I\'m not seeing the advertised trickle behavior on my system.

I started climateprediction.net about 3 days ago, my current CPU time is 65:37:22 on my HADCM3 model, my current timestep is 50556 out of 4147560, and I see only ONE trickle up on my BOINC manager messages tab and that occured almost two days ago. That\'s it.

I have one of my 2.8 ghz cores running my model 100% of the time.

Also, my BOINC manager\'s projects tab shows ZERO avg. work done.

However, my account webpage at climateprediction.net shows 259.2 total credit with a recent average credit of 24.82

1) I thought trickle UPs are supposed to occur much more frequently (every 10802 timesteps according to http://boinc-wiki.ath.cx/index.php?title=Climateprediction_Size_Table)

2) I thought trickle DOWNs were supposed to occur daily to update my BOINC manager/client.

Does anyone know what is happening? I apologize for my unfamiliarity here as I just joined, but all available doc indicates that what I\'m seeing is not normal.

Thanks,
Ed

---

Hi Ed, welcome to the forum.

I\'m not seeing the advertised trickle behavior on my system.

Yes you are! Your model has just trickled again:
http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=6665557
It\'s doing fine. It will trickle each model year on 5 Dec or as soon after that as you give it internet access and within boinc manager allow network activity.

I started climateprediction.net about 3 days ago, my current CPU time is 65:37:22 on my HADCM3 model, my current timestep is 50556 out of 4147560, and I see only ONE trickle up on my BOINC manager messages tab and that occured almost two days ago. That\'s it
I have one of my 2.8 ghz cores running my model 100% of the time..

I think your model should be able to run faster on your computer than what it\'s doing. Check in your Climate Prediction preferences that you\'ve allowed the model to use 100% CPU. Ie allow it 100% of one core. On a dual-core it would then be using 50% of the whole. If you allow it less, it won\'t progress as fast. You can also speed the model up a bit by disabling the screensaver and just watching the globe when you actually want to using the graphics button in boinc manager.

My dual-core is using 100% of both cores by running 2 models.

Also, my BOINC manager\'s projects tab shows ZERO avg. work done.

However, my account webpage at climateprediction.net shows 259.2 total credit with a recent average credit of 24.82

It can take the trickles a day or two to update and award the credits - the server only updates the stats once each 24 hours. But you always get the credits in the end.

2) I thought trickle DOWNs were supposed to occur daily to update my BOINC manager/client.

There are no trickle downs. Boinc doesn\'t need anything extra; it got everything it needed when you installed it. In a few months time you may think about updating to a new version of it, but your present version should be good for at least a year.

No trickle-downs from the project or model in Oxford either. Your model contains everything it needs to continue. But what you do keep receiving are the messages which are boinc\'s dialogue with the cpdn server in Oxford (and its own internal mutterings and ramblings).

Take a look at the project READMEs reachable thro my signature. For a cpdn newbie I\'d recommend

*in the one about running the model, the top tips

*in the one about avoiding model crashes, item #5 by Mike about avoiding model crashes, plus item #5 by Les about how to back up the model. It\'s important to do this regularly because if the model does crash, the only way to get it back and continue it is by restoring a backup. (Everybody crashes a model sooner or later....)

Hope that helps.

Mo

---

Cpdn news

Mo,

Thank you for taking the time to give such a detailed response.

Thanks for pointing out that my model will trickle every Dec 5 - somehow I missed that tidbit of information in the project doc. Now I know what to expect.

I\'m not sure that I can wring any more performance out of my machine short of doing drastic things I\'m not inclined to do (like overclocking). My preferences are set to the maximum - 100% of the CPU, and Windows Task Manager shows a solid 50% usage on both cores by the BOINC clients (core one is always CPDN, core two is either SETI or World Community Grid). It runs 24/7.

An hardware-specific, optimized app like the SETI \"Chicken\" app would be nice to have!

I only have the screensaver up when I choose to look at it via the \"show graphics\" button in BOINC manager.

I\'m studying the links you have in your signature for tips that I\'ve missed.

Ed

---

This is my trickle record Here but it appears to of lost a couple!

missing trickle
2007-07-18 13:47:19 [climateprediction.net] Sending scheduler request: To send trickle-up message
2007-07-18 13:47:19 [climateprediction.net] (not requesting new work or reporting completed tasks)
2007-07-18 13:47:23 [climateprediction.net] Scheduler RPC succeeded [server version 509]

recorded trickle
2007-07-18 20:08:43 [climateprediction.net] Sending scheduler request: To send trickle-up message
2007-07-18 20:08:43 [climateprediction.net] (not requesting new work or reporting completed tasks)
2007-07-18 20:08:48 [climateprediction.net] Scheduler RPC succeeded [server version 509]

missing trickle
2007-07-19 12:09:44 [climateprediction.net] Sending scheduler request: To send trickle-up message
2007-07-19 12:09:44 [climateprediction.net] (not requesting new work or reporting completed tasks)
2007-07-19 12:09:49 [climateprediction.net] Scheduler RPC succeeded [server version 509

recorded trickle
2007-07-20 15:52:18 [climateprediction.net] Sending scheduler request: To send trickle-up message
2007-07-20 15:52:18 [climateprediction.net] (not requesting new work or reporting completed tasks)
2007-07-20 15:52:23 [climateprediction.net] Scheduler RPC succeeded [server version 509]

awaiting trickle
2007-07-20 17:16:36 [climateprediction.net] Sending scheduler request: To send trickle-up message
2007-07-20 17:16:36 [climateprediction.net] (not requesting new work or reporting completed tasks)
2007-07-20 17:16:41 [climateprediction.net] Scheduler RPC succeeded [server version 509]

How do you figure? As of 1555Z, 20 July, they\'re all there. (There\'s one every 10,802 TS.)

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

How do you figure? As of 1555Z, 20 July, they\'re all there. (There\'s one every 10,802 TS.)

These are the only ones showing on my account.
20 Jul 2007 15:55:08 706688 6641627 hadsm3fub_0210_005893772_0 1 205,238 489,607 2.3856
18 Jul 2007 20:11:31 706688 6641627 hadsm3fub_0210_005893772_0 1 194,436 463,577 2.3842

Yet I sent 2 on the 18th, 1 on the 19th and a further 2 on the 20th! a total of 5 it just seems strange when all others are about 3 mins after they were sent.

What\'s the current timestep on your globe?

What\'s the current timestep on your globe?

timestep 213990 of 259248

Then all of your trickles are on the server, even if the server timestamp is a bit odd.
Perhaps the server was out at a protest meeting at the time (Too much work! Not enough storage space!), but there are times when it takes a while for a trickle to show up.
Your next trickle timestep will be 216040.

Happy crunching. :)

Thanks Les.

---

Crunching for Natures Team

I am certainly missing a timestep 75,614 here

I have trickled 3x today yet this timestep has still not appeared. Present state on globe is timestep 99,216. I know I have a mememory issue since loading a program I need for work but if this was a serious hiccup then an error would surely be showing.

Todays trickles.
2007-08-25 07:20:29 [climateprediction.net] Sending scheduler request: To send trickle-up message
2007-08-25 07:20:29 [climateprediction.net] (not requesting new work or reporting completed tasks)
2007-08-25 07:20:34 [climateprediction.net] Scheduler RPC succeeded [server version 509]
2007-08-25 08:35:58 [climateprediction.net] Sending scheduler request: To send trickle-up message
2007-08-25 08:35:58 [climateprediction.net] (not requesting new work or reporting completed tasks)
2007-08-25 08:36:03 [climateprediction.net] Scheduler RPC succeeded [server version 509]
2007-08-25 19:12:29 [climateprediction.net] Sending scheduler request: To send trickle-up message
2007-08-25 19:12:29 [climateprediction.net] (not requesting new work or reporting completed tasks)
2007-08-25 19:12:34 [climateprediction.net] Scheduler RPC succeeded [server version 509]

---

Crunching for Natures Team

It is odd, but that\'s a slab model, and there\'s no data in the trickles.

<variety>orig</variety>
<wu>2r71_200149856</wu>
<ph>2</ph>
<ts>248446</ts>
<cp>1694952</cp>

Do I continue, abort or load a backup? even though there are no errors showing.
I have backups for 21st, 22nd, 23rd and 24th August.

Does the missing data make the model invaluable to those that require it? Sorry to be a pain!

---

Crunching for Natures Team

Continue.

As I said, there\'s no data in trickles. These models are from the early days of the project, when ALL science data was collected at the end of a phase, zipped up, and sent in one big file. This philosophy changed with the introduction of the Coupled models in 2006, so that now small amounts of data are returned in each trickle.
The only thing that these slab trickles do, is say to the server: \"I\'m still alive, and I\'m now at this timestep\". If a trickle goes missing, then the next trickle will tell the server that the model is still alive.

And if you should be looking at the computer at the time that a phase ends, you\'ll see the cursor start to flicker. This will continue for a few minutes.
DON\'T PANIC. It\'s caused by the program collecting all the data to create the zip file, and then removing the now un-needed small files.
It\'s best to let the computer get on with it without you using it.
And DEFINITELY don\'t reboot in the middle of this, as some people have done! You\'ll lose the data!

Hi Martin
Just to add, this may well sort itself out in the early hours when the servers update.At the moment the slab has accumulated 51030.94 credits hopefully tomorrow will read 51030 +283 if it has.
Good luck
Chris.

Cheers Both

\"If you don\'t ask one never knows\"

Martin

---

Crunching for Natures Team

Done at last Here complete with a long list of messages.

---

Crunching for Natures Team

Well done!

---

Cpdn news

Good work, congratulations :-)

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

Whew! Well done, Martin! ;-)

Best regards, MM

---

Visit the Scotland team