Configuration Recommendations

W

Wharf_Rat

[H]ard|Gawd
Joined
Mar 15, 2000
Messages
1,151
Hi all. I just got back into folding...used to run in back when it first came on the scene. I don't seem to remember the name I used, it was when the team was first started. Does Stanford purge for inactivity?

Anyway, onto the machines.

I have a e6400 and e6600 both running 2 consoles as a service, allowing greater than 5mb units, with -forceasm and -advmethods. Is this the best configuration for the gear?

I am about to add a 3.0Ghz p4 w/HT to the mix and was wondering how I should set it up. I am assuming a single console, but should I accept larger workunits?

Thanks for any help!

 
Wharf_Rat said:
Hi all. I just got back into folding...used to run in back when it first came on the scene. I don't seem to remember the name I used, it was when the team was first started. Does Stanford purge for inactivity?

Anyway, onto the machines.

I have a e6400 and e6600 both running 2 consoles as a service, allowing greater than 5mb units, with -forceasm and -advmethods. Is this the best configuration for the gear?

I am about to add a 3.0Ghz p4 w/HT to the mix and was wondering how I should set it up. I am assuming a single console, but should I accept larger workunits?

Thanks for any help!

Click to expand...

STOP TIS HAMMERTIME...

For your C2D I would say that u should be running the smp clients on them big wu.
For your P4 run a single reg client big wu w/ -forceasm.

Anything else?

 
Thanks for the fast responses.

Is there a significant difference in crunching w/ the SMP client?

I held off on the SMP client for now, due to the reports I have read on a couple boards with windows as a service. Also, I didn't really want to have to keep updating it every couple months. I may revisit it here soon...

Thanks for the name search link, Ill give it a whirl. Doesn't really matter, as [H] got the points, but my inner stat whore keeps thinking about it!!
 
Wharf_Rat said:
Thanks for the fast responses.

Is there a significant difference in crunching w/ the SMP client?

I held off on the SMP client for now, due to the reports I have read on a couple boards with windows as a service. Also, I didn't really want to have to keep updating it every couple months. I may revisit it here soon...
Click to expand...

Differences - big bonus on points obtained...
As services and etc - it'll work but understandable if u don't want to run beta and maintain it until there's a stable one out. In the meantime, run dual single clients with big wus and I don't recommend advmethods but that's me.

Good luck and also feel free to ask anything :)

 
The -advmethods, i should remove from all the clients?

Thanks for the advice. I will probably take another look at the SMP client, try it on one of the machines and see how it goes.

Got the 3.0Ghz running now.
 
The main difference between running the SMP vs console is that you need your machine to be able to transfer WU every several days. SMP deadlines are short, usually 2 or 3 days.
That obstacle conquered the point differential is extreme, for now.
 
The SMP client will give you around 4 or 5 times the Points per Day that running twin CPU clients will.
I'm only running 10 SMP clients for ~17k PpD.

As for the "-advmethods" flag, I aways ran the CPU clients with one on and one off on a dual core box.
But it all depends when the new protiens drop and how they fold, which way is better.

Luck ............ :D
 
Enough arm twisting! Ill do it!

I just conifigued one machine and waiting for my 2 consoles to finish the current work...then ill start the service. Have to see what the other machines status is.

Thanks everyone of your replies.
 
I got the SMP client running on both of the core2 machines.

One question. It is normal for the log to just say working, as opposed to counting down the checkpoints like the normal console version?

Thanks
 
no .........

Your FAHlog.txt file should look similar to the standard clients one with only a few extra line as each workunit starts and finishes.
Mine looks like ..........
[12:25:40] + Processing work unit
[12:25:40] Core required: FahCore_a1.exe
[12:25:40] Core found.
[12:25:40] Working on Unit 01 [September 30 12:25:40]
[12:25:40] + Working ...
[12:25:41]
[12:25:41] *------------------------------*
[12:25:41] Folding@Home Gromacs SMP Core
[12:25:41] Version 1.74 (March 10, 2007)
etc, etc.

If its not passed the working line then its probably hung.

Did you set it up as a service.
If you did you need to manualy add you user name and password to the service call.
Its why Stanford does not recomend running it as a service.

If not try rerunning the install.bat.
If that does not work try a reboot.

Luck ............ :D
 
Reading: 1 - Wharf_Rat: 0

Jon855 said:
8. Close the registry editor and reboot. Trying to start the service without rebooting will not work, you must reboot
Click to expand...

I never rebooted...

Then to top it off, I replied to the wrong thread!!!

Thanks anyway for the reply.
 
