iMac processing questions.
22 Dec 2011 23:29:10 UTC
Topic 196116
(moderation:
I have the latest generation of an iMAC (3.06 GHz intel Core Duo), running 10.6.8. I have the preference set to run on both processors, but it only runs on one. On my other Mac (Mac Book) Boinc runs on both.
How can I get the iMac to run Boinc on both cores.
A second question:
I am not getting any work unit on the iMc, while the MacBook gets as many as it needs. What is going on there?
Peter
iMac processing questions.
)
What kind of messages are you seeing in the event log in Boinc manager? I didn't see anything obvious in comparing your computers and their last communication on the webpage.
I get the following
)
I get the following "Communication Deferred" (followed by a count down clock)
RE: I get the following
)
That's not the event log - that's the status of a particular project on the projects tab. That gives no useful information for working out what is wrong.
If you look at your host details, you can see that BOINC recognises that your machine has two cores. BOINC will use both if it is allowed to do so by your preference settings. Please check those very carefully in case you are preventing BOINC from using both.
Also find your event log (what used to be the old messages tab of BOINC Manager) and look there for clues as to what is going on. When this machine works for Seti, do both cores get used? If not, it most likely is to do with how you have configured your machine. If you can't work it out, stop BOINC and restart. After everything settles, go to the event log and copy and paste all the startup message (maybe 20-50 lines) into a reply here. Those messages will probably contain the answer.
Two possibilities I can think of:
1. You have NNT (No New Tasks) set for Einstein.
2. You have a large debt owing to Seti and BOINC is preventing E@H tasks from being downloaded until the debt is repaid.
Questions for you. You have no tasks at all in the online database for that host. That means you haven't been sent a new task for probably 3-4 weeks or more. It would seem to be unlikely that a debt issue could last that long. Has that machine had regular Seti work all during the last several weeks? Have you checked what the debt values are (the values are in your state file) for both projects to see if there is anything unusual going on?
One last point. On your hosts page on the website, there is a 'Last contact' link for each of your hosts. Clicking on the relevant link gives the following info that was sent back to your host at the time of last contact.
You can see that your BOINC client is telling the server that it only has one effective CPU and that it is requesting 0.00 seconds of work. You have to work out why it's doing that.
Cheers,
Gary.
Two possibilities I can think
)
Two possibilities I can think of:
1. You have NNT (No New Tasks) set for Einstein.
I do not have NNT set . It shows "No New Tasks" as an option to click on
2. You have a large debt owing to Seti and BOINC is preventing E@H tasks from being downloaded until the debt is repaid.
I could not find a "Debt" anywher.
Questions for you. You have no tasks at all in the online database for that host. That means you haven't been sent a new task for probably 3-4 weeks or more. It would seem to be unlikely that a debt issue could last that long. Has that machine had regular Seti work all during the last several weeks?
Yes
Have you checked what the debt values are (the values are in your state file) for both projects to see if there is anything unusual going on?
I do not understandt what you want me to do here. Each time i request work from Einstein at home I get the "Communication Deferred (and a count dow clock).
One last point. On your hosts page on the website, there is a 'Last contact' link for each of your hosts. Clicking on the relevant link gives the following info that was sent back to your host at the time of last contact.
Which "Host"?
2011-12-23 03:48:04.2544 [PID=26423] Request: [USER#xxxxx] [HOST#1012921] [IP xxx.xxx.xxx.157] client 6.12.35
2011-12-23 03:48:04.2589 [PID=26423] [send] effective_ncpus 1 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999
2011-12-23 03:48:04.2590 [PID=26423] [send] effective_ngpus 0 max_jobs_on_host_gpu 999999
2011-12-23 03:48:04.2590 [PID=26423] [send] Not using matchmaker scheduling; Not using EDF sim
2011-12-23 03:48:04.2590 [PID=26423] [send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
2011-12-23 03:48:04.2590 [PID=26423] [send] work_req_seconds: 0.00 secs
2011-12-23 03:48:04.2590 [PID=26423] [send] available disk 81.23 GB, work_buf_min 0
2011-12-23 03:48:04.2590 [PID=26423] [send] active_frac 0.999949 on_frac 0.992236 DCF 1.465386
2011-12-23 03:48:04.2614 [PID=26423] Sending reply to [HOST#1012921]: 0 results, delay req 60.00
2011-12-23 03:48:04.2616 [PID=26423] Scheduler ran 0.013 seconds
You can see that your BOINC client is telling the server that it only has one effective CPU and that it is requesting 0.00 seconds of work. You have to work out why it's doing that.
Check in Boinc Manager for
)
Check in Boinc Manager for the Event Log. It is located under Advanced menu. What do you see for the Einstein project in the notes? Have you checked your project preferences under 'Your Account' at the bottom of the forums page? What application are you trying to run? Are you trying CPU and GPU or just CPU?
Have you tried resetting the project in Boinc Manager?
RE: 2. You have a large
)
BOINC Manager->Advanced view->Projects tab->Select Einstein@Home->Click Properties in the left menu.
Values of interest here are the scheduling priority and the CPU/GPU work fetch deferred values. Check those against the other projects you run as well.
The project with the largest scheduling priority will be eligible to fetch work next.
I have done what you
)
I have done what you suggested, and I now am getting Einstein an home to load new work.
Thanks.
However I still can not get both processors running Boinc.
Peter
What computer did you reset
)
What computer did you reset the project on? It appears you did the reset on the older computer according to the last communication logs. Is the older computer still processing on both cpu cores? You may need to exit the Boinc Manager and all running processes on your newer iMAC and restart it. You also need to copy and past the log text, about the first 50 lines, in this forum. You need to switch your BOINC manager from the simple view to the Advanced View. It should be a button in the lower right corner if you are still in Simple View. After that, just follow the instructions to get to the event log that were posted earlier