New "folder"

nonstop301

Hello everyone,

I have been interested in Folding at Home and happily joined Team 93 after visiting the forums here regularly and finding lots of useful information both with respect to the Folding at Home project and other more general computer issues.

I have only started using the folding software for a day, so I would like to ask a few things about what to expect and how to maintain an efficient operation.

Currently the program seems to be halfway through it's first task (completed 250,000 of the 500,000 steps necessary to finish the frame) and the expected time it will finish this task is sometime tomorrow in the evening.

Is this a good rate of performance and will all the tasks it will engage in require approximately the same number of steps ?

Thanks for your time and I'm looking forward to your replies.

csimon

Hey nonstop. One thing you can do is to ensure that you are utilizing the optimizations. You can do this by opening the fahlog.txt with notepad or any editor and look for the first line of production ...you should see "Extra SSE boost OK!" or something to that effect.

Might look something like this ...all you're really looking for is what I have highlighted in bold print.


--- Opening Log file [December 13 16:15:41]


# Windows Console Edition #####################################################
###############################################################################

Folding@Home Client Version 5.04beta

http://folding.stanford.edu

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

Launch directory: C:\Program Files\f@h\f@h1
Executable: C:\Program Files\f@h\f@h1\FAH504-Console.exe
Arguments: -advmethods

[16:15:41] - Ask before connecting: No
[16:15:41] - Use IE connection settings: Yes
[16:15:41] - User name: csimon (Team 93)
[16:15:41] - User ID: 17BCBE2B6AFCACC1
[16:15:41] - Machine ID: 1
[16:15:41]
[16:15:41] Loaded queue successfully.
[16:15:41] + Benchmarking ...
[16:15:44]
[16:15:44] + Processing work unit
[16:15:44] Core required: FahCore_78.exe
[16:15:44] Core found.
[16:15:44] Working on Unit 04 [December 13 16:15:44]
[16:15:44] + Working ...
[16:15:44]
[16:15:44] *

---

*
[16:15:44] Folding@Home Gromacs Core
[16:15:44] Version 1.90 (March 8, 2006)
[16:15:44]
[16:15:44] Preparing to commence simulation
[16:15:44] - Looking at optimizations...
[16:15:44] - Files status OK
[16:15:44] - Expanded 78910 -> 420497 (decompressed 532.8 percent)
[16:15:44]
[16:15:44] Project: 2125 (Run 237, Clone 10, Gen 2)
[16:15:44]
[16:15:44] Assembly optimizations on if available.
[16:15:44] Entering M.D.
[16:16:05] (Starting from checkpoint)
[16:16:05] Protein: p2125_lambda_5way_melt_4_10011
[16:16:05]
[16:16:05] Writing local files
[16:16:05] Using table 1
[16:16:05] Completed 12440310 out of 20000000 steps (62)
[16:16:05] Extra SSE boost OK.
[17:04:45] Writing local files
[17:04:45] Completed 12600000 out of 20000000 steps (63)
[18:04:56] Writing local files
[18:04:56] Completed 12800000 out of 20000000 steps (64)

nonstop301

Thanks for your reply csimon.

I had a look at the log but there was no mention of the SSE boost you're referring to.
I also get several checkpoint notifications as it going along.

Here's a small sample from the log file:

--- Opening Log file [December 13 18:48:47]


# Windows Console Edition #####################################################
###############################################################################

Folding@Home Client Version 5.04beta

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

Launch directory: C:\Program Files\Folding@Home
Executable: C:\TEMP\FAH504-Console.exe


[18:48:47] - Ask before connecting: Yes
[18:48:47] - User name: nonstop301 (Team 93)
[18:48:47] - User ID: 7CA3ACE8038E3609
[18:48:47] - Machine ID: 1
[18:48:47]
[18:48:48] Loaded queue successfully.
[18:48:48] + Benchmarking ...
[18:48:50]
[18:48:50] + Processing work unit
[18:48:50] Core required: FahCore_82.exe
[18:48:50] Core found.
[18:48:50] Working on Unit 01 [December 13 18:48:50]
[18:48:50] + Working ...
[18:48:50]
[18:48:50] *

