Lots of Computers Need a Use

A

antoniohawk

Limp Gawd
Joined
Nov 3, 2003
Messages
326
I've got a bunch of older computers that are just sitting around, because there's really no use for them. I was wondering how best to go about setting them up to fold. Should they all be networked, what OS, and what specs are really required to make even a little difference? I really don't know much about folding, but I figure that it's time to learn and make some oldies some goodies. :)
 
Well depending on speed, you can F@H on them, anything much less than 733 or 866 will probably not turn units in on deadlines, so you can fold timeless on them. You might consider running the fold-serve as it is all linux and server based.

Or depending on your desire, you can also run the [H]orde sanction UD project as it has no unit deadlines. Ultimately the choice is up to you.

I would recommend reading through the FAQ's and the project guides on www.hardfolding.com to determine what is best for the equipment you have.

 
I have a P233 folding, and it seems to manage to do half of what this thing manages to do in a day (mind you, I have it set to only use 66% of the CPU).

I run a stripped down version of Mandrake Linux on the P233. My P4 has done 2025 frames and the P233 has managed to do 1125. Although, the P4 goes off at night, and the client is off when I play games.

edit: in ~4 days time.
edit 2: Due to having to restart the P233 due to Internet on it going tits-up, It's now at 1184.
 
I would do a network boot on them, that way you can make changes to all the systems and you need them networked so they can freely turn in and get new WUs (I've done it manually, trust me it sucks!). Setup a huge list in Electron Microscope with all your machines so you can make sure they are meeting the deadlines.

Good luck with the Farm!

Fold on fellas!
 
If a system is folding 24/7 and that's all it does, it will easily make the deadlines of the normal WUs if it is a PIII class system with SSE. I had a PIII 700 system at work that ran folding 24/7 plus it was my work machine. It had no trouble at all turning in WUs on time. I would not try to run bigpackets on it, though.

For a Micrsoft OS, I would personally run either Win2k or WinXP and make sure they have at least 256 meg of RAM. Anything running 2k or XP with at least 256 meg of RAM and 1 Ghz+ speeds w/ SSE I would run bigpackets WUs on. That's how I have my Morgan core 1 Ghz Duron setup.

 
I attempted to install F@H following the guide at www.hardfolding.com, but I'm too sure that it's working. In task manager is says that the Fahcore is running at near 100%, but in EMIII I'm not making any progress. It says 0% complete and all of the running times are at 00:00:00. Maybe I just haven't started it yet or something, but how do I do that?
 
EMIII takes a while to refresh. What happens is that EMIII bases its calculations on the first step completed. So once you're boxen start completing steps, you will see the graphs and times show. Depending on the speed of the boxen this could take a fair amount of time.

 
Now I'm quite positive that there is an error because it says "No". Then when I double-click on that it says "waiting..." and resides on "Error". Anyone know what's up?
 
What shows up in the FAHlog.txt? It should be in the same folder as the executable.

If you have a bunch of computers, try out fold-server; it lets you boot the computers over network, and everything is stored on one network drive, so you don't have to have hard drives on every machine. Pretty cool.

<- fold-server developer. ;) Email me if you have questions, comments, whatever.
 
I've given up on EMIII Working properly.

As long as it shows a new line every percent (1%, 2%, 3&, 4%, etc). Should be something like this:

[12:47:36] Assembly optimizations on if available.
[12:47:36] Entering M.D.
[12:48:01] (Starting from checkpoint)
[12:48:02] Protein: p884_p53dimer884
[12:48:02]
[12:48:02] Writing local files
[12:48:05] Completed 979140 out of 2500000 steps (39)
[12:48:06] Extra SSE boost OK.
[13:48:34] Writing local files
[13:48:34] Completed 1000000 out of 2500000 steps (40)
[15:04:03] Writing local files
[15:04:03] Completed 1025000 out of 2500000 steps (41)
[16:18:49] Writing local files
[16:18:49] Completed 1050000 out of 2500000 steps (42)
[17:31:04] Writing local files
[17:31:04] Completed 1075000 out of 2500000 steps (43)
[18:44:17] Writing local files
[18:44:17] Completed 1100000 out of 2500000 steps (44)
[19:38:06] Writing local files
[19:38:06] Completed 1125000 out of 2500000 steps (45)
[20:26:39] Writing local files
[20:26:39] Completed 1150000 out of 2500000 steps (46)
[22:31:53] Writing local files
[22:31:57] Completed 1175000 out of 2500000 steps (47)
 
I second the Fold-Server, once it's set up, adding and removing boxen is so easy :)
And stats for all the machines are easy to keep track of.

