Ok it pains me to ask

T

Tormond

[H]ard|DCer of the Month - June 2007
Joined
May 4, 2003
Messages
1,406
Ok Marty you seem to have the uber sekret l33t setup for point production. Since I just built my little Core2 box (that folds 100% on both cores of course) gimme the low down on exactly what I need to set in the config file to maximize whatever it is to maximize. Since my 4400+ machine went down (I need to get a replacement MB damn lightning) my big machine at home (not counting the servers) has been down for almost a month and trying to get caught back up. I just don't want to shortchange myself anymore than I have to (I never shut down folding) and hoping this chip will overclock worth a damn (I have had teh worst luck with overclocking). Anyway any help from anybody would be greatly appreciated.

Thanks

T-
 
Easy, have 100 of those systems running!

Na, config for big WU (accept > 5mb packets). From there it's all up to the -adv methods game. Basically there are trends where the odds of high PPD WU's lie in the -adv group some weeks and some weeks they don't.

Most of the time it's a safe bet to run w/o -adv. However taking that gamble for 1495 series WU's and scoring ~ 900+ PPD off >each< (o frigging lot) core from the box is always a plus.

Only real way to track this is from forums or monitoring a certain numbers of boxes set one way or the other and watch for trends. The classifications are listing on the psummaries but it's a mess and next to worthless, especially since the same WU can exist in both groups (normal/-adv late stage beta group).

Those would be the legit ways to max production.
 
Well I always config for big WU (with 4GB RAM that just isn't an issue for me). Do I need to run the Advmethods flag or just leave it as

fah504-console -configonly (to start the setup)

Tormond
33
yes to start as service
yes to accept big WU
eerything else (except Machine ID) using the defaults that are given (IE no proxy, dont use IE settings, and whatever the extra stuff that 504 added that I don't know what it does)

Thanks again
 
90% of the time your score pretty well w/o -adv flag. It's the most miss used flag there is, try it w/o.
 
the WUs change all the time. Big WU and Advmethod tag control this so just try with both on, off and either one on until you hit what you like :)
 
The deal is WU's change all the time and there is no one right config for all bases, all times.
 
i leave them both on... dunno if it makes a difference?
 
FLECOM said:
i leave them both on... dunno if it makes a difference?
Click to expand...

I've been meaning to ask, you borged all those systems with -big on? I figured you did from the points per WU you do. If QMD's were to come back you would be in a world of hurt on those systems, in my opinion.
 
I brought up the idea, in hopes others would try and keep it up to date. Honestly for my own systems I rarely keep track of what they are folding.
 
marty9876 said:
I've been meaning to ask, you borged all those systems with -big on? I figured you did from the points per WU you do. If QMD's were to come back you would be in a world of hurt on those systems, in my opinion.
Click to expand...

not a clue... i installed it on one computer and made a ghost image from it... now whenever anyone installs a new computer at work it gets ghosted automagically and bam, there is F@H :)
 
FLECOM said:
not a clue... i installed it on one computer and made a ghost image from it... now whenever anyone installs a new computer at work it gets ghosted automagically and bam, there is F@H :)
Click to expand...

I always leave FAH disabled in the images, worry wart I am.

Keep in mind if WU's with large RAM requirments come back around you could be burning 900+ RAM (across 2 instances) and some of these might suck up all available RAM bandwidth.

Not trying to be a buzz kill, just protecting the mother of all borgs!
 
Thanks for the assist marty (and others). What are you guys finding to be the best solution for tracking PPD output of a single box? I used to use EM before I quit caring about points so much and just folded :) I am just curious to see how well the Conroe does moreso than I am really interested in actual PPD.


 
roftranspo said:
Yurt would be your answer....search this forum :)
Click to expand...
Thanks for pumping YURT, roftranspo.

By the way, will you be upgrading your machines to the new client soon?
 
You might also try the SMP client and compare for points, which are always consistent 587 :) (there is one monster 1760, but rarely shows up). I know its meant for 4 cores but does run on 2......just my 2 cents
 
Ok after letting it run for a few days it seems that I have done something horribly wrong acording to yurt. Or maybe I just n't figure out what Yurt is saying. Can someone who understands this look at the core2 box under my username (the only one there) and see if I am doing something wrong or do those numbers look right?
 
marty9876 said:
I always leave FAH disabled in the images, worry wart I am.

Keep in mind if WU's with large RAM requirments come back around you could be burning 900+ RAM (across 2 instances) and some of these might suck up all available RAM bandwidth.

Not trying to be a buzz kill, just protecting the mother of all borgs!
Click to expand...

well, the machines have a gb of ddr2... if they go slow, oh well, i will just blame it on "the system must be slow today... probably amazon has a sale again" and the drones will continue on doing whatever it is they do
 
FLECOM said:
well, the machines have a gb of ddr2... if they go slow, oh well, i will just blame it on "the system must be slow today... probably amazon has a sale again" and the drones will continue on doing whatever it is they do
Click to expand...
LOL... Drones...
 
The numbers seem low, times are slow.
At 2.2 Ghz p2124 takes me very close to 3 days not the 5.5 days your doing it in.
Open your FAHlog.txt file and look at the start of a protien.
Do you see the SSE Boost line ?

[13:04:45] Writing local files
[13:04:50] Extra SSE boost OK. <------ This one.
[13:04:50] Writing local files
[13:04:50] Completed 0 out of 250000 steps (0)

If not then your running at half speed due to it being turned off.
Stop the client.
Add the -forceasm switch to the start shortcut or registry entry if started that way.
Restart the client and you should now run at full speed.

Luck ............. :D
 
Tigerbiten said:
Add the -forceasm switch to the start shortcut or registry entry if started that way.
Restart the client and you should now run at full speed.

Luck ............. :D
Click to expand...

How exactly do you add -forceasm to the registry? SSE Boost should show up in the log every time the client restarts no?
 
If the client is shut down incorrectly due to a power cut, system freeze, reboot useing the reset switch, etc, etc then the client thinks it's unstable and sets the stable/unstable flag to unstable.
Therefore the next time it start, without the -forceasm switch, the SSE boost is switched off untill the client is shut down correctly.
Once the client shuts down correctly then the stable/unstable flag is set to stable and SSE boost is engaged.
The -forceasm switch overrides the stable/unstable flag and henceforth SSE boost is always engaged.

Not 100% sure how to set it in the registry as non of my clients run as a service.
But it needs to go in the line that starts the service.

Luck ................ :D
 
Not sure what it is doing. I am showing the SSE boost as OK but will go in and force it to on just in case. I thought the numbers looked pretty weak considering tha the Core2 CPUs are supposed to be gods gift to computing
 
You must log in or register to reply here.
Back
Top