1) Forums : Technical Support : Can NOT upload completed work. Server problems? (Message 21757)
Posted 30 Jun 2018 by Aurum
Post:
Where can we find this log??? I looked but don't see anything. I do see that the C@H Notices file is blank. Sending out a notice would reassure everyone.
2) Forums : Technical Support : Can NOT upload completed work. Server problems? (Message 21754)
Posted 30 Jun 2018 by Aurum
Post:
I'm not getting or sending any C@H WUs.
3) Forums : Technical Support : GridCoin Team is gone??? (Message 21752)
Posted 30 Jun 2018 by Aurum
Post:
Today I see my Team was blank. I joined Gridcoin again. Hope it sticks.
BTW, UL/DL servers are not working.
4) Forums : Technical Support : GridCoin Team is gone??? (Message 21739)
Posted 17 Jun 2018 by Aurum
Post:
I noticed that my Cosmo team had changed from Gridcoin to INTERNATION. I quit and then tried to rejoin the Gridcoin Team but it has vanished. I don't see any announcements.
5) Forums : Technical Support : C@H hit or miss... (Message 21728)
Posted 7 Jun 2018 by Aurum
Post:
I tried changing
Max # of jobs for this project:  1
and apparently it means how many WUs can be DLed. I set it to one and I get one. I changed it to 8 and I got 7 more WUs.
6) Forums : Technical Support : C@H hit or miss... (Message 21727)
Posted 7 Jun 2018 by Aurum
Post:
I just tried deleting my app_config.xml file, exiting BOINC and just using
Max # of jobs for this project:  1
Max # of CPUs for this project:  1
The jobs seem to only use a single CPU thread but C@H took over all CPUs NOT just one. No idea what "Max # of jobs for this project" actually does.
7) Forums : Technical Support : C@H hit or miss... (Message 21726)
Posted 7 Jun 2018 by Aurum
Post:
It was on the same 3 headless rigs as above:
Rig-01 327541 {has 4 1080 Ti GPUs}
Rig-13 327537 {has 4 1080 Ti GPUs}
Rig-15 330327 {has 4 1080 Ti GPUs}
The ones that have been stable and seem to run C@H with no problem are all desktop PCs with only a single Nvidia GPU. I assumed that the BOINC client_state.xml file was created when BOINC is loaded. But there must other events that trigger writing a new file. Also, all rigs have World Community Grid CPU WUs running.
C@H tries to use all CPU threads and does not play well with others. It's also a RAM hog so I've had to limit it on all rigs to a single CPU and no more than 4 concurrent WUs:
<app_config>
    <app>
        <name>camb_boinc2docker</name>
        <max_concurrent>1</max_concurrent>
    </app>
    <app_version>
        <app_name>camb_boinc2docker</app_name>
        <plan_class>vbox64_mt</plan_class>
        <avg_ncpus>1</avg_ncpus>
    </app_version>
    <project_max_concurrent>4</project_max_concurrent>
</app_config>

Anecdotally I've noticed that if I let C@H use multiple CPUs then it greatly increases the likelihood that a WU will fail due to Computation_error.
I just noticed that C@H Preferences has the option to limit CPUs which may work better than my attempt at writing an app_config.xml file.
Max # of jobs for this project:  1
Max # of CPUs for this project:  1
8) Forums : Technical Support : C@H hit or miss... (Message 21719)
Posted 6 Jun 2018 by Aurum
Post:
They worked for awhile and then when I checked later in the day they'd reset to:
<p_vm_extensions_disabled>1</p_vm_extensions_disabled>
9) Forums : Technical Support : C@H hit or miss... (Message 21718)
Posted 5 Jun 2018 by Aurum
Post:
Thanks Marius, The second approach fixed it. client_state now says:
<p_vm_extensions_disabled>0</p_vm_extensions_disabled>
10) Forums : Technical Support : C@H hit or miss... (Message 21714)
Posted 4 Jun 2018 by Aurum
Post:
From their client_state.xml files:
Rig-01 <hostid>327541</hostid>
Rig-13 <hostid>327537</hostid>
Rig-15 <hostid>330327</hostid>

I tagged other BOINC CPU projects as "No new work" and let them run a day. I even let one entirely run out of CPU work and these did not get a single C@H WU.
Others get as many C@H WUs as they can handle delivered fast. E.g., <hostid>328562</hostid>.
11) Forums : Technical Support : C@H hit or miss... (Message 21701)
Posted 30 May 2018 by Aurum
Post:
I've tried everything I can think of to get C@H WUs into my computers. I have one that DLs lots of them and the others get nothing. I set all BOINC projects to "No new tasks" and increased the number of days of work to request. I've let work run down to less the number of threads and still no Cosmo WUs. There's a bug somewhere.
12) Forums : Technical Support : C@H hit or miss... (Message 21676)
Posted 8 Apr 2018 by Aurum
Post:
I have a number of Win7-64 rigs running BOINC and C@H. Some DL WUs continuously while some never DL a single WU. I've searched for a reason. I find that limiting it to 2, 4, 6 or 8 CPUs works best for me. I've tried suspending any other BOINC job. I can find no differences that might explain why some PCs don't ever DL a C@H WU but have no problem with any other BOINC project.
Any suggestions of what I can do to get them running C@H ???
13) Forums : Technical Support : Computation Error (17 CPUs) (Message 21535)
Posted 9 Oct 2017 by Aurum
Post:
I have C@H running on several rigs with Xeon E5 CPUs (24-28 cores). Some get only 6-CPU WUs, others 8_CPU and some get 17-CPU. The small ones seem to run without any problems but the big ones may run the first but then dump all the rest as Computation Errors. Could it have anything to do with trying to use a prime number of cores??? I've heard tales that was a problem on another project.
14) Forums : Technical Support : Only getting camb_legacy, virtualbox installed and detected (Message 21528)
Posted 7 Oct 2017 by Aurum
Post:
Thanks Jonathon, I didn't see the camb_legacy option until I clicked on All Applications. Giving it a try now.
15) Forums : Technical Support : Only getting camb_legacy, virtualbox installed and detected (Message 21526)
Posted 7 Oct 2017 by Aurum
Post:
How can I stop getting legacy WUs??? All my rigs are Win7-64 with mostly Xeon 24-28 core CPUs. One rig only gets 6-CPU WUs but several only get legacy jobs. I have boinc_7.8.2_windows_x86_64_vbox installed on all rigs.