Focusing on World Community Grid

[Smartazz]

My rig completed about a dozen work units then didn't send them in nor did it get more work units so it sat idle most of the night. Any idea how to keep this from happening?

 

[somethingsketchy]

My rig completed about a dozen work units then didn't send them in nor did it get more work units so it sat idle most of the night. Any idea how to keep this from happening?

Do you have the network setting in BOINC to "Network Always Available"? That would eliminate any variables about daily WU download/upload limits.

 

[Smartazz]

Do you have the network setting in BOINC to "Network Always Available"? That would eliminate any variables about daily WU download/upload limits.

It wasn't set to that. Thanks for the advice. Hopefully this fixes it.

 

[ZipSpeed]

Top 20 producers based on yesterday's data:

Looking good for overtakes! Chance to move up 9 positions in ~1 month.

 

[ZipSpeed]

Courtesy of SETI.GERMANY, they have a few neat tools under their "Challenge Gifts".

Here is "All badges of the members of TeAm AnandTech"

http://www.seti-germany.de/wcg/team_badges.php?team=SHNW8R84S1

This one lets you display your badges on one picture.

http://www.seti-germany.de/wcg/show_badge_picture_bild.php

My badge:

 

[petrusbroder]

Not bad at all!

 

[GLeeM]

Do you have the network setting in BOINC to "Network Always Available"? That would eliminate any variables about daily WU download/upload limits.

Hey Smartazz - let me know if this works.

I have the same problem on my main cruncher. I did a Reset and then a Remove. If I remember right the Remove helped for a couple results
One time a reboot got a couple results sent in.

 

[Smartazz]

Hey Smartazz - let me know if this works.

I have the same problem on my main cruncher. I did a Reset and then a Remove. If I remember right the Remove helped for a couple results
One time a reboot got a couple results sent in.

Haven't had the problem since so I think it worked.

 

[Uppsala9496]

For some odd reason, I have been getting computation errors on one of my machines the past couple of days. I've restart, detached, uninstalled, reinstalled, etc. Ran mprime overnight and it came back with no errors. I even scaled back on the overclock some.

For the life of me, I can't figure out why I'm suddenly getting errors. This same machine ran SIMAP error free earlier in the month. It's not a temperature issue. Never gets above 50*C.

I'm at wits end why WCG suddenly hates this machine.

 

[ZipSpeed]

For some odd reason, I have been getting computation errors on one of my machines the past couple of days. I've restart, detached, uninstalled, reinstalled, etc. Ran mprime overnight and it came back with no errors. I even scaled back on the overclock some.

For the life of me, I can't figure out why I'm suddenly getting errors. This same machine ran SIMAP error free earlier in the month. It's not a temperature issue. Never gets above 50*C.

I'm at wits end why WCG suddenly hates this machine.

Has it just been a few errors or a page full of errors? I've hit a few errors myself over the last few weeks but from what I gather, they are errors on the WCG end and not the user end. All the errors I got have zero or small amount of runtime. Got quite a few server aborted WUs too.

Maybe run Memtest and see if it's the RAM.

 

[blckgrffn]

I currently have 0 errors, 0 invalid, and 1 inconclusive task over the last week+.

I am only doing GO Fight Malaria and Clean Energy.

 

[Uppsala9496]

Running help fight cancer and fight childhood cancer on all 5 machines. Only the one has issues. And it is page after page of them.

Here is an example:

X0910058060147200509231214_ 0-- Cassandra Error 2/8/12 06:41:41 2/8/12 13:38:17 1.28 41.8 / 0.0
X0910058060151200509231214_ 0-- Cassandra Error 2/8/12 06:41:41 2/8/12 13:38:17 1.28 42.1 / 0.0

I'll keep it running the remainder of the day to see what happens. If it continues, I'll move that machine on to correlizer or something.

That particular machine doesn't have a lot of memory. Only 4 GB of some cheap ddr3. Hmmm, might have to run memtest.

 

[Rattledagger]

Running help fight cancer and fight childhood cancer on all 5 machines. Only the one has issues. And it is page after page of them.

Here is an example:

X0910058060147200509231214_ 0-- Cassandra Error 2/8/12 06:41:41 2/8/12 13:38:17 1.28 41.8 / 0.0
X0910058060151200509231214_ 0-- Cassandra Error 2/8/12 06:41:41 2/8/12 13:38:17 1.28 42.1 / 0.0

I'll keep it running the remainder of the day to see what happens. If it continues, I'll move that machine on to correlizer or something.

