FR Folding@Home Project Update - We're in the Top 190 teams (A Tribute to Ronald Reagan)

soccer_maniac ^ | soccer_maniac

[soccer_maniac]

Time for a new FreeRepublic folding@home thread.

Our FreeRepublic team of 250+ members comprised primarily of Free Republic members in good standing have banded together to donate their excess CPU cycles to a worthy cause. Via distributed computing, millions of computers around the world, contribute directly to scientific research, in the quest for a greater understanding of diseases such as Alzheimer's, Cancer, and Mad Cow (BSE).

Currently, the team is in 188th place (with 940 CPUs - 15,725 completed Work Units and 2,330,000 points)

This is an entirely voluntary program, and if you want to learn more, please see the links posted below (or read one of the previous 8 folding threads)

[soccer_maniac]

That's worth a cool 404 points

See: http://fah-web.stanford.edu/psummary.html

[soccer_maniac]

3PM update - 8,083 points (35,487 points for today so far)

[abner]

p2107 20000 frame it'll never be finished bump.

You are not alone... That one just started on this machine... 148 minutes per frame??? That can't be right, can it?

[Stentor]

148 minutes per frame??? That can't be right, can it?

Nope. It'll take a while to sort itself out, I think. Mine is 20 seconds per frame on a 3gig pentium.

[ahayes]

I don't think frames are always the same size. I had a 10,000 frame protein that was folding at 1 frame/22 s. Then I've had 500 frame proteins folding at 1 frame/3.5 min. Once it benchmarks properly it should settle down at a more reasonable value of probably 3-7 minutes.

[Petronski]

The salad days of quickie 153-point folds are over I think. Both my CPUs have been issued GROMAC 2107s. Sure, they're worth 404 points, but the seem to take a week to calculate.

I think I'll be slipping into the 60s or 70s in the rankings.



It's all good though....except for the competitiveness factors.

[soccer_maniac]

A 404-pointer should take exactly 2.64 times a 153-pointer.

Points are benchmarked against standard performance, so all thing being equal your CPU should turn in the same avg point per day over time.

[abner]

Right now it says it will finish on Sunday, October 23, 2011...
That is a bit discouraging for 404 points. LOL!

[abner]

Ok, it just changed to something close to reasonable. . . Now down to 21 seconds per frame and says it will end on Saturday, March 4th, 2006.. . Whew.

[dfwddr]

Monday afternoon "Hey y'all, come check this out and join us" bump.

[Petronski]

I will be testing this assertion. I hope you're right.

[houeto]

Both my CPUs have been issued GROMAC 2107s.

Ditto. Several in the past week of so.

[HangThemHigh]

You'd think so, but I had a project 2107 (404 points) that took only seemed to take a little longer than a project 2106 (202 points). Wish I had a record of the exact times.

[BreitbartSentMe]

You can go looking back through your log, and see how much time each of them took, IIRC.

[HangThemHigh]

I just learned something, some of you may know, but I just found out and figured it's knowledge gained the hard way and worth posting. I am sure this applies to the console version, which I use, but it likely applies to the GUI version, as well.

I recently had a power outage and had to restart. On reboot, I got the dread "checksums don't match" error and the project restarted from the beginning (which wasn't too bad as it was in the early stages).

But, after the reboot, I noticed progress was painfully slow - it was taking an hour to complete 1%. On checking the logs, I noticed the line "Extra SSE Boost OK" I usually see in the log was missing.

I rebooted and, this time, got the SSE boost and the time to complete 1% dropped to 25 min. The SSE boost more than doubled the folding speed for this problem.

Reading a few logs I learned that after a bad checksum, the core assumes the problem may have been with the SSE boost and disables it. To get it back, you have to reboot again (as I did) or use the -forceasm flag after the power loss to get the current problem back up to full speed.

I assume that if you let a problem run to completion without the SSE boost, the next problem would have it applied automatically, but I'm not sure on this.

Bottom line, if a problem is running slowly you might want to look for the "Extra SSE Boost OK" in the log, especially after a power loss or blue screen reboot. If you normally see this in the log, a reboot or the -forceasm flag may give a big help.

[HangThemHigh]

You recall correctly, but when I tried that, it was too late and the portion of the log referring to the 2107 project was already gone (too little, too late should be my tagline)

[SC Swamp Fox]

[SirTaurus]

#279 got his WU sir

[SC Swamp Fox]

All three of my CPUs are now working on 2107, my two slow ones are choking. They are going to take 10 to 15 days.

[HangThemHigh]

Ouch, 404 points in 15 days is less than 30 points per day.

I don't know what your "slow" machines normally score, but I'd be tempted to try to give that problem to a better machine and get different problems for the slower guys. Not sure how hard that would be though.