Performance=0? help with SLOW folding

[ibilly]

well, here's my terminal feedback (below). I was wondering if I was doing something wrong: people speak of 10 min frames, and I gaet an hour and tem min a frame. If you were going to say it is your hardware, I ma running folding on a 12" PB
(Rev A)... I seem to get this perfomance (or close to it) on all of our PB's... any recomendations/diagnosis(es)?? it's bugging me and i was wondering what i could do to speed the folding. here's the log:

Last login: Fri Oct 24 23:49:59 on ttyp3
Welcome to Darwin!
[artemis:~] billy% /Users/billy/Desktop/OSX-3.25 j

Note: Please read the license agreement (OSX-3.25 -license). Further
use of this software requires that you have read and accepted this agreement.

Bad argument: j
Bad Argument(s) received

Folding@home Version 3.25
Usage: OSX-3.25 [-mode]

modes:
-config Configure user information
-queueinfo Get information on queued work units
-delete x Delete item #x from work queue
-send x Send result #x to server. Use x=all to send all results
-verbosity x Sets the output level, from 1 to 9 (max). The default is 3
-pause Pause after finishing & trying to send current unit
-forceasm Force core assembly optimizations to be used if available
-advmethods Use new advanced scientific menthods (beta)

Please note the above recognized flags and try again.
Press any key to exit.

Folding@home Client Shutdown.
[artemis:~] billy% /Users/billy/Desktop/OSX-3.25 -forceasm

Note: Please read the license agreement (OSX-3.25 -license). Further
use of this software requires that you have read and accepted this agreement.



--- Opening Log file [October 25 05:50:25]


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

Folding@home Client Version 3.25

http://foldingathome.stanford.edu
email:help@foldingathome.stanford.edu

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

Arguments: -forceasm

[05:50:25] - Ask before connecting: No
[05:50:25] - User name: billy@flanigan.us (Team 34051)
[05:50:25] - User ID = 7CB1FF41248D1E49
[05:50:25] - Machine ID: 1
[05:50:25]
[05:50:25] Loaded queue successfully.
[05:50:25] + Benchmarking ...
[05:50:34]
[05:50:34] + Processing work unit
[05:50:34] Core required: FahCore_65.exe
[05:50:34] Core found.
[05:50:34] Working on Unit 04 [October 25 05:50:34]
[05:50:34] + Working ...
Warning: Ignoring unknown arg
[05:50:34] Folding@Home Client Core Version 2.47 (June 14, 2002)
[05:50:34]
[05:50:34] Proj: work/wudata_04
[05:50:35] Done: 22962 -> 142970 (decompressed 622.6 percent)
[05:50:35] nsteps: 5000000 dt: 2.000000 dt_dump: 250.000000 temperature: 298.000000
[05:50:35] xyzfile:
[05:50:35] " 393 p692_L939_K12M_int
[05:50:35] 1 N 67.061705 29.338506 85.308336 20..."
[05:50:35] keyfile:
[05:50:35] "parameters ./proj692.prm
[05:50:35] NOVERSION
[05:50:35] ARCHIVE
[05:50:35]
[05:50:35] cutoff 16.0
[05:50:35] taper 12...."
[05:50:35]
[05:50:35] Hashes matched on file work/wudata_04.dyn
[05:50:35] ARC file integrity verified
[05:50:35] Restarting from checkpointed files.
[05:50:35]
[05:50:35] Protein: p692_L939_K12M_int
[05:50:35] - Run: 16777216 (Clone 234881024, Gen 0)
[05:50:35] - Frames Completed: 7, Remaining: 393
[05:50:35] - Dynamic steps required: 4912500
[05:50:35]
[05:50:35] Writing local files:
[05:50:35]
[05:50:35] parameters work/wudata_04.prm
[05:50:35] - Writing "work/wudata_04.key": (overwrite) successful.
[05:50:38] - Writing "work/wudata_04.xyz": (overwrite) successful.
[05:50:38] - Writing "work/wudata_04.prm": (overwrite) successful.
[05:50:41] - Writing "work/wudata_04.key": (append) successful.
[05:50:41]
[05:50:41] PROJECT="work/wudata_04", NSTEPS=4912500, DT=2.0000, DTDUMP=25.000000, TEMP=298.00
[05:50:42] TINKER: Software Tools for Molecular Design
[05:50:42] Version 3.8 October 2000
[05:50:42] Copyright (c) Jay William Ponder 1990-2000
[05:50:42] portions Copyright (c) Michael Shirts 2001
[05:50:42] portions Copyright (c) Vijay S Pande 2001
[06:51:42] Finished a frame (8)
[07:57:23] Finished a frame (9)