That particular machine doesn't have a lot of memory. Only 4 GB of some cheap ddr3. Hmmm, might have to run memtest.

Hmm, I don't for the moment remember any specific problems with HCC, if it had been HPF2 problems with some wu's is common but AFAIK not HCC...

So, just some general things to try:
For virus-scanners, disable scanning of the boinc data-directory including sub-directories.
Try uninstall & re-install of BOINC-client (just in case it's a permission-error).
Since the computer is overclocked, try with default speeds instead and see if this fixes the problems. Since different applications uses different parts of a cpu, and rest of the computer, being "stable" with application X and Y doesn't mean application Z won't use the "wrong" part of the system and this part happens to be the weakest in the system so shows it's not stable afterall...

 

[Uppsala9496]

memtest has cleared pass 1. Going to let it run for another 4 passes.
Was fine running mprime last night (some 8 hours).
I detached and uninstalled boinc (running ubuntu 11.10 on this machine) and did a complete removal of the data. Installed boinc fresh, and the two errors I posted are after the installation.

This machine was error free in the past running WCG. Only difference is a new motherboard (the silly GA-78lmt-sp2 board, except this is rev 4.0 and before it was rev3.1).

I'll run mprime again after the memtest completion and let it go for 12 hours.
After that if I get any WCG errors, I'll just move the machine to a different project. It's a X2 555BE with 4 cores unlocked. Was running at 3.6 and now down to 3.5.

 

[VirtualLarry]

This machine was error free in the past running WCG. Only difference is a new motherboard (the silly GA-78lmt-sp2 board, except this is rev 4.0 and before it was rev3.1).

I'll run mprime again after the memtest completion and let it go for 12 hours.
After that if I get any WCG errors, I'll just move the machine to a different project. It's a X2 555BE with 4 cores unlocked. Was running at 3.6 and now down to 3.5.

That might be it. That board is only good for 95W CPUs. Might be overloading the VRMs, and they deliver unstable power to the CPU, when you have the CPU both unlocked, and overclocked, AND most importantly, you are running 100% load on the CPU.

 

[Uppsala9496]

The 555 is a callisto chip, so it's only 80w.
Passed 8 passes of memtest. Going to give mprime some 12-13 hours to run.

 

[GLeeM]

I used my sons laptop to get over 250,000. I installed the WCG customized BOINC version 6.10.58 downloaded from the WCG site. It's an old version but at least it worked

 

[Uppsala9496]

Well, I haven't had any errors in the past 24+ hours on the machine that was flaking out.
I didn't change a single thing either. Just bizarre.

 

[ZipSpeed]

Good to hear Uppsala. Finally got rid of that poltergeist!

I got a bunch of errors myself when I checked this morning. If anyone else is getting errors on Go Fight Malaria, looks like there's a bad batch of WUs.

From the WCG forum:

Re: GFAM_ x3lt4_ PfENR units all erroring out
We have stopped sending out new work for GFAM.

It appears as though the target sent with the work units does not match the expected name in the work unit file. We will keep you updated as things occur.

Thank you for your patience,
-Uplinger

 

[blckgrffn]

Good to hear Uppsala. Finally got rid of that poltergeist!

I got a bunch of errors myself when I checked this morning. If anyone else is getting errors on Go Fight Malaria, looks like there's a bad batch of WUs.

From the WCG forum:

Re: GFAM_ x3lt4_ PfENR units all erroring out
We have stopped sending out new work for GFAM.

It appears as though the target sent with the work units does not match the expected name in the work unit file. We will keep you updated as things occur.

Thank you for your patience,
-Uplinger

Haha, that's a lot of projects I haven't seen before cluttering up my BOINC task manager

Good thing I have that "send me other WU's in case what I have selected is not available" or I would be nearly idling along...

 

[somethingsketchy]

Haha, that's a lot of projects I haven't seen before cluttering up my BOINC task manager

Good thing I have that "send me other WU's in case what I have selected is not available" or I would be nearly idling along...

So that explains why I've been getting a lot of the HFCC WUs. I haven't seen any of the GFAM in quite a while, though I didn't notice any of the WUs erring out.

 

[Smartazz]

Weird, I didn't notice any errors and my results log doesn't report any rejected WUs.

 

[biodoc]

Added 1/2 my cores to the cause.

 

[ZipSpeed]

Added 1/2 my cores to the cause.

Very nice! Thanks for the addition. :biggrin:

 

[ZipSpeed]

All expectations have been surpassed. So much for getting into the top 100 by summer. At this rate, we might get there before spring! Well done and crunch on!

Top 20 producers based on yesterday's data: