• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

+30 ghz FTW

ND40oz said:
So if I have 2000 machines out there, how do you normally go about installing. Is there a silent switch that I can drop a config file in with or am I manually installing it? Anyone configure an SMS push for the client?

They on a domain?
 
marty9876 said:
Short by about 20 CPUS at least.

176 field borgs @ 75%. 2.4 GZH > 2.66 GZH call it 2.5 average. Call it 1.875 X 176= 330 GZH 24x7 (production units).

~ 25 systems @ shop @ 100% 24/7 2.5 GZH average = 62.5 GZH

6 GZH home/personal notebook= 6 GZH

1 GZH borged= 1 GZH

Misc corp borged, scheduled tasked while system idle = Who knows

400 GZH 24/7 plus things I've forgotten about.
and uh, you have... written approval for all that, right?
 
Oldbenwa said:
Verbal, or Oral? :D


D00d, you forgot the crapper farm, which I noticed are now back under your name...

Yea, I forgot to talk to you about that.... I decided, all on my own... Mine in the summer since I have to deal with the heat and yours in winter. Deal?

Ok, I got sick of FLECOM out scoring me. Real reason.

Add about 50 Gzh to the number above I guess.
 
marty9876 said:
They on a domain?

Definitely. We mainly use SMS for pushes, but I guess I could use AD with group policy as well if someone has a good way to do it. These are just workstation type kiosks that employees can come up and use. They're autologged in and have limited access to select applications. I need to have a good written plan with deployment procedures and testing done before I can even bother getting approval. Just trying to do my homework first...
 
ND40oz said:
Definitely. We mainly use SMS for pushes, but I guess I could use AD with group policy as well if someone has a good way to do it. These are just workstation type kiosks that employees can come up and use. They're autologged in and have limited access to select applications. I need to have a good written plan with deployment procedures and testing done before I can even bother getting approval. Just trying to do my homework first...


ocforums has domain installers, let us know if you interested we can link it for ya. someoen tried in the past, no luck. I've used the "get computer names" kinda step one VB script, works-- never pushed anything out via this.

I stole this idea from OC-AMD- run a scheduled task locally on each box under admin account to run folding only when machine is in idle and store the files on a network share. Works ok for me, get delay write errors (along with applicaiton pop up messages....) on the workstations when share goes offline. Still need to clean that up. Bonus sellign point of this is FAH is actually not running while machine is in use. Or just set a flat out schedule tasks and run after hours. Huge pain setting this up, OC-AMD found a way yo change passwords at least under some batch process I think.
 
roftranspo said:
ask unhappy_mage...i think he made an installer for the console version. but i dont know where it is at
I made a prettified installer for the command line client here. It's not suitable for deploying things, though - it is best suited to a single machine, with a single processor (!). I'll modify it to support multiprocessor machines if there's demand.

As for deploying this... not pretty. You can pretty much figure out what the .nsi is doing and replicate it in a vbs or whatever you're using to roll things out; if you get stuck on it (having only the low-level constructs nsi allows makes for some pretty weird code sometimes) let me know and I can usually figure out what I did. If you can specify what language it should be written in, I'll even take a whack at it myself. My time is pretty limited these days, between 8+ hours a day at work and no computer to myself, but I can make time for 2000 machines :drool:

 
unhappy_mage said:
I made a prettified installer for the command line client here. It's not suitable for deploying things, though - it is best suited to a single machine, with a single processor (!). I'll modify it to support multiprocessor machines if there's demand.

As for deploying this... not pretty. You can pretty much figure out what the .nsi is doing and replicate it in a vbs or whatever you're using to roll things out; if you get stuck on it (having only the low-level constructs nsi allows makes for some pretty weird code sometimes) let me know and I can usually figure out what I did. If you can specify what language it should be written in, I'll even take a whack at it myself. My time is pretty limited these days, between 8+ hours a day at work and no computer to myself, but I can make time for 2000 machines :drool:


Excellent, this gives me some good ideas. I'm going to talk to the guys that are the admins these systems (3 guys support these 2000 workstations) and see if I can get it running in their test environment. Thanks all.

Edit: Go Retrievers!, small world.
 