It seems like my e6400 is taking an awful long time to complete work units...anything I can take a look at? Bad work unit? Bad config? Thanks

Launch directory: C:\FAH
Service: C:\FAH\fah
Arguments: -svcstart -forceasm

Launched as a service.
Entered C:\FAH to do work.

Warning:
By using the -forceasm flag, you are overriding
safeguards in the program. If you did not intend to
do this, please restart the program without -forceasm.
If work units are not completing fully (and particularly
if your machine is overclocked), then please discontinue
use of the flag.

[17:24:37] - Ask before connecting: No
[17:24:37] - User name: WRat (Team 33)
[17:24:37] - User ID: ***************
[17:24:37] - Machine ID: 1
[17:24:37]
[17:24:37] Loaded queue successfully.
[17:24:37]
[17:24:37] + Processing work unit
[17:24:37] Core required: FahCore_a1.exe
[17:24:37] Core found.
[17:24:37] Working on Unit 01 [September 30 17:24:37]
[17:24:37] + Working ...
[17:24:37]
[17:24:37] *------------------------------*
[17:24:37] Folding@Home Gromacs SMP Core
[17:24:37] Version 1.74 (March 10, 2007)
[17:24:37]
[17:24:37] Preparing to commence simulation
[17:24:38] - Ensuring status. Please wait.
[17:24:55] - Assembly optimizations manually forced on.
[17:24:55] - Not checking prior termination.
[17:24:56] - Expanded 1140871 -> 5811661 (decompressed 509.4 percent)
[17:24:56]
[17:24:56] Project: 2652 (Run 0, Clone 582, Gen 20)
[17:24:56]
[17:24:57] Assembly optimizations on if available.
[17:24:57] Entering M.D.
[17:25:04] Calling FAH init
[17:25:07] Read topology
[17:25:07] files
[17:25:07] ing from checkpoint)
[17:25:07] Read checkpoint
[17:25:07] Protein: Protein
[17:25:08] Writing local files
[17:25:09] Extra SSE boost OK.
[17:25:10] Writing local files
[17:25:10] Completed 0 out of 1000000 steps (0 percent)
[19:08:39] Writing local files
[19:08:40] Completed 10000 out of 1000000 steps (1 percent)
[20:52:12] Writing local files
[20:52:12] Completed 20000 out of 1000000 steps (2 percent)
[22:35:34] Writing local files
[22:35:34] Completed 30000 out of 1000000 steps (3 percent)
[00:19:13] Writing local files
[00:19:14] Completed 40000 out of 1000000 steps (4 percent)
[02:02:41] Writing local files
[02:02:41] Completed 50000 out of 1000000 steps (5 percent)
[03:46:45] Writing local files
[03:46:45] Completed 60000 out of 1000000 steps (6 percent)
[05:30:24] Writing local files
[05:30:24] Completed 70000 out of 1000000 steps (7 percent)
[07:14:01] Writing local files
[07:14:01] Completed 80000 out of 1000000 steps (8 percent)
[08:57:23] Writing local files
[08:57:24] Completed 90000 out of 1000000 steps (9 percent)
[10:41:42] Writing local files
[10:41:43] Completed 100000 out of 1000000 steps (10 percent)
[12:26:22] Writing local files
[12:26:23] Completed 110000 out of 1000000 steps (11 percent)
[14:12:30] Writing local files
[14:12:31] Completed 120000 out of 1000000 steps (12 percent)
[14:43:56] Service stop request received.
 
Wharf_Rat said:
It seems like my e6400 is taking an awful long time to complete work units...anything I can take a look at? Bad work unit? Bad config? Thanks

Launch directory: C:\FAH
Service: C:\FAH\fah
Arguments: -svcstart -forceasm

Launched as a service.
Entered C:\FAH to do work.

Warning:
By using the -forceasm flag, you are overriding
safeguards in the program. If you did not intend to
do this, please restart the program without -forceasm.
If work units are not completing fully (and particularly
if your machine is overclocked), then please discontinue
use of the flag.