---

*
[18:48:50] Folding@Home PMD Core
[18:48:50] Version 1.03 (September 7, 2005)
[18:48:50]
[18:48:50] Preparing to commence simulation
[18:48:50] - Ensuring status. Please wait.
[18:49:07] - Looking at optimizations...
[18:49:07] - Working with standard loops on this execution.
[18:49:07] - Previous termination of core was improper.
[18:49:07] - Going to use standard loops.
[18:49:07] - Files status OK
[18:49:07] - Expanded 82284 -> 557656 (decompressed 677.7 percent)
[18:49:07]
[18:49:07] Project: 1814 (Run 1, Clone 576, Gen 52)
[18:49:07]
[18:49:07] Entering M.D.
[18:51:07] Protein: p1814_Collagen_POG10more_refolding
[18:51:07]
[18:51:07] Completed 178200 out of 500000 steps (35)
[18:57:00] Writing local files
[18:57:00] Completed 180000 out of 500000 steps (36)
[19:04:12] Writing checkpoint files
[19:13:33] Writing local files
[19:13:33] Completed 185000 out of 500000 steps (37)
[19:19:16] Writing checkpoint files
[19:29:48] Writing local files
[19:29:48] Completed 190000 out of 500000 steps (38)
[19:34:18] Writing checkpoint files
[19:47:51] Writing local files
[19:47:51] Completed 195000 out of 500000 steps (39)
[19:49:19] Writing checkpoint files

Folding@Home Client Shutdown.

csimon

Try restarting the client and recheck.

[18:49:07] - Looking at optimizations...
[18:49:07] - Working with standard loops on this execution.
[18:49:07] - Previous termination of core was improper.

SPIKE09

2 points nonstop 1 why is the folding client in your temp file danger danger and how did you close the client last time, as that is the reason for the using standard loop's message. Unless you had a system crash looks like you did not close the client with control c. And if you clear your cache you may lose all folding as it will clear your temp file's.

nonstop301

Hi again csimon and spike

I started it once again earlier and it seems to be going well now.
I moved the console.exe to the correct location now too
I used the -advmethods flag when I launched it so I hope that will keep it going without any glitches.
There was a system crash in that instance spike but the cache wasn't cleared so I resumed from where it was interrupted.
How frequent should those checkpoints be in your opinion ?
I get one after every 5000 steps it completes in this task. I don't think it's hampering the overall process though, so I'll leave them as they are.

SPIKE09

Checkpoint's are your friend they allow the WU to restart from saved data when you use control c to halt the client. And on another point why would you halt the client:) Oh yes and you have picked a top team to join here.

Oriane

nonstop301 wrote:

Hi again csimon and spike
... How frequent should those checkpoints be in your opinion ?
I get one after every 5000 steps it completes in this task. I don't think it's hampering the overall process though, so I'll leave them as they are...

That is OK. There is only limited control you have over these anyhow- as described by Stanford Folding@Home here:

Periodically, the core writes data to your hard disk so that if you stop the client, it can resume processing that WU from some point other than the very beginning. With Tinker, that happens at the end of every frame. With GAH it happens at the end of each sequence.

As proteins become more complex and run longer, it is better to have more frames in a WU so that you don't loose so much progress if you have to restart - - hence WUs that have 400 frames instead of 100. That still doesn't take the speed of the machine into account. A fast machine completes a frame in a few minutes while a slow one may take hours, and the donor with the slow machine still doesn't want to lose 99% of those "hours" yet the fast machine doesn't really want the overhead of writing the checkpoints every "few minutes" - - and neither of them wants the upload time associated with results containing many frames.

With Gromacs, these checkpoints can happen almost anywhere and they are not tied to the data recorded in the results. Initially, this was set to every 1% of a WU (like 100 frames in Tinker) and then a timed checkpoint was added every 15 minutes, so that on a slow machine, you never loose more that 15 minutes work.

In the 4.x (and later) version of the client, you can set the 15 minute default to another value.

This is really a pretty good read if you're interested.

edcentric

I always set the check point at 30min. Why waste time saving since my machines fold 24/7

nonstop301

Thank you for replies and I appreciate the assistance you are providing.

This is how the console is currently configured and it looks like it's running well and I can keep it up all day now without any crashes.

I'm just using the 5 MB assignments at the moment since I've only started this for a couple of days and I'm evaluating how to run it without any problems.