Ya... got some points posted last night, looks like about 260 made it to completion and posted while I was sleeping. I put another 600 ppd online last night so the results from those should be hitting the site around lunch time and this afternoon. I can't wait to have a solid amount of work to take a look at, it will be most cool.

On a related note, I'm running a truely lovely Lian Li v2000 case with a built in Koolance Inex 3 watercooling system. When I started folding on my opteron 170 it only raised my water loop temps 1 degree celcius. :) I love wc.
 
ND40oz said:
Excellent, this gives me some good ideas. I'm going to talk to the guys that are the admins these systems (3 guys support these 2000 workstations) and see if I can get it running in their test environment. Thanks all.

Edit: Go Retrievers!, small world.

2000 computers? was that an exageration? :eek:
 
dekard said:
Ya... got some points posted last night, looks like about 260 made it to completion and posted while I was sleeping. I put another 600 ppd online last night so the results from those should be hitting the site around lunch time and this afternoon. I can't wait to have a solid amount of work to take a look at, it will be most cool.

On a related note, I'm running a truely lovely Lian Li v2000 case with a built in Koolance Inex 3 watercooling system. When I started folding on my opteron 170 it only raised my water loop temps 1 degree celcius. :) I love wc.

I am working on cutting up the same case right now. I love it :cool:
 
ND40oz said:
Edit: Go Retrievers!, small world.
Small indeed! My brother and I will both be there next semester.

I'll be glad to help you out with this project. Whatever language it's in, I can probably figure out :p

 
ahh yes, I'm finally starting to throw down some points. I hope in the next couple of days to get my production stabilized and working at 100%.

Its pretty funny, the main system in my sig, a opteron 170, wasn't overclocked anymore and I just threw it back on. It raised my home ppd from 700 to 1200. Shesh!
 
dekard said:
Oh, and the dual xeon 2.8 server is pushing over 1k by it self. Lol...

That's incredible, my dual prestonia box at 2.8 is only throwing out 350 a day according to EMIII. How many instances are you running, 4?
 
ND40oz said:
That's incredible, my dual prestonia box at 2.8 is only throwing out 350 a day according to EMIII. How many instances are you running, 4?
3, totaling about 80 % of the cycles on the server. I can't go to 4 because the OS will complain about a lack of idle cycles.
 
you could do 4, but run the 4th at 50% instead of 100%.... or 75% or 86% or whatever would keep the OS from complaining about the lack of idle cycles..... (what OS is it? I've never heard of that before?)


Keep on Folding!! For the [H]orde!!

 
dekard said:
Oh, and the dual xeon 2.8 server is pushing over 1k by it self. Lol...

Ultra Nice! :) If all my boxen were this, I'd change my handle to Silicon Celerator.

 
OSUguy98 said:
you could do 4, but run the 4th at 50% instead of 100%.... or 75% or 86% or whatever would keep the OS from complaining about the lack of idle cycles..... (what OS is it? I've never heard of that before?)


Keep on Folding!! For the [H]orde!!

Hey, nice idea... Its running Windows Server 2003 Small Business Server... I get admin alerts via email when I install any distributed client that will peg the cpu at 100% for hours at a time.. I'll add a third and set it to a about 60%... should be safe enough. If not, I'll drop it back to whatever it takes to get it to shut up.
 
dekard said:
Hey, nice idea... Its running Windows Server 2003 Small Business Server... I get admin alerts via email when I install any distributed client that will peg the cpu at 100% for hours at a time.. I'll add a third and set it to a about 60%... should be safe enough. If not, I'll drop it back to whatever it takes to get it to shut up.

Eh alerts are for suckers...as long as it's not production, disable the alerter service...
 
ND40oz said:
Eh alerts are for suckers...as long as it's not production, disable the alerter service...
it is a production box.. Can't and dont' want to disable them.
 
dekard said:
it is a production box.. Can't and dont' want to disable them.

Oh shit, you fold on production servers...you're a better man then I...
 
dekard said:
Hey, nice idea... Its running Windows Server 2003 Small Business Server... I get admin alerts via email when I install any distributed client that will peg the cpu at 100% for hours at a time.. I'll add a third and set it to a about 60%... should be safe enough. If not, I'll drop it back to whatever it takes to get it to shut up.