[17:24:37] - Ask before connecting: No
[17:24:37] - User name: WRat (Team 33)
[17:24:37] - User ID: ***************
[17:24:37] - Machine ID: 1
[17:24:37]
[17:24:37] Loaded queue successfully.
[17:24:37]
[17:24:37] + Processing work unit
[17:24:37] Core required: FahCore_a1.exe
[17:24:37] Core found.
[17:24:37] Working on Unit 01 [September 30 17:24:37]
[17:24:37] + Working ...
[17:24:37]
[17:24:37] *------------------------------*
[17:24:37] Folding@Home Gromacs SMP Core
[17:24:37] Version 1.74 (March 10, 2007)
[17:24:37]
[17:24:37] Preparing to commence simulation
[17:24:38] - Ensuring status. Please wait.
[17:24:55] - Assembly optimizations manually forced on.
[17:24:55] - Not checking prior termination.
[17:24:56] - Expanded 1140871 -> 5811661 (decompressed 509.4 percent)
[17:24:56]
[17:24:56] Project: 2652 (Run 0, Clone 582, Gen 20)
[17:24:56]
[17:24:57] Assembly optimizations on if available.
[17:24:57] Entering M.D.
[17:25:04] Calling FAH init
[17:25:07] Read topology
[17:25:07] files
[17:25:07] ing from checkpoint)
[17:25:07] Read checkpoint
[17:25:07] Protein: Protein
[17:25:08] Writing local files
[17:25:09] Extra SSE boost OK.
[17:25:10] Writing local files
[17:25:10] Completed 0 out of 1000000 steps (0 percent)
[19:08:39] Writing local files
[19:08:40] Completed 10000 out of 1000000 steps (1 percent)
[20:52:12] Writing local files
[20:52:12] Completed 20000 out of 1000000 steps (2 percent)
[22:35:34] Writing local files
[22:35:34] Completed 30000 out of 1000000 steps (3 percent)
[00:19:13] Writing local files
[00:19:14] Completed 40000 out of 1000000 steps (4 percent)
[02:02:41] Writing local files
[02:02:41] Completed 50000 out of 1000000 steps (5 percent)
[03:46:45] Writing local files
[03:46:45] Completed 60000 out of 1000000 steps (6 percent)
[05:30:24] Writing local files
[05:30:24] Completed 70000 out of 1000000 steps (7 percent)
[07:14:01] Writing local files
[07:14:01] Completed 80000 out of 1000000 steps (8 percent)
[08:57:23] Writing local files
[08:57:24] Completed 90000 out of 1000000 steps (9 percent)
[10:41:42] Writing local files
[10:41:43] Completed 100000 out of 1000000 steps (10 percent)
[12:26:22] Writing local files
[12:26:23] Completed 110000 out of 1000000 steps (11 percent)
[14:12:30] Writing local files
[14:12:31] Completed 120000 out of 1000000 steps (12 percent)
[14:43:56] Service stop request received.
Click to expand...

Bad config for sure. It looks like you started the SMP service then run the console as well. Make sure you say "no" to the question asking if you want to start in service mode.

 
I have it configured to run as a service and would like to keep it that way if possible.. I havent opened the console at all, other than to config it initially.

What line tips you off to the console running?

Thanks
 
Wharf_Rat said:
I have it configured to run as a service and would like to keep it that way if possible.. I havent opened the console at all, other than to config it initially.

What line tips you off to the console running?

Thanks
Click to expand...

I suggest you remove the service since the SMP is extremely buggy when run in service mode. Open the taskmanager then see how much Fahcore_A1.exe threads you have ?

 
It has 4 threads running for Fahcore_A1.exe.

Only thing I can think that caused this was the fact that I installed it on this machine, via RDP, over a Citrix teminal server session. I used shutdown -r to reboot the box and maybe it didnt stop the service properly. I am not sure, but i will keep an eye on it.

I went ahead and rebooted the machine again...and the following is what i am looking at now. Trimmed down from 1:45 checkpoints to ~:16.

14:46:36] Completed 122861 out of 1000000 steps (12 percent)
[14:46:36] Extra SSE boost OK.
[14:59:09] Writing local files
[14:59:09] Completed 130000 out of 1000000 steps (13 percent)
[15:16:44] Writing local files
[15:16:44] Completed 140000 out of 1000000 steps (14 percent)
 
Wharf_Rat said:
It has 4 threads running for Fahcore_A1.exe.

Only thing I can think that caused this was the fact that I installed it on this machine, via RDP, over a Citrix teminal server session. I used shutdown -r to reboot the box and maybe it didnt stop the service properly. I am not sure, but i will keep an eye on it.

I went ahead and rebooted the machine again...and the following is what i am looking at now. Trimmed down from 1:45 checkpoints to ~:16.

14:46:36] Completed 122861 out of 1000000 steps (12 percent)
[14:46:36] Extra SSE boost OK.
[14:59:09] Writing local files
[14:59:09] Completed 130000 out of 1000000 steps (13 percent)
[15:16:44] Writing local files
[15:16:44] Completed 140000 out of 1000000 steps (14 percent)
Click to expand...

This is better. It's on the normal averages :)

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