User name [nonstop301]?
Team Number [93]?
Launch automatically at machine startup, installing this as a
service (yes/no) [no]?
Ask before fetching/sending work (no/yes) [yes]?
Use Internet Explorer Settings (no/yes) [yes]?
Use proxy (yes/no) [no]?
Allow receipt of work assignments and return of work results greater than
5MB in size (such work units may have large memory demands) (no/yes) [no]?
Change advanced options (yes/no) [no]? yes
Core Priority (idle/low) [idle]?
CPU usage requested (5-100) [100]?
Disable highly optimized assembly code (no/yes) [no]?
Pause if battery power is being used (useful for laptops) (no/yes) [no]?
Interval, in minutes, between checkpoints (3-30) [15]?
Memory, in MB, to indicate (511 available) [511]?
Request work units without deadlines (no/yes) [yes]?
Minimize networking -- get and submit work in batches (no/yes) [no]?
Ignore any deadline information (mainly useful if
system clock frequently has errors) (no/yes) [no]?
Machine ID (1-8) [1]?

If you think there's something there that should be modified please let me know.

Thanks again for all your help.

profdlp

Launch automatically at machine startup, installing this as a service (yes/no) [no]?

A service start means that you won't have to manually start it should you reboot the computer.

Ask before fetching/sending work (no/yes) [yes]?

Unless you're on dialup, I'd set this to "no".

Use Internet Explorer Settings (no/yes) [yes]?

I'd go "no" here, too, especially if you've upgraded to IE7.

Allow receipt of work assignments and return of work results greater than 5MB in size (such work units may have large memory demands) (no/yes) [no]?

It looks like you have 512MB of RAM, which is right on the borderline between going for the big Work Units or not. You might give them a try once your computer has knocked off a few of the smaller ones, just to see how they do.

Request work units without deadlines (no/yes) [yes]?

I'd change this to "no". The deadlineless WU's are scarce at many times, meaning that your computer will just be sitting there doing nothing.

Glad to have you on the team.

Shorty

Huzzah another UKer! Welcome in chap

nonstop301

Hello to both profdlp and shorty

Thank you for the help with the configuration settings profdlp and it's my pleasure to participate in Folding at Home with Team 93.

It's still very early stages for me, but the potential is always there for a respectable contribution to the cause of this important project in the long run

nonstop301

Hello to everyone again,

Thanks to your all your good tips, the Folding at Home program has been running well and completed it's first assignment.

Initially I had my doubts whether it would complete it correctly because there were many interruptions but it came through in the end and now has moved on to another task.

I noticed that after completing the first task and sending off the results, it automatically downloaded a new core too and then started off again. It is the Gromacs core (core 78) it is using now whereas previously it was operating with the PMD core (core 82).

Should I stop and start another task with the PMD core or should I leave the current assignment to complete?
It has only done 4% of it at the moment and it's due to complete within 6-7 days.

Thanks once again for all your helpful comments.

SPIKE09

No it can only run one instance per physical core for maximum efficeincy on the standard client.

Trogan

Hi nonstop301!

Nice to see someone else from London posting here.

csimon

nonstop301 wrote:

Hello to everyone again,

Thanks to your all your good tips, the Folding at Home program has been running well and completed it's first assignment.

Initially I had my doubts whether it would complete it correctly because there were many interruptions but it came through in the end and now has moved on to another task.

I noticed that after completing the first task and sending off the results, it automatically downloaded a new core too and then started off again. It is the Gromacs core (core 78) it is using now whereas previously it was operating with the PMD core (core 82).

Should I stop and start another task with the PMD core or should I leave the current assignment to complete?
It has only done 4% of it at the moment and it's due to complete within 6-7 days.

Thanks once again for all your helpful comments.

Leave it most definitely. The client will download new cores from time to time.

nonstop301

Hello csimon

I will continue with this task and just allow the program to choose the assignments on its own once it completes a job.

Sure enough now that it started this task for the Gromacs core, it also initiated the SSE boost that you mentioned to me in a previous thread.

It seems that all is going well now and it can run without any faults throughout the day

nonstop301

Trogan wrote:

Nice to see someone else from London posting here.

Greetings to you as well Trogan

I'll have to disappoint you however in terms of the soccer team I support

"Anfield glory and Pride in the Five" is what I'm about in that respect

csimon

nonstop301 wrote:

Greetings to you as well Trogan

I'll have to disappoint you however in terms of the soccer team I support

"Anfield glory and Pride in the Five" is what I'm about in that respect

Not all wu's will utilize SSE but all gromacs do. Gouble gromacs (dgromacs) will utilize SSE2 instead and Tinkers will utilize 3DNow! and yet some may use no optimization at all.

Also ...the -advmethods flag will request public bete wu's while no flags at all will request public wu's only. I mention this because if there ever comes a time when the server is not sending you work you can reverse the request and initially request from a different server. It can be quite handy in that respect.