You can change the threasholds on these alerts I think.

edit: yep. Server Management>Monitoring and Reporting>Change Alert Notifications>Preformance Counters> Processor Activity> either set to 1% or flat out disable it.
 
ND40oz said:
Oh shit, you fold on production servers...you're a better man then I...
Lol...
i just bought the dang thing with lots of good money. Got a solid dell support package and lots of registered ram.

I better be equipped to handle it after all the money I spent on it.
 
marty9876 said:
You can change the threasholds on these alerts I think.

edit: yep. Server Management>Monitoring and Reporting>Change Alert Notifications>Preformance Counters> Processor Activity> either set to 1% or flat out disable it.
Marty, what would you do if this was a pretty critical box? Fold away using 80% of available cycles on a production box or turn them off and run one more instance of it?
 
Oh, i just added a dual piii 750 box I've got as a media server. Not going to be the fastest folder in my barn but it should be reliable enough.
 
dekard said:
Oh, i just added a dual piii 750 box I've got as a media server. Not going to be the fastest folder in my barn but it should be reliable enough.
slow and steady always wins the race my friend :) i had at one point a P2 400 folding :)
 
roftranspo said:
slow and steady always wins the race my friend :) i had at one point a P2 400 folding :)
I've added 7.15 ghz today to the folding.. i can't wait to see my stats in a week or so once things settle down a bit. 30.883 ghz online as we speak. I've got another 2800+ in my laptop which is only up a few hours a week and another 933 box to install at the office, so i should settle in at 31.816 + laptop sometimes....
 
dekard said:
Marty, what would you do if this was a pretty critical box? Fold away using 80% of available cycles on a production box or turn them off and run one more instance of it?

Not run FAH on the box period. If it's a Domain Controller, keep FAH off.

Look at things this way, all the rest of your borgs vs. this one. Say something goes wrong and you get the rest of the borgs pulled over this one. Years of folding on all but this one, or x amount till something goes wrong on this one.

Not saying something will go wrong, shit happens (and has).

That aside, if it's a dual CPU (not dual CPU and dual core)- only run two instances anyways. Running on the hyper-threading side of things provides only a slight increase in production (15% max). FAH and hyper-threading a long confusing conversation leading to much on nothing.

Anyways, my opinion.
 
marty9876 said:
Not run FAH on the box period. If it's a Domain Controller, keep FAH off.

Look at things this way, all the rest of your borgs vs. this one. Say something goes wrong and you get the rest of the borgs pulled over this one. Years of folding on all but this one, or x amount till something goes wrong on this one.

Not saying something will go wrong, shit happens (and has).

That aside, if it's a dual CPU (not dual CPU and dual core)- only run two instances anyways. Running on the hyper-threading side of things provides only a slight increase in production (15% max). FAH and hyper-threading a long confusing conversation leading to much on nothing.

Anyways, my opinion.
food for thought.. im gonna sleep on it..
 
marty9876 said:
Not run FAH on the box period. If it's a Domain Controller, keep FAH off.

Look at things this way, all the rest of your borgs vs. this one. Say something goes wrong and you get the rest of the borgs pulled over this one. Years of folding on all but this one, or x amount till something goes wrong on this one.

Not saying something will go wrong, shit happens (and has).

I concur.

If it's small business server, most likely it is your DC as well as a bunch of other things. And the bad thing about SBS is you can't have multiple DCs, meaning if it goes down, you're SOL until you can bring it back up. Best advice is, if you can't afford to have it go down, don't put anything on it that you can't get tech support for. And [H] doesn't count as tech support in the boss's eyes ;)
 
ND40oz said:
I concur.

If it's small business server, most likely it is your DC as well as a bunch of other things. And the bad thing about SBS is you can't have multiple DCs, meaning if it goes down, you're SOL until you can bring it back up. Best advice is, if you can't afford to have it go down, don't put anything on it that you can't get tech support for. And [H] doesn't count as tech support in the boss's eyes ;)
the only good thing I can say is that I own the company. So, my server, my employees etc... I won't have to explain what I'm doing, my only concern is how would folding affect the reliability or longevity of the server. I'm having trouble coming up with reasons its going to break it or wear it out faster...

But, if you toons are encourging me to limit my folding you all must think its serious.
 
Back
Top