So far 012 has been running for over 2 solid days (no stops of any kind). I have seen no hangs or halts, it seems to be playing well with BOINC 4.43 (Standard release), CP@H 4.12, S@H 4.30 (Mac NN optimized), P@H 4.29 (Standard release). All of the system running E@H are Mac os 10.3.9. I have seen no bad behavior of a serious nature, and what I have seen is more likely to be BOINC then E@H.
So far all of the E@H WU have validated ok.
The only strange stuff I have seen is that on occasion P@H and S@H will share a CPU. This can (and has) happened with any of the other apps running on the other processor. Usually at the next app swap it fixes itself. I am running with the apps retained in memory (mostly because of CP@H). So it might not be a fair test of a full app swap out. So far the system runs in any combination of apps on the two processors. So far it looks real good from here.
The WUs run fairly consistently at around 7:10 hours per. A lot better than the standard app.
Good Work
Phil
We Must look for intelligent life on other planets as,
it is becoming increasingly apparent we will not find any on our own.
Can anyone report on the performance of the 0.12 test app with older (eg, 4.19) core clients? If all continues to go well I think I may soon switch the 0.12 test app to being the standard Mac OS X app.
Bruce
What the heck, I'll give it a try here tonight with 4.19.
Had one hang while a .11 work unit was still in queue. None with the .12 so far. I'll switch to 4.19, and run it with a little predictor and should be able to give results in 24hours.
Snake Doctor: This sounds like the suspend problem which is probably still present in the SETI App.
Bruce, Shaktai: I just switched my Mac mini to the old 4.19 client. We should have some results by tomorrow.
BM
Well that makes two of us then. I should be able to complete about 4 work units by this time tomorrow night. That should give a good sample between both of us. Had to run it in terminal though.
Snake Doctor: This sounds like the suspend problem which is probably still present in the SETI App.
Bruce, Shaktai: I just switched my Mac mini to the old 4.19 client. We should have some results by tomorrow.
BM
Yea, I would agree. But this does not seems to be a E@H issue. So far I have not seen an E@H WU sharing the CPU with anything else. Now to be truthful I am not staring at the monitor all the time (not a lot of a life but I do have one). It is possible that it happens when I am not looking. But in any case, it does not seem to cause a problem beyond slowing processing to 50% speed for the apps involved. As I said it self corrects.
Phil
We Must look for intelligent life on other planets as,
it is becoming increasingly apparent we will not find any on our own.
Don't know if this is important or even related to E@H but thought I should report it just the same.
This morning the entire system was hung. Not just BOINC but the entire system. This is a a very rare event for us. It happens but it is rare. I have the screen saver set to something other than BOINC and last night I let it go to screen saver mode for the first time in a while. It may not be related to the hang, but it stopped just after reporting a "trickle" to CP@H. Also, P@H is not sending WUs right now and the system was asking for work every 50 min or so, and not getting any. Unfortunately, I could not view the screen to see the exact condition of the system at the time of the hang. I do not think it is related to E@H, but I think there may be a problem with running other screen savers while BOINC is running CP@H. Based on the condition of the system when the screen save kicked on (I was watching) it processed for about 55 min. before it "trickled" and hung.
When I restarted, the system Hung again when BOINC loaded. It was trying to process S@H on both CPUs when that happened. It was up just long enough to start working, then it hung. Rebooted a second time and it seems to be fine now.
Regards
phil
We Must look for intelligent life on other planets as,
it is becoming increasingly apparent we will not find any on our own.
This morning the entire system was hung. Not just BOINC but the entire system. This is a a very rare event for us. It happens but it is rare.
Phil sound like you had a kernel panic. It has happened to me a few times as well. It seems to be a Boinc issue. And it doesn't happen often, but it does happen. For we It happens over night most of the time.
Don't know if this is important or even related to E@H but thought I should report it just the same.
This is what I would expect if you are attached to projects whose applications have NOT been built using the very latest version of the BOINC API library (August 12, 2005 or later).
