I have an 2x p4 820d & a amd64 3200+. Why cant i beat sm27 we just go back and forth moving up at the same time but i can stay ahead like 1 day then ill be above.
alright i scrounged through some parts i gotta buy a new power supply, 3 power cords and a dvd drive and ill have 3 new systems. amd 2200+, 1Ghz celeron, & PIII 500MHz all up and running. I should be able to get sm27 then!
Fresh (warm) air, vegemite, being driven by Team spirit all plays part of the success story that SM27 has become. She is hooked up to a 29 computer network and they are all her BEST friends.
When someone tries to beat her steady run, the others chip in with moral support and do what they can to keep her climbing the folding ladder.
Just having a look at the stats and me thinks poor SM27 will have a hard time keeping up with the challenge.
How many PPD are you doing? With a pair of 820D's and an A64 3200 you should be smoking a single system. Are you running 2 clients on the P4's? Each set for a different cpu id?
the p4's are both running 2 clients, the amd64 i think is acting up i think i need to reinstall windows. It was locking up today so i might not have seen it since i havent looked at the screen in like a week. One of the p4's get shut off each night though (damn work)
Yoy can't really request the type you want, however through the use of flags you can point in a general direction. (also how your client is configured will help)
With Intel CPU's and signing up (and being accepted) to test beta WU's, you can do quite well with the QMD-core WU's. Oh, yeah - you'll need a Gig of RAM, too.
Checking the advmethods flag under Advanced Options in the new 5.04 Console helps some if you're an AMD man.
Personally, I don't worry about the "stinker" WU's. My reasoning is that Stanford wouldn't send them out if they weren't necessary. Who knows - some of the answers we're hoping to find might well lie in those low-point-yield Work Units. I just try and keep my rigs running at top efficiency and let the rest take care of itself.
Bud, how much ram do you have installed on those 820 machines? If you are running 2 consoles, I recommend 1GB. To get a steady stream of QMD's on my 2 P4 machines, I have the clients configured with bigpackets=yes and start them with the -advmethods flag. I don't know how long ago it's been since either of my P4 machines have drawn a non-QMD from the normal assignments while configured like that.
were still going at it. Im adding a 3GHz tommorow and i know i shouldnt but i think ill be buyin a 246 opteron or a new intel d820 for my fileserver and that will be a good folder too. I better be taking sm27 dang it
0
LeonardoWake up and smell the glaciersEagle River, AlaskaIcrontian
edited January 2006
Bud, the advice given above to you about setting up your 820D for Folding was good. Set it up right, and it should bury SM27. You should be getting 600-800 points a day with an 820D. Formula:
two Folding clients
each "Big Packets - Yes"
each -advmethods
I would not use these settings though, without a 1GB of RAM. The same settings also work beautifully for HT-enabled P4 machines.
BTW, check out your heatsink thread. I highly, highly recommend the Big Typhoon. BTW, if you ever decide to sell your 820, just let me know!
looks like i finally got my p4d setup right cause i finally gained some ground from sm27. I cheated though right now I have my 820d and a 930d at work, hahaha. I need to reinstall windows on my amd64 rig and just a mobo for my p4 3GHz and it should be over. But know bluknight is on my tail
Comments
It shows two active processors.
On top of that, there's the Vegemite factor.
BTW: It looks like you're about to make your move. See the "do great" link.
its actually powered by human blood pumped by an orphan heart. thats why its unstopable. not even your local priest can exercise sm27 out of its spot
If they broke, a few extra parts, a little know how, and an electrical storm (for effect) would bring them back to life. (insert manical laugh)
PS: I love snow, but I'm a little jealous nevertheless... :o
When someone tries to beat her steady run, the others chip in with moral support and do what they can to keep her climbing the folding ladder.
Just having a look at the stats and me thinks poor SM27 will have a hard time keeping up with the challenge.
If only by one point today.
I'd love to fine tune a few of the machine I have running, i know one of them estimated a Wu to take like 26 days once...that seemed a bit long no?
Checking the advmethods flag under Advanced Options in the new 5.04 Console helps some if you're an AMD man.
Personally, I don't worry about the "stinker" WU's. My reasoning is that Stanford wouldn't send them out if they weren't necessary. Who knows - some of the answers we're hoping to find might well lie in those low-point-yield Work Units. I just try and keep my rigs running at top efficiency and let the rest take care of itself.
Most of the QMD core WU's are worth 450 points. They do extremely well on P4's. With HT enabled it can crank them out in two days.
2 X 450 = 900
Plus, there's always the VEGEMITE!
two Folding clients
each "Big Packets - Yes"
each -advmethods
I would not use these settings though, without a 1GB of RAM. The same settings also work beautifully for HT-enabled P4 machines.
BTW, check out your heatsink thread. I highly, highly recommend the Big Typhoon. BTW, if you ever decide to sell your 820, just let me know!