Folding@Home

Mercutio said:
If Buck can compete with Newegg's prices, I'll even buy from him exclusively. :)
Click to expand...

Mercutio, if I could sell at NewEgg's prices, I'd be impressed. I can beat mWave by a small margin on some things (usually when outfitting a whole system). But NewEgg, they're tough to beat. Of course, if I sold hundreds of every item I stock each day, it would be easy! :D
 
Buck said:
Mercutio said:
If Buck can compete with Newegg's prices, I'll even buy from him exclusively. :)
Click to expand...

Mercutio, if I could sell at NewEgg's prices, I'd be impressed. I can beat mWave by a small margin on some things (usually when outfitting a whole system). But NewEgg, they're tough to beat. Of course, if I sold hundreds of every item I stock each day, it would be easy! :D
Click to expand...

Well, there is your goal! Sell hundreds of items to reduce cost.
 
Groltz said:
How about these two lines in the client.cfg?
Code: 
[core]
priority=96
cpuusage=100

Were they set differently?
Click to expand...

No. But since I left the computers alone, they did apparently start folding sometime this morning. At least according to their log files. Very strange.

Or annoying. Whichever.
 
Don't know, but they've been going down, then up, then down,... I would expect them to get their act together eventually.
 
Who knows, maybe they will try to remove the random members who spread invontarily accross many teams a while ago. I doubt it though.

On a different subject, I have a good and a bad news for you. The good news is that Buck will be able to pass me without much trouble. The bad news is that I'll soon stop crunching F@h 24x7 on most of my systems. Why? Because since last March (back when we started our G@h team), I have completly dedicated my computers to @home projects. My use of my many computers was dictated by the completion of F@h units. I've almost always waited the end of a step prior to reboot and or install/launch applications that pertubated the smooth working of the F@h client. I did everything possible to make sure that I always return completed units in time, so as not to make the Pande Group waste time for having the results of the units they sent me. Basicly, I fell now that I'm renting my computers instead of owning them. Everything I do on them is done to minimised the impact on F@h. And I'm tired of it.

Also another thing that doesn't help my motivation is the fact that we have hit our summit (32nd position overall, last week IIRC) and that I don't fell we'll ever climb higher in the ranking than that. And since the Pande Group has already published a major piece of their results not so long ago tells me that the biggest part of this project might have already been done.

There are many things I want to try on my computers these days. I have just began messing with Linux, still haven't touched BSD, not had the time to install my one year old Special Edition BeOS CD, still haven't installed Solaris 8.1 just for the kicks, etc. And not to mention the substantial amount of cash it cost me in electricity since last March to let 3 to 7 machines work unstopped.

Some 8000 points later, I don't feel too guilty about slacking down for a while. I think I did more than my part for that project. Like Eric Blair once wrote after the MBF at SR : "It's been good, but now it's time to move on".

I won't uninstall the client from my computers, I'll just stop running it 24x7 and using my systems based on its progression. Expect my weekly production to drop from my current 400-450 points to maybe 100-150 at most. I just have higher priorities in my computing life than F@h right now.
 
Coug, I'm sorry to see that your crunching strength will lessen. But I can appreciate the priorities that you've outlined for yourself. I think we all appreciate the dedication that you put into this project.
 
CougTek,
I think you were being too diligent in returning work units.
I just setup the client as a service and mostly forget about it.
If it doesn't send results while I'm on my dial-up, too bad.
It is the client's fault not mine that it wasn't ready to send the results.
If I have to put any real effort into it, it isn't worth that much to me.
I even hesitated installing it because I had to follow more than a few steps (install firedaemon then install client then configure them together).
I already leave my computer on 24x7 so it was no biggie.

That is why I liked distributed.net's RC5 project:
Install was like three steps. Configuration was a bit harder but Secondary installs were even easier since you could just supply them with a configuration file.
 
Will Rickards WT said:
CougTek,
I think you were being too diligent in returning work units.
I just setup the client as a service and mostly forget about it.
If it doesn't send results while I'm on my dial-up, too bad.
It is the client's fault not mine that it wasn't ready to send the results.
If I have to put any real effort into it, it isn't worth that much to me.
I even hesitated installing it because I had to follow more than a few steps (install firedaemon then install client then configure them together).
I already leave my computer on 24x7 so it was no biggie.

That is why I liked distributed.net's RC5 project:
Install was like three steps. Configuration was a bit harder but Secondary installs were even easier since you could just supply them with a configuration file.
Click to expand...

I agree with Will, you have been too diligent. My values are the same as Will's. I just setup the F@H client as a service and forget about it. I use my machine for other tasks and if F@H doesn't get 100% CPU...too bad.

My machines do stay up 24x7 and for the majority of the day they crunch at full power. I think that is all that should be asked of anyone willing to fold for the team. I find it insane to dedicate a machine entirely to folding; there should be some middle ground where the machine can perform more useful tasks.
 
I don't even bother to run it as a service. It just sits there minimized. I check them once or twice a day as time permits (usually once before going to work & once again in the evening) and whenever results need uploading I dial in and send them.

The servers I'm running it on at work are also just running it minimized on screen via MS Terminal Services. I use TSC to connect to the servers, fire it up, and let it run. I use a spare laptop (which also runs F@H) to control the TSC sessions so I can leave them up 24x7. They run like that most of the time but I've no problem if I need to shut them down overnight or for any other reason.

Oh, I just passed JMP to claim the #10 spot. Tea, I'm gunnin' fer ya!