-E

 
The whole file was too big to paste so here's this part. I think that it says that I've completed at least a few parts of the work unit.
[17:54:10] Proj: work/wudata_01
[17:54:10] Done: 22905 -> 142986 (decompressed 624.2 percent)
[17:54:10] nsteps: 5000000 dt: 2.000000 dt_dump: 250.000000 temperature: 355.000000
[17:54:10] xyzfile:
[17:54:10] " 393 p1112_L939_K12M_nat_min1_355K
[17:54:10] 1 N -67.710747 70.609528 87.5..."
[17:54:10] keyfile:
[17:54:10] "parameters ./proj1112.prm
[17:54:10] NOVERSION
[17:54:10] ARCHIVE
[17:54:10]
[17:54:10] cutoff 16.0
[17:54:10] taper 12..."
[17:54:10]
[17:54:10] - Couldn't get size info for dyn file: work/wudata_01.dyn
[17:54:10] Starting from initial work packet
[17:54:10]
[17:54:10] Protein: p1112_L939_K12M_nat_min1_355K
[17:54:10] - Run: 103 (Clone 15, Gen 2)
[17:54:11] - Frames Completed: 0, Remaining: 400
[17:54:11] - Dynamic steps required: 5000000
Click to expand...
 
Is that from the top or the bottom of the fahlog.txt file? The most recent stuff if appended to the bottom, and it appears according to that, that no work has been done yet, at least no full steps. If it has done a full step, you would see a line similar to this:

[13:48:34] Completed 1000000 out of 2500000 steps (40)

Also, you might want to check the unitinfo.txt file. It isn't as in depth, but it will provide a % of the work completed on the current unit.

 
Here's FAHlog.txt, from the bottom. I looked in that other file and it says that I'm at 1%, so I guess that I'm making some kind of progress.
# Windows Console Edition #####################################################
###############################################################################

Folding@Home Client Version 5.02

http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: C:\Program Files\Folding@Home
Service: C:\Program Files\Folding@Home\FAH502-Console
Arguments: -config.exe -svcstart

Launched as a service.
Entered C:\Program Files\Folding@Home to do work.

[20:22:12] - Ask before connecting: No
[20:22:12] - User name: antoniohawk (Team 33)
[20:22:12] - User ID: 47224A354BBCBEFD
[20:22:12] - Machine ID: 1
[20:22:12]
[20:22:13] Loaded queue successfully.
[20:22:13] + Benchmarking ...
[20:22:18]
[20:22:18] + Processing work unit
[20:22:18] Core required: FahCore_65.exe
[20:22:18] Core found.
[20:22:18] Working on Unit 01 [June 22 20:22:18]
[20:22:18] + Working ...
[20:22:19] Folding@Home Client Core Version 2.53 (June 29, 2004)
[20:22:19]
[20:22:19] Proj: work/wudata_01
[20:22:20] Done: 22905 -> 142986 (decompressed 624.2 percent)
[20:22:20] nsteps: 5000000 dt: 2.000000 dt_dump: 250.000000 temperature: 355.000000
[20:22:20] xyzfile:
[20:22:20] " 393 p1112_L939_K12M_nat_min1_355K
[20:22:20] 1 N -67.710747 70.609528 87.5..."
[20:22:20] keyfile:
[20:22:20] "parameters ./proj1112.prm
[20:22:20] NOVERSION
[20:22:20] ARCHIVE
[20:22:20]
[20:22:20] cutoff 16.0
[20:22:20] taper 12..."
[20:22:20]
[20:22:20] Hashes matched on file work/wudata_01.dyn
[20:22:20] ARC file integrity verified
[20:22:20] Restarting from checkpointed files.
[20:22:20]
[20:22:20] Protein: p1112_L939_K12M_nat_min1_355K
[20:22:20] - Run: 103 (Clone 15, Gen 2)
[20:22:20] - Frames Completed: 4, Remaining: 396
[20:22:20] - Dynamic steps required: 4950000
[20:22:20]
[20:22:20] Writing local files:
[20:22:20]
[20:22:20] parameters work/wudata_01.prm
[20:22:20] - Writing "work/wudata_01.key": (overwrite) successful.
[20:22:20] - Writing "work/wudata_01.xyz": (overwrite) successful.
[20:22:22] - Writing "work/wudata_01.prm": (overwrite) successful.
[20:22:24] - Writing "work/wudata_01.key": (append) successful.
[20:22:25]
[20:22:25] PROJECT="work/wudata_01", NSTEPS=4950000, DT=2.0000, DTDUMP=25.000000, TEMP=355.00
[20:22:25] TINKER: Software Tools for Molecular Design
[20:22:25] Version 3.8 October 2000
[20:22:25] Copyright (c) Jay William Ponder 1990-2000
[20:22:25] portions Copyright (c) Michael Shirts 2001
[20:22:25] portions Copyright (c) Vijay S Pande 2001
[21:57:22] Finished a frame (5)
Click to expand...

I just noticed too, that I'm at 48% of a frame in EMIII, so I guess that that is now working also.
 
Yep. :D

Looks good, EMIII just takes some time after a client reboot, or restarting as it bases its numbers on that first frame completed. What you'll come to notice after awhile, is that it can take a fair amount of data completed before times will stabilize and such.

 
You must log in or register to reply here.
Back
Top