 

[Plutronics]

My B&W #1 G3 450MHz is getting 50 Min. a Frame and B&W #2 G4 450 MHZ is getting 49 Min a frame with the p692 WU. This is a Tinker WU and they don't use the Altivec code available to a G4. so -forceasm won't help on a p692, but won't hurt either. If you use the :
-forceasm < Force core assembly optimizations to be used if available and >
-advmethods < Use new advanced scientific menthods (beta)>
Flags , You'll get more of the Gromac WU that run faster on a G4. For Ten minute frames, try using a 2.7 GHz Pentium 4 ! You couldn't pay me to use one ( well, I'd be happy if you tried! ) . Window boxes make up 95 % of the folding Rigs , so stanford only puts 2.5 % effort into making the Macs run as fast as the could. I'm shure a 200 % increase is possible, but what can one Do ? Panther should speed things up a Bit !!

 

[britboy]

Originally posted by Plutronics
Panther should speed things up a Bit !!

I shouldn't think that would be the case. F@H relies on the coding skills of the guys at stanford, and the speed of your processor. The OS isn't going to change how well those two things perform.

ibilly> My 500MHz powerbook G4 is taking about 58mins a frame on a tinker WU. Mind you, that's because most of the time I'm not using it, leaving it free to concentrate on folding. Do you have other apps open, which might be taking up processor time?

 

[Plutronics]

I've heard reports of three folders who say Panther gave them a 20% Increase !
Speed up the system and everything running under it will be a bit faster!

 

[Decembermouse]

I know this thread is OLD, but I know Plutronics and want to ask a few questions if the Mac gurus are listening.
____________________________________________
400MHz iMac G3 Ruby edition with 128mb 100MHz RAM
____________________________________________

that's the machine I'm dealing with. I'm getting freaking slow frames. it's a tinker core, and after a DAY one frame out of 400 was done. How can I set the verbosity and other flags on a Mac to see more about what's going on, and how to speed it up? I don't see how to set the max. CPU usage either. Thanks!

-edit: benchmark result is 150, now that I've activated -verbosity 9 I can see that. I'll try to get more memory, but that's slow, even for a 400MHz processor, even without SSE/Altivec, isn't it?

 

[Mudbug]

I know this thread is OLD, but I know Plutronics and want to ask a few questions if the Mac gurus are listening.
____________________________________________
400MHz iMac G3 Ruby edition with 128mb 100MHz RAM
____________________________________________

that's the machine I'm dealing with. I'm getting freaking slow frames. it's a tinker core, and after a DAY one frame out of 400 was done. How can I set the verbosity and other flags on a Mac to see more about what's going on, and how to speed it up? I don't see how to set the max. CPU usage either. Thanks!

-edit: benchmark result is 150, now that I've activated -verbosity 9 I can see that. I'll try to get more memory, but that's slow, even for a 400MHz processor, even without SSE/Altivec, isn't it?

send him a private message - he probably won't see this.

 

[Dreadnought]

A 400 mhz G3 is pretty slow for folding, I know, because I have one! I have a G3 B&W @400mhz. Tinkers may take about 1,5 hours to finish a frame, depends on the tinker. The same tinker on my dual G5 1.8 takes about 7 minutes per frame per processor or even faster. I have calculated it several times, the G5 1.8 was about 20 times as fast as the G3 400 mhz. So buy a knew Mac and you will see quite a boost in your production! It was unbelievable last year when I bought mine, within two months I was in the top 100.