Associate
Ah interesting stuff 
DC Vault
- Thread starter hanluc
- Start date
DIMES stats go up and down like the proverbial yo-yo !
In June, Ido said the server had crashed and they had to move it onto a new one and the Vault stats haven't updated since then, though Free-dc did figure out what was needed to get it working and report them again.
Usually the stats on the local PCs keep track and when the server gets rebooted everything gets updated again. Friday's crash looks like the June one, stats on the local PC aren't incrementing so it looks like the backend has gone awol again
In June, Ido said the server had crashed and they had to move it onto a new one and the Vault stats haven't updated since then, though Free-dc did figure out what was needed to get it working and report them again.
Usually the stats on the local PCs keep track and when the server gets rebooted everything gets updated again. Friday's crash looks like the June one, stats on the local PC aren't incrementing so it looks like the backend has gone awol again
Thanks, I realised others contacted him so I stopped adding to his inbox, call it patience or laziness on my part
Rusty has posted this at the Vault today...
http://www.mersenne.org/report_top_5...&B1=Get+Report
http://www.netdimes.org/data.php?getAll=GroupsByMeas
GIMPS and DIMES are stuffed because the stats link is stuffed.
I think Dimes needs to be this...
http://www.netdimes.org/new/data.php?getAll=GroupsByMeas
I don't have a Vault account so haven't posted it there.
http://www.mersenne.org/report_top_5...&B1=Get+Report
http://www.netdimes.org/data.php?getAll=GroupsByMeas
GIMPS and DIMES are stuffed because the stats link is stuffed.
I think Dimes needs to be this...
http://www.netdimes.org/new/data.php?getAll=GroupsByMeas
I don't have a Vault account so haven't posted it there.
Well that's going to take a few weeks to get back to position 10 if we have to do it the hard way 
I don't think the fault is at the Vault !
I'm a poet and I didn't even know it
Team scoring for OGR-28 is pivoting at team 41 so teams are being awarded points of 243.90 per position above and -243.90 per position below the pivot. The OGR-28 score is shown as 0.000000 but I think that is what is being displayed only, it isn't meant to handle negative numbers. Our actual score is about -66,585 which is what has dragged us down 49 places in the table
Our team table shows us as having 83,453...
83,453 - 66,585 = 16,868 the value shown in the overall table
66,585 /243.90 = 273, add 41 gives 314 which is just 1 below our team position of 315. I'm assuming there is a -1 factor to put into the calculation.
Whilst on OGR-28, it is the worst project I have come across for setting up and working out how to get started. I'm guessing it is based on a really old system and they haven't got round to making it more user-friendly
I don't think the fault is at the Vault !
I'm a poet and I didn't even know it
Team scoring for OGR-28 is pivoting at team 41 so teams are being awarded points of 243.90 per position above and -243.90 per position below the pivot. The OGR-28 score is shown as 0.000000 but I think that is what is being displayed only, it isn't meant to handle negative numbers. Our actual score is about -66,585 which is what has dragged us down 49 places in the table
Our team table shows us as having 83,453...
83,453 - 66,585 = 16,868 the value shown in the overall table
66,585 /243.90 = 273, add 41 gives 314 which is just 1 below our team position of 315. I'm assuming there is a -1 factor to put into the calculation.
Whilst on OGR-28, it is the worst project I have come across for setting up and working out how to get started. I'm guessing it is based on a really old system and they haven't got round to making it more user-friendly
Well that's going to take a few weeks to get back to position 10 if we have to do it the hard way
I don't think the fault is at the Vault !
I'm a poet and I didn't even know it
Team scoring for OGR-28 is pivoting at team 41 so teams are being awarded points of 243.90 per position above and -243.90 per position below the pivot. The OGR-28 score is shown as 0.000000 but I think that is what is being displayed only, it isn't meant to handle negative numbers. Our actual score is about -66,585 which is what has dragged us down 49 places in the table
Our team table shows us as having 83,453...
83,453 - 66,585 = 16,868 the value shown in the overall table
66,585 /243.90 = 273, add 41 gives 314 which is just 1 below our team position of 315. I'm assuming there is a -1 factor to put into the calculation.
Whilst on OGR-28, it is the worst project I have come across for setting up and working out how to get started. I'm guessing it is based on a really old system and they haven't got round to making it more user-friendly
Well spotted on the negative score, I've been posting in the DC Vault forums about the problems and didn't click that's why we have such a bad score.
Do you need any help setting up OGR? I think there's a guide buried in this thread somewhere.
Associate
I tried Atlas@home errr i now see why all of us members have 0 points !!!! lol
I think I know what I'm doing now but they sure don't make it easy, helpful links to a step by step guide to installing and setting it up take you to a blank page !
