• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

DC Vault 2

Keep an eye on those VM's. Two issues I run into is: a) Virtualbox crashing and needing started again b) DIMES client crashing without any indication and needing started again. I basically just check my daily numbers and look for any that haven't reported for the day.
 
Woohoo - I moved up to 15th place on the team! And I will hold that position for exactly TWO DAYS until Grandpa mows me and I'm back to 16th... :)
 
VMs seem to have run fine overnight, using a utility called VBoxHeadlessTray to keep the VMs in the tray, seems to be working OK so far.
 
Hit my one year goal on WCG/MCM, moving on to Cosmology for January...
 
Even though POGS is worth more points and we are tearing it up right now, I am switching everything over to Cosmology since that was the majority vote for projects focus this month. Perhaps POGS next month.
 
I had trouble running Cosmology on my main rig, constant computation errors.
I'll see what I can switch over to it, the units seem to have a pretty long runtime which isn't ideal for some of the systems I run BOINC on.
 
ABC@Home's website can't be reached. They let the domain expire. Hopefully they get this resolved in short order.
 
After a few days, I've been seeing some random issues like CaptainUnlikely - for me, I think it has to do with not having enough memory to run 4 WU at once on a machine with only 2GB RAM... Knocked it down to 2 cores Cosmology and 2 cores Asteroids to see if it fixes anything. Will report back after a few full WU runs.
 
Last edited:
I see, that would explain why my Atom server with 2GB hasn't been doing so well with Cosmology, in fact right now it's only using 3/4 cores. I think I'll drop it down to use 2 instead, and see how it goes.
Still not sure about my main rig though as it has 8GB, and >4GB free, so it shouldn't be due to a lack of RAM. Also, it was constant on this rig, not just some WUs but all of them, whereas the other machines have had no problems. Would be interesting to see what happens with yours.
 
Captain, is this the troublesome PC? http://www.cosmologyathome.org/show_host_detail.php?hostid=204820

The error units say there were too many exits. Are these tasks being suspended a lot? Also, is this rig running a bunch of VM's for DIMES?

Check your BOINC manager settings. Things that can cause BOINC to pause are where I would begin. When a work unit pauses too many times or repeatedly in a very small amount of time, the WU can error out.

Some of the other work units that failed indicate possibly a disk issue. If this machine is running a bunch of VM's for DIMES, stop them for a while and see if the problem clears up. You may be over taxing your HD.

Also, check your virtual memory size. One of your errors had this listed in it: forrtl: severe (41): insufficient virtual memory

Jathanis, is this the troublesome PC? http://www.cosmologyathome.org/show_host_detail.php?hostid=179979

You look to have a very similar issue to some of Captains. I'm seeing insufficient virtual memory on yours as well. Double check the virtual memory on your system and check to see what settings you are using in BOINC. I would definitely take a look at your "leave application in memory" section and see how much memory you are allocating when the system is in use. If your virtual memory isn't large enough when the RAM goes to dump the tasks in VM, it may error out. I noticed this system has 6GB of RAM total in it.

According to WUProp, the work units use ~652.5MB each under Windows. http://wuprop.boinc-af.org/results/projet.py?projet=Cosmology@Home&application=CAMB
 
Last edited:
Interesting. I believe these issues were before I started loading up DIMES VMs. Virtual memory is set to auto, and nothing should pause BOINC apart from me suspending it, it's set to run while CPU usage is <85% which I've never seen my usage get up to.
DIMES VMs are now on a separate SSD to my main SSD, and right now I have 3GB RAM free with a whole bunch of stuff open, as well as my VMs and some POGS units.

I'll try shutting down those VMs and see if it helps, thanks.
 
Well, if they are on a different disk I wouldn't think they would be taxing the BOINC work in that regard. I would make sure your Virtual Memory is at least as large as your RAM. I normally have it much larger in case I get hit with multiple RNA work units at a time unexpectedly. Those work units can be 4GB of RAM or more a piece at times. What is your setting for LAIM set to? How much memory is setup for when system is in use?

The other thing to try is to set BOINC to always run and see if the problem continues with the errored wu's that state too many exits. I'm guessing something was causing BOINC to stop. Possibly even an anti virus or something like that could do it.You didn't have many of those errors. I mostly saw the ones mentioning Virtual Memory.
 
LAIM is unchecked, memory usage is set to use up to 75% when in use and 90% when idle, so 6GB/7.2GB respectively. I also don't have scheduled AV scans, I do those manually and pause BOINC myself when doing that or gaming to avoid too many exits, after you educated me about it before :)
It's 18% through one Cosmology unit and 12% through two more right now (leaving a core free for Milkyway@Home as it knocks off a significant amount of time per unit even though it only uses 2-3% of a core sporadically) so I'll crunch a few and see how it goes, if they don't error out I'll bring back some VMs one at a time and see how it goes. I also changed my page file to a larger, manual size, just to be sure. It's not like I don't have the free space.
 
