I'm running BM 6.6.36 on a quad-core Intel machine under XP SP3. I'm attached to the following projects:

Climate Prediction, resource share 200 (50%)
Leiden Classical, resource share 100 (25%)
World Community Grid, resource share 100 (25%)

This configuration in earlier BM releases has always resulted in two CPDN tasks running and one each of Leiden and WCG.

But with 6.6.36, CPDN gets ALL FOUR CORES assigned (four tasks) and I get no new Leiden or WCG tasks. The messages from the server state that the tasks "won't finish in time". This is not true.

When I force Leiden or WCG tasks to download by suspending CPDN, when they finish the behavior always reverts back to what I described above.

Is the 200/100/100 resource allocation causing this behavior? Should I reset the profiles to 100/50/50 ? Or is something else going on?

Thanks.

Something else is going on. I suspect you're running into that line that says something alike "(won't finish in time) BOINC runs 93% of the time, computation enabled 100% of that."

It's the 100% that's causing this, but we can't get the correct information from the projects as to why it is doing this. It is something in the software on the server, though, while it's client version independent.

Now, for some people (including me), a reset of the project worked magic.

But before you do so, may I request you post some info from your sched_request*.xml files on the affected projects?

Go to your BOINC Data directory and open sched_request_www.worldcommunitygrid.org.xml, scroll down through it to underneath the <cross_project_id> tags. You'll find the time stats here.
Can you post the lines for <on_frac>, <connected_frac> and <active_frac>, please?

Close the sched_request file and open the sched_request_boinc.gorlaeus.net.xml file and do the same for this one.

After you posted that information you can reset these projects.

---

I suspect you're running into that line that says something alike "(won't finish in time) BOINC runs 93% of the time, computation enabled 100% of that."

It's the 100% that's causing this, but we can't get the correct information from the projects as to why it is doing this. It is something in the software on the server, though, while it's client version independent.

I have the same thing on Milky way (though it says 99.9% of that.)

One user reported:

I started to see that message when I upgraded BOINC to 6.6.36

I went back to 6.6.20 and all is well.

I reset the project, didn't help.

Hey there, Jord,

Yes, that is exactly the message I'm receiving, except the numbers are different:

"7/2/2009 1:48:24 PM Leiden Classical Message from server: (won't finish in time) Computer on 99.3% of time, BOINC on 97.8% of that, this project gets 25.0% of that"

I'll try resetting Leiden and WCG, but I cannot reset CPDN as I have four long-running models currently.

Here's the lines you requested from sched_request_www.worldcommunitygrid.org.xml:

<on_frac>0.993443</on_frac>
<connected_frac>0.995929</connected_frac>
<active_frac>0.977402</active_frac>

And from sched_request_boinc.gorlaeus.net.xml:

<on_frac>0.993494</on_frac>
<connected_frac>0.995961</connected_frac>
<active_frac>0.977579</active_frac>

Hope this helps.

Ed

I reset both WCG and Leiden. I received two WCG tasks but no Leiden tasks. I received the same message:

"7/3/2009 8:33:04 AM Leiden Classical Message from server: (won't finish in time) Computer on 99.4% of time, BOINC on 97.7% of that, this project gets 25.0% of that"

It could be that BM considers it now to be over-scheduled so I'll wait and see what happens when some of the CPDN tasks complete.

If it appears that something is still amiss I'll revert back to an earlier release of BM.

Hello. Just a few notes on the above. I am running Windows XP SP 3 and upgraded to BOINC 6.6.36 on 6-20-2009. I have WCG and Climateprediction running. After the upgrade, I received the same kinds of messages - and received no new work from WCG (for about 10 days).

I finally reinstalled BOINC 6.4.5 and received a project immediately. So, right now, I'll stick with 6.4.5 and wait until the next upgrade or two after 6.6.36. I contacted WCG about the problem, don't know what actions they can take.

S. Johnson

P.S. Looks like at least one other person went back to an earlier version.

Hello. Just a few notes on the above. I am running Windows XP SP 3 and upgraded to BOINC 6.6.36 on 6-20-2009. I have WCG and Climateprediction running. After the upgrade, I received the same kinds of messages - and received no new work from WCG (for about 10 days).

I finally reinstalled BOINC 6.4.5 and received a project immediately. So, right now, I'll stick with 6.4.5 and wait until the next upgrade or two after 6.6.36. I contacted WCG about the problem, don't know what actions they can take.

S. Johnson

P.S. Looks like at least one other person went back to an earlier version.

A better choice might have been 6.6.31 which is what I did.

---

Boinc V 7.4.36
Win7 i5 3.33G 4GB NVidia 470

Note: Received a response from World Community Grid. They are currently supporting the following versions of BOINC.