I see nothing got updated/downdated on the Vault tonight
What problems are you having with Atlas ?
I've gone straight to the top of the team leaderboard tonight
The expected time to completion came and went but it finished successfully eventually. Did have lots of time to read the message boards to see people seem to have problems if they run out of memory (too little to begin with or try and run too many tasks at the same time).
Associate
What problems are you having with Atlas ?
I've gone straight to the top of the team leaderboard tonight
The expected time to completion came and went but it finished successfully eventually. Did have lots of time to read the message boards to see people seem to have problems if they run out of memory (too little to begin with or try and run too many tasks at the same time).
i got 16gig memory and it was doing 8 threads at a time, took 2 days to get to 99.9% and that was when it looked like it wasnt going anywere so got bored lol
Gona stick with Folding on main pc for a while, easier on the pocket
Get in the top 1000 pcs is my target for my 1 machine will be nicesig less AGAIN grrr
Last edited:
i got 16gig memory and it was doing 8 threads at a time, took 2 days to get to 99.9% and that was when it looked like it wasnt going anywere so got bored lol
Gona stick with Folding on main pc for a while, easier on the pocket
Yeah, that's probably too much, they reckon 2.5gigs per wu. I'd suggest you only run 4 at a time to play safe and then top up with some other jobs from something light on memory. If you did that you'd streak past me in less than 4 hours
Associate
ok 4 threads @ 99.91% after 21 hours , Virtual box seems to be set up ok, cant see any problem.
What percent of system memory is reported as being used/free ?
Are the WUs still using cpu time ?
Might be worth suspending 3 of the tasks and then restarting Boinc Manager to see if 1 can make it on its own past the finish line. They do checkpoint and I remember reading on the forums that others had restarted and had some success with that.
'Magic' on the Atlas forum has said...
You need at least 6GB of Ram to run ONE task at a time here.
If you want to run 2 of these tasks you need 8GB Ram or more.
If you have less than 6GB the VirtualBox will fail every time.
And as I mentioned before here you can check your Boinc Manager
Task-Properties and check the task compared to the actual run time you see on that task and if the *last checkpoint* is close to what it has on the Boinc running time.
If it says a few hours and yet your task has been running for over 40 hours that is a good time to abort.......then make sure you clean up the VirtualBox (file-manager)
Must admit I've only run one at a time simply because the other cores are chasing Vault points but they have all finished with no problems.
If you've tried suspending 3 and the 4th didn't finish I'd check what the run times are as Magic suggests and probably abort them. Then try just 1 at a time to see if that works and then move up to 2 concurrently to see how that goes. I only have 8Gb in the one I'm running it on with no GPU work either to complicate matters.
You need at least 6GB of Ram to run ONE task at a time here.
If you want to run 2 of these tasks you need 8GB Ram or more.
If you have less than 6GB the VirtualBox will fail every time.
And as I mentioned before here you can check your Boinc Manager
Task-Properties and check the task compared to the actual run time you see on that task and if the *last checkpoint* is close to what it has on the Boinc running time.
If it says a few hours and yet your task has been running for over 40 hours that is a good time to abort.......then make sure you clean up the VirtualBox (file-manager)
Must admit I've only run one at a time simply because the other cores are chasing Vault points but they have all finished with no problems.
If you've tried suspending 3 and the 4th didn't finish I'd check what the run times are as Magic suggests and probably abort them. Then try just 1 at a time to see if that works and then move up to 2 concurrently to see how that goes. I only have 8Gb in the one I'm running it on with no GPU work either to complicate matters.
Associate
cheers, trying 1 at a time now, on 99.955% @ 23 hours lol
Out of interest are you using an app_config.xml in the project area so that you can control how many WUs run ?
If you haven't created one in the BOINC/projects/atlasathome.cern.ch/ folder then I expect as soon as you download them then you could have 8 start to run. To stop that happening you should create a file using Notepad and call it app_config.xml with just the text below in it...
<app_config>
<app>
<name>ATLAS</name>
<max_concurrent>1</max_concurrent>
</app>
</app_config>
Once you've done this click on Advanced, Read config files in Boinc Manager and this will only let one Atlas WU run at a time. Not sure if having started another 7 that you manually suspend will have an impact on the 1 running task though.
If you haven't created one in the BOINC/projects/atlasathome.cern.ch/ folder then I expect as soon as you download them then you could have 8 start to run. To stop that happening you should create a file using Notepad and call it app_config.xml with just the text below in it...
<app_config>
<app>
<name>ATLAS</name>
<max_concurrent>1</max_concurrent>
</app>
</app_config>
Once you've done this click on Advanced, Read config files in Boinc Manager and this will only let one Atlas WU run at a time. Not sure if having started another 7 that you manually suspend will have an impact on the 1 running task though.