I would turn LAIM on or check it. If your system does PAUSE, you will lose your work all the way back to the last check point. This can also cause the issue with it erroring with too many exits. However, the biggest question would be why it is pausing or exiting to begin with. Since most of your issues seemed to be related to virtual memory, I would just keep an eye out for the changes you have already made for those errors.
 
I just added 32 more threads of Intel goodness to Cosmology.........
 
DooKey said:
I just added 32 more threads of Intel goodness to Cosmology.........
Click to expand...

Yeah...I accidentally had YoYo attached still and after suspending POGS, most of my systems pulled work from YoYo. So, once those units get sent in, I should be mostly on Cosmology.

How many threads do you have total? Thanks to Kai, I should have a few more coming online within the next week or so. That is of course if UPS isn't totally shut down due to whether.
 
I'm interested in YoYo again sometime soon. I could probably knock a lot of work out there for us.
 
YoYo has a new sub project that I haven't earned the first badge yet for. So, I need to get there soon to at least keep my collection going. :)
 
DooKey currently has the two top host positions at the moment in Cosmology.

I agree, either a YoYo or perhaps POGS for the next month. Sounds like a plan Stan!!
 
POGS has the greatest potential for Vault points. I would suggest hitting that before a push at YoYo.
 
Cosmology is odd...hasn't errored out on me again yet, but one unit was 20% complete after 25 minutes or so, and now after just over 5 and a half hours it's at 87%. The other two units have near identical run times and are at 51% and 47%. Not too sure if that's normal, but I'll check up on them after the units finally finish and make sure I get credit properly. I've seen various run times from my gf's laptop, from 12 hours to 20 hours per unit and many in between. Is that normal? Looks like I've been granted 420 credit per unit which seems a little unfair.
 
I think they go with a flat point per work unit system. Faster systems finish more work units and therefore earn more points. I agree at this project that system isn't fair because the work units do vary. And yes, what you are seeing is correct. Typically this type of rewards system is used when work units are approximately the same size.
 
Gilthanis said:
Yeah...I accidentally had YoYo attached still and after suspending POGS, most of my systems pulled work from YoYo. So, once those units get sent in, I should be mostly on Cosmology.
Click to expand...

Funny you mention that - somehow my i7 had Primegrid still on with 0 Resource Share that pulled 8 WUs, one for each thread. The problem is that they are Genefer WUs, and by the time I noticed the lack of production today, they are currently at 27.6% to 35.1% - at 66 to 80 hours of run time! Seeing as the average for Genefer WUs is around 200 hours, I'll be doubling my Cosmology work about this time next week... :rolleyes:
 
Yeah..PrimeGrid is still on a few boxes here too. I just let them run. I get tired of changing profile settings around all of the time. And since these boxes have GPU's as well, I just let them run.
 
Looks like WUProp is down today - can't access their webpage and WUs have stopped uploading...
 
I noticed the uploads issue. I didn't think anything of it since there is usually an hour or so each night that they server is down for updates/maintenance. Hopefully they get it up quick so that we don't lose run time towards badges. ;)
 
Gilthanis said:
I think they go with a flat point per work unit system. Faster systems finish more work units and therefore earn more points. I agree at this project that system isn't fair because the work units do vary. And yes, what you are seeing is correct. Typically this type of rewards system is used when work units are approximately the same size.
Click to expand...

Ah I see, at least it's not my system this time then, I just got concerned that they seemed to be progressing fast at first and then slowed way down. Seems like a crazy points system - I guess if everyone gets a mixed bag of units then it's "fair" in the sense that two identical systems should average the same credit per day, but it does seem really odd that an 8 hour and 16 hour unit are both worth the same amount of credit.
 
I agree Captain, the points system isn't fair in that regard. Unfortunately, there isn't much of a response from the admins any more and I doubt you would get much productive discussion out of bringing it up over there. It is a shame really.
 
Yeah, I've been on the forums there just to look around, seems quite dead and there's definitely not much input from those involved with the project, and not much discussion at all really. Like I say though, everyone is in the same boat really, so it's not a huge deal, I just wanted to make sure it wasn't my setup being weird.
 
Gilthanis said:
ABC@Home's website can't be reached. They let the domain expire. Hopefully they get this resolved in short order.
Click to expand...

ABC's domain is good again. Now if only they could get work units on a more consistent manner. :rolleyes:
 
Good job on Cosmology you guys. We have made some pretty good progress and still have a good amount of time to go.
 
Cosmology is still being a bit flaky for me...yesterday or the day before I had a bunch of units error out with "too many exits" - there's literally nothing that can pause BOINC right now except from me, and I was at work so I can't blame myself :p
It's been mostly alright though and apart from this one blip, since making changes to my page file etc it's been alright. No problems on any of my other crunchers either.
 
Try to change the setting under Activity to Run Always and see if the behavior continues...
 
Already have :) like I say, there's nothing at all that should stop BOINC from running. My GPU projects, Collatz and Milkyway, had continued on as normal, too.
 
You must log in or register to reply here.
Back
Top