs of RAM and running at 50% core. I might up the ram and core depending on how much it impacts the server's normal functions (it's a fileserver for about 15 people). So with that out of the way, on to the questions.

On the four instance I just started, they have different total steps (one is 2mil, the other are 4 mil). What causes that/does it matter? They should all have exactly the same configs.

You are assigned work units from an assignment server. You will get a range of unit sizes and speeds. That is complete out of our control what they assign us.

How often does it update the number of steps completed? They've each been running for about 30 mins and there has been no change on the screen.

This is work unit dependant. Some will go every 15 minutes, other every hour. I do recomend that you set checkpoints at 30 minutes to shave off some time. It is a server, uptime should be its middle name.

I tried searching the Hardfolding members list for myself, and even though I've created a profile, I'm not getting anything when searching for myself. Is something wrong, or does it just take a while to update?

Once you submit a work unit and hardfolding updates with you submission, you should show up on there. Welcome to the team!

Would I be better off running just two instances (one per physical core), or should I keep it at 4 (one per virtual core)?

Stanford is currently running a beta on SMP processing. It will utilize all 4 cores at 100% (or real close). The work and points are outstanding. Use this config if you plan to only do folding or light use on your server. If you want to do less, use a VM and still use the SMP work units at full on only two processors.

If I need to reboot or shutdown, how should I close each instance of folding so as not to lose results? Is there a pause or halt command?

Ctrl-C

I think that's about all for now, but I might have more questions as time passes. Thanks for any help or insight that you might be able to provide.

I hope this helps. Again welcome to the team!

 

[holli4pirating]

This is work unit dependant. Some will go every 15 minutes, other every hour. I do recomend that you set checkpoints at 30 minutes to shave off some time. It is a server, uptime should be its middle name.

I had set the checkpoints to 20 mins, which is why I thought I'd have seen something by now. Or does the checkpoint denote something else?

EDIT: It just started updating them, the first one took one hour for 20k steps.

 

[Pocatello]

Welcome to the [H]ardest Team in the World! Team 33 FTW!



Nice to have you around. Keep on Folding!

 

[ICE_9]

I had set the checkpoints to 20 mins, which is why I thought I'd have seen something by now. Or does the checkpoint denote something else?

EDIT: It just started updating them, the first one took one hour for 20k steps.

If you are running it with the -verbosity 9 flag, I think you would see when the checkpoint writes to the disk.

 

[Tigerbiten]

The checkpoint timer only really shows up when running a protien that needs an Amber core.
Then it show up in the FAHlog.txt file, otherwise it saves the data but does not write the info in the log file unless you set the vebosity flag to 9.

On a 1.6 Ghz HT Xeon you can run either ..........
2 CPU clients, you'll return the work the quickest.
Or ...............
4 CPU clients, you'll return the work slower but get around 15% more PpD.

Staford likes you to run only 2 copies as it returns results faster but knows 4 copies gives more points.
I dont think the CPU's are fast enough to run the SMP client.

Depending on the actual protiens you've got, it could take anything from about 4 days upto around 15 days to finish it.
So 1 hour a frame is not bad.

Luck ............