Windows (XP, Me, 2000): 6.2.28

Linux (x86): 6.2.15

(Apple) Mac (PowerPC, x86): 6.2.18

I'm running on WinXP SP3 on a Core2Duo laptop with 4GB RAM, and dedicate one processor to BOINC all the time. I belong to 13 projects. When I upgraded to 3.3.36 on Monday, 7/6, I had 9 tasks in progress. Because my laptop had been off for part of the weekend (it is usually on 24x7), the 3.3.36 scheduler started to refuse new work because my machine was only up ~50% of the time, based on its calculation. By Tuesday morning I had completed all of my tasks except for Climate Prediction, and couldn’t get any more. I downgraded to 6.6.20, and within 2 minutes had 9 more tasks across 8 projects to process. I would strongly suggest that the percentage uptime calculation be modified in the BOINC manager.

This problem with 6.6.36 isn't about percent on time. It affects laptops to 4 core units. 2% project percent to 50%. Various projects. Mine is Milkyway (short WU/short deadlines)

Appears that projects think available time/project percent > WU deadline and won't give any work (unless the cache is totally empty)

Previous version 6.6.28 gets work fine. I get 6 Milkyway, Boinc goes into EDF mode for 24 hours processing 5.5 of them and then returns normal.

I am in a similar situation -- I had a 40+ wu before the weekend covering 6 projects and now all I have left is AQUA and CPDN. The remaining 6 workunits (2 AQUA and 4 CPDN) all have expirations of Oct 2009 and later. I am running Ver 6.6.36 on Mac Pro (2X 2.66 Dual Core Xeon) which is up 24/7. It looks like to me that the scheduler is giving too much weight to completing long term work units in the short term which is not allowing shorter wu's to run. (JUST A GUESS)

heres a clip of my scheduler messages

Mon Jul 20 07:20:48 2009 SETI@home Message from server: No work sent
Mon Jul 20 07:20:48 2009 SETI@home Message from server: No work is available for Astropulse v5
Mon Jul 20 07:20:48 2009 SETI@home Message from server: (won't finish in time) BOINC runs 99.9% of time, computation enabled 97.0% of that
Mon Jul 20 07:20:53 2009 Milkyway@home Sending scheduler request: To fetch work.
Mon Jul 20 07:20:53 2009 Milkyway@home Requesting new tasks
Mon Jul 20 07:20:58 2009 Milkyway@home Scheduler request completed: got 0 new tasks
Mon Jul 20 07:20:58 2009 Milkyway@home Message from server: No work sent
Mon Jul 20 07:20:58 2009 Milkyway@home Message from server: (won't finish in time) BOINC runs 99.9% of time, computation enabled 97.0% of that
Mon Jul 20 07:22:04 2009 Milkyway@home Sending scheduler request: To fetch work.
Mon Jul 20 07:22:04 2009 Milkyway@home Requesting new tasks
Mon Jul 20 07:22:09 2009 Milkyway@home Scheduler request completed: got 0 new tasks
Mon Jul 20 07:22:09 2009 Milkyway@home Message from server: No work sent
Mon Jul 20 07:22:09 2009 Milkyway@home Message from server: (won't finish in time) BOINC runs 99.9% of time, computation enabled 97.0% of that
Mon Jul 20 07:31:16 2009 Milkyway@home Sending scheduler request: To fetch work.
Mon Jul 20 07:31:16 2009 Milkyway@home Requesting new tasks
Mon Jul 20 07:31:21 2009 Milkyway@home Scheduler request completed: got 0 new tasks
Mon Jul 20 07:31:21 2009 Milkyway@home Message from server: No work sent
Mon Jul 20 07:31:21 2009 Milkyway@home Message from server: (won't finish in time) BOINC runs 99.9% of time, computation enabled 97.0% of that
Mon Jul 20 07:33:43 2009 Einstein@Home update requested by user
Mon Jul 20 07:33:46 2009 Einstein@Home Sending scheduler request: Requested by user.
Mon Jul 20 07:33:46 2009 Einstein@Home Requesting new tasks
Mon Jul 20 07:33:51 2009 Einstein@Home Scheduler request completed: got 0 new tasks
Mon Jul 20 07:33:51 2009 Einstein@Home Message from server: No work sent
Mon Jul 20 07:33:51 2009 Einstein@Home Message from server: (won't finish in time) BOINC runs 99.9% of time, computation enabled 97.0% of that


Thanks

my p4 boinc 6.6.37 started doing this just recently too.... like today...i come home and just went over to look at it and see what was running and the project throwing that message is dd@H.

Boinc V6.6.36 and .37 both have a bug in the way it reports how much work you have on your system. This causes the project to reject work requests. I've had problems getting work from low RS projects and even a project with nearly a 50% resource share. I only have a 1.5 day cache so this is not factor, I opted to revert to .31 which was the last semi stable version I've used.

---

Boinc V 7.4.36
Win7 i5 3.33G 4GB NVidia 470

i'll go to .38 on that pc..., no bugs in .38 other than a gpu request on projects with no gpu app out...

i'll go to .38 on that pc..., no bugs in .38 other than a gpu request on projects with no gpu app out...

Yet again - every BOINC V6.6 has always asked for GPU work from every available project, just in case a GPU application may have been deployed since last time.

v6.6.38 has a bug fix that puts information back into the message log so you can see what it's doing.