So far 012 has been running
)
So far 012 has been running for over 2 solid days (no stops of any kind). I have seen no hangs or halts, it seems to be playing well with BOINC 4.43 (Standard release), CP@H 4.12, S@H 4.30 (Mac NN optimized), P@H 4.29 (Standard release). All of the system running E@H are Mac os 10.3.9. I have seen no bad behavior of a serious nature, and what I have seen is more likely to be BOINC then E@H.
So far all of the E@H WU have validated ok.
The only strange stuff I have seen is that on occasion P@H and S@H will share a CPU. This can (and has) happened with any of the other apps running on the other processor. Usually at the next app swap it fixes itself. I am running with the apps retained in memory (mostly because of CP@H). So it might not be a fair test of a full app swap out. So far the system runs in any combination of apps on the two processors. So far it looks real good from here.
The WUs run fairly consistently at around 7:10 hours per. A lot better than the standard app.
Good Work
Phil
We Must look for intelligent life on other planets as,
it is becoming increasingly apparent we will not find any on our own.
RE: Can anyone report on
)
What the heck, I'll give it a try here tonight with 4.19.
Had one hang while a .11 work unit was still in queue. None with the .12 so far. I'll switch to 4.19, and run it with a little predictor and should be able to give results in 24hours.
Team MacNN - The best Macintosh team ever.
Snake Doctor: This sounds
)
Snake Doctor: This sounds like the suspend problem which is probably still present in the SETI App.
Bruce, Shaktai: I just switched my Mac mini to the old 4.19 client. We should have some results by tomorrow.
BM
BM
RE: Snake Doctor: This
)
Well that makes two of us then. I should be able to complete about 4 work units by this time tomorrow night. That should give a good sample between both of us. Had to run it in terminal though.
Team MacNN - The best Macintosh team ever.
RE: Snake Doctor: This
)
Yea, I would agree. But this does not seems to be a E@H issue. So far I have not seen an E@H WU sharing the CPU with anything else. Now to be truthful I am not staring at the monitor all the time (not a lot of a life but I do have one). It is possible that it happens when I am not looking. But in any case, it does not seem to cause a problem beyond slowing processing to 50% speed for the apps involved. As I said it self corrects.
Phil
We Must look for intelligent life on other planets as,
it is becoming increasingly apparent we will not find any on our own.
Don't know if this is
)
Don't know if this is important or even related to E@H but thought I should report it just the same.
This morning the entire system was hung. Not just BOINC but the entire system. This is a a very rare event for us. It happens but it is rare. I have the screen saver set to something other than BOINC and last night I let it go to screen saver mode for the first time in a while. It may not be related to the hang, but it stopped just after reporting a "trickle" to CP@H. Also, P@H is not sending WUs right now and the system was asking for work every 50 min or so, and not getting any. Unfortunately, I could not view the screen to see the exact condition of the system at the time of the hang. I do not think it is related to E@H, but I think there may be a problem with running other screen savers while BOINC is running CP@H. Based on the condition of the system when the screen save kicked on (I was watching) it processed for about 55 min. before it "trickled" and hung.
When I restarted, the system Hung again when BOINC loaded. It was trying to process S@H on both CPUs when that happened. It was up just long enough to start working, then it hung. Rebooted a second time and it seems to be fine now.
Regards
phil
We Must look for intelligent life on other planets as,
it is becoming increasingly apparent we will not find any on our own.
First Result from 4.19 client
)
First Result from 4.19 client and 0.12 App dropped in. No problems, no noticable change in computation time (8,5h on the Mac mini).
BM
BM
RE: This morning the entire
)
Phil sound like you had a kernel panic. It has happened to me a few times as well. It seems to be a Boinc issue. And it doesn't happen often, but it does happen. For we It happens over night most of the time.
RE: Don't know if this is
)
This is what I would expect if you are attached to projects whose applications have NOT been built using the very latest version of the BOINC API library (August 12, 2005 or later).
Bruce
Director, Einstein@Home
RE: First Result from 4.19
)
I've got three in, pending validation and should have a fourth shortly. No problems indicated so far. This is on my iMac G5 running the terminal.
http://einsteinathome.org/host/390527/tasks
Team MacNN - The best Macintosh team ever.