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)

[Paladin2]

Average points/unit are rising (check team history at fahstats). When the WUs that are being handed out suddenly shift to being bigger and taking longer to finish, one would expect the point rate to drop off until many of the team's computers have digested a couple and get spread out from each other in completing WUs. Then the rate should come back.

There are a number of FAH servers that are rejecting connections so that completed WUs can not be returned. I think I'm up to three of those (two for more than a week, one for today).

[FreeAtlanta]

Sounds like you really have something up your sleeve. You might pass me before I make it to 100.

[processing please hold]

I lurk here several times a day. When they start talking about computers it's like they're speaking Chinese. Goes right over my head. But I'm still chugging along. Slooooooowly but surely.

[soccer_maniac]

Our average is 55k, a 15% variation is 8,250.

We had 68,842 on 2/28 and 44,710 2 days later. That's a 35% swing.

[ahayes]

If the major players are adding computers in batches it's likely output will surge and fall periodically. Wait a week and see what the new pattern looks like.

[DocRock]

Friday PM bookmark

[abner]

I was a little nervous too about the watercooled thing... But, it is the best computer I have ever had. It is awesome and very speedy. Quiet as a mouse too. :-)

[FreeAtlanta]

Are you two using the console/process or the screen saver?

[FreeAtlanta]

One of the guys said in an earlier thread that it seems that we have just wrapped up a bunch of moderately sized projects and appear to be loading a bunch of really big ones. I know I got a huge one early today, but I haven't checked my other downloads.

His (or, maybe hers) theory is that we will see our numbers come back dramatically as these new job start finishing, and that everything should smooth out again as everyone starts processing these big puppies.

[texas booster]

Last week I was posting 1,000 point days, very unusual for me. Now every system is at least a day from completing.

We were producing lots of points for about a week. Lets see what happens over the next week. I expect that we may have our first 70,000 point day.

[texas booster]

I just check several other teams and all are showing a decline in production. It may be related to a server issue or just to the size of work units being handed out.

Even DU has dropped in production. Imagine that.

[ahayes]

I'm using the graphic version but it does not run as a screen saver? It sits there in the system tray unless I pull it up to look at the protein. I just turn off the monitor and leave the computer on to fold.

[FreeAtlanta]

Has anyone determined if there is much of a speed difference?

I tried the graphical version and thought the pictures were cool, but then switched all my processing to the console (I had to do it on my dual to use both CPUs). I found it was so easy, that I put it on my other processors. I like that if my workstation reboots, I am pretty sure the process gets started and does it's work. I might be wrong. I will do some checking.

[Clara Lou]

I just downloaded a Gromacs p2106. Below is a copy of my log. I keep getting the "timered checkpoint triggerd" entry. It doesn't seem to be processing beyond step 48790. Would you know anything about this. I've done17 other units, and none has ever behaved like this one. Does the checkpoint message mean that there's no folding going on?

[00:19:48]

---

[00:19:48] Folding@Home Gromacs Core
[00:19:48] Version 1.86 (August 28, 2005)
[00:19:48]
[00:19:48] Preparing to commence simulation
[00:19:48] - Assembly optimizations manually forced on.
[00:19:48] - Not checking prior termination.
[00:19:48] - Expanded 79074 -> 419621 (decompressed 530.6 percent)
[00:19:48]
[00:19:48] Project: 2106 (Run 79, Clone 12, Gen 3)
[00:19:48]
[00:19:48] Assembly optimizations on if available.
[00:19:48] Entering M.D.
[00:20:08] (Starting from checkpoint)
[00:20:08] Protein: p2106_lambda_5way_melt_4_10011
[00:20:08]
[00:20:08] Writing local files
[00:20:08] Using table 3
[00:20:08] Completed 48790 out of 10000000 steps (0)
[00:20:09] Testing CPU type...
[00:20:09] Done testing.
[00:20:09] Extra AltiVec boost OK.
[00:26:09] Timered checkpoint triggered.
[00:32:09] Timered checkpoint triggered.
[00:38:09] Timered checkpoint triggered.
[00:44:08] Timered checkpoint triggered.
[00:50:08] Timered checkpoint triggered.

[Clara Lou]

Would you please read my post number 152 and see if you know anything about "timered checkpoint triggered"?

[abner]

Screen saver version. Does it make a difference?

At work- projects working on--- 1. 404 points, 2. 404 points

At home- projects working on--- 1. 404 points, 2. 153 points

Today I talked to a friend of mine who has several CPUs under his command. He is doing an experiment on one of his systems to see what it does to his computer. If it gives him good results, he might put some more under my handle. :-)

He said a few years ago, he did some SETI stuff and it wreaked havoc on the systems it was running on. Hopefully, he will contribute if this is better.

[Malsua]

Clara, checkpointing doesn't necessarily mean no folding is occuring. Do you have it set for 6 minutes?

It's entirely possible that you've got such a bear, and those 5waymelts are that it's taking a long time and checkpointing before hitting the next reporting step.

Every core reports slightly differently but a recent log for one of those looks like this to me.

---

Folding@Home Gromacs Core
[15:33:19] Version 1.86 (August 28, 2005)
[15:33:19]
[15:33:19] Preparing to commence simulation
[15:33:19] - Looking at optimizations...
[15:33:19] - Created dyn
[15:33:19] - Files status OK
[15:33:19] - Expanded 76362 -> 419621 (decompressed 549.5 percent)
[15:33:19] - Starting from initial work packet
[15:33:19]
[15:33:19] Project: 2106 (Run 93, Clone 23, Gen 0)
[15:33:19]
[15:33:19] Assembly optimizations on if available.
[15:33:19] Entering M.D.
[15:33:25] Protein: p2106_lambda_5way_melt_4_10011
[15:33:25]
[15:33:25] Writing local files
[15:33:25] Using table 3
[15:33:25] Extra SSE boost OK.
[15:33:25] Writing local files
[15:33:25] Completed 0 out of 10000000 steps (0)
[15:54:47] Writing local files
[15:54:47] Completed 100000 out of 10000000 steps (1)
[16:16:13] Writing local files
[16:16:13] Completed 200000 out of 10000000 steps (2)

---

Notice, it's 22 minutes for 100,000 steps. This is on an AMD 64 4000+ cpu with 2gb ram. In other words, it's a pretty hefty burden and if you're running much less CPU, you could take 2-3 times to finish the same unit and with checkpointing at 6 minutes, it's just reporting in.

[Drango]

Weekend folding bump

[ahayes]

The timered checkpoint is a backup save frequency that you can set in the configuration menu. It probably doesn't seem to have progressed in the log file because it is such a large job that it won't report the progress until probably 100,000 steps are completed. It's only been about 30 min in that log, I would give it another hour and check on it. My computer is running the corresponding p2107 job with 20,000,000 steps reporting every 200,000 steps in the log file about every 20 minutes. Your computer is saving the progress (see your beginning step value in the log file showing you'd restarted the program and it picked up folding that job at ~49,000) but just not displaying it to you.

These are whopping huge WU, mine won't wrap up for at least 3 more days.

[Malsua]

I also noticed something mine reports that your's does not.

[15:33:25] Extra SSE boost OK.

What CPU are you running? If it doesn't have SSE or it's disabled for some reason(can occur if a protein errors out), completion times are going to be much longer than my 22 minutes.