- Fushigi
 
It's one thing to have a few machines to fold, and another thing to buy numerous machines that do nothing but fold. It’s just a matter of preference I guess. :)
 
Mercutio,

Since I don't have the resources, I won't be able to contribute to the folding at home project. However, I've resigned myself to be counted amongst the crunching power of our dear Bartender. I hope he doesn't mind.

CL
 
I might not have had the 1000th post, but I own the 10000th view of this thread. :D

My stats probably won't drop as much as I anticipated afterall. I'll stop letting my SCSI box work 24x7 though, as it probably is the one that eats the most electricity. It's just a T'bird 800MHz, so it's not the end of the world.
 
DC can get pretty adictive and expensive. I remember I was content running a Celeron 700@900 or so, then when I started messing with Seti and saw my friends P3-933 finishing a unit three full hours ahead of me I started upgrading... again and again. Buying more hardware. Faster hardware. Overclocking. Delta 7000 rpm fans to keep that AXIA athlon cool at 1.4GHz. Then dual! And so on.

The good thing about F@H, and the fact that I moved to it, is that you don't get that much feedback about your performance. With SETI, there were so many settings and hardware configs that you could mess with, and see the results right there and then - swapping from a BX board with that Celeron to a KT133A with Athlon dropped my WU times from 8.5-9 hours down to 7. Overclocking with a huge fan brought it down to under 6 hours.

Changing to DDR (KT266A) and XP1800 overclocked to 1667 got me around 3 hours. It was then that I moved to F@H and haven't replaced my processor since. I dont know how long it takes for a unit to be completed, and each one can take a totally different amount of time to another.

And thats good, otherwise I would probably be running an XP3000 by now, overclocked, and as a result having to eat bread and drink water every day to finance it :mrgrn:

Back then, I made everyone I knew run it to get my stats up. Now its just me. Heck, I don't even look at the stats any more. No more overclocking either. It just runs.

Certainly more boring, but better than upgrading every 3 month just to see a number rising, right?
 
The news I'm waiting for is the release of a fixed Linux client that'll work fine with glibc 2.3.1. Now, we have to install a proxy and lure the client regarding ports if we want to make it work without a segmentation fault. It's lame. The problem has been known since last Fall and they still havent fixed it. This isn't serious.
 
coug, i mentioned to you that this problem is easier to fix than you thought.. it took me literally 10 seconds to fix it...

just use your ISP's proxy, no need to install your own... perhaps you dont even need a proxy, you just need to change "localhost" to "127.0.0.1" in the client.cfg.... not sure if that one will work, but it's a 2second fix...
 
Yes James, I saw this page a few months ago, but I haven't read it in detail yet. This solution would be even better if it would include a how-to guide to make all the clients run the windows version through Samba, since the Win32 FAH client is much more performant than the Linux one.

But I like the principle.
 
Incase no one is paying attention out folding has dropped like a rock recently. Any explanations? If Cliptin is less than 1000/week.
 
Damn.. I just realised I've not been folding for the past one and a half months. Actually I've been folding, but who knows under what ID... certainly not mine!

:evil:
 
P5-133XL said:
Incase no one is paying attention out folding has dropped like a rock recently. Any explanations? If Cliptin is less than 1000/week.
Click to expand...

I suppose you mean "even Cliptin". I ran out of machines at work. Whereas I used to have 15 machines at work and 5 machines at home, I now have only one machine at work and 3-4 machines at home. Frankly, I'll bet most of my productivity these days is donated.
 
Yes, I ment even.

An elimination of 16-17 machines should definately drop the teams folding capacity. Tis a shame, but who knows what will happen in the future.
 
Mainly to increase my folding, but also because I had a cooling problem with my XP1900, I slipped in a 2400 the other week. (0.13 instead of 0.18u - i.e., cooler.) Alas, I got stupid and tried a Volcano 9, and in order to keep the temp reasonable without drowning the radio out, I needed to kill the folding process. I've replaced it with a more sensible plain vanilla thing now, so I should be back on the stats page soon.
 
My potential threat is JMP and he will overtake me on Jun 12, 2020 03:00 AM. :lol:
 
Clocker - Jun 19, 2003 02:00 PM
The_JoJo - Dec 12, 2003 01:00 PM
Cliptin - Apr 21, 2004 12:00 AM

Clocker, I'm gunnin fer ya!

I like how it even estimates the time of day for an event nearly a year out.

- Fushigi
 
CougTek said:
Look at our team graph at statsman. Look at it carefully. We are now 32nd overall. This is the highest ranking we've ever been...and this is the highest ranking we'll ever be. Our team is now at his apogee. Enjoy it while it last (it will be short).

Be proud of what we've done.
Click to expand...
Every time I look at our stats, I remember that post I made. How has it changed since then...

In order to compete again with teams around the 30-35th position, we would need no less than 100 additional high-end Athlon XP. I made the count. That would bring our weekly total around 20000 points. Unfortunately, this seems like an impossible task. And since we don't get much new blood around here, those 100 additional processors would have to come from our members who are already generously contributing. This too is impossible, especially on the financial side.

We are doomed to slide, and slide, and slide...

Something must be done...but what? We originally started the team partly to attract new members to our forum. I think this point can be called a failure. The way I see it, F@H and other @home projects are a curse. You always have to increase your productivity to remain competitive and not fall to oblivion. I'm sorry to have started the whole thing with the G@h idea I threw in February of last year.
 
You must log in or register to reply here.
Back
Top