Okay, backed off to the unoptimized BOINC 4.43 client. Same problem, Einstein shows in BOINC Manager as being paused, but in the background the Einstein app is still running with the Predictor app.
I'm running both Einstein 0.05 and SETI 4.18 through the standard BOINC Manager 4.43, on a G4 PowerMac with OSX 10.3.9, and with those, Einstein is pausing as it should do, and not running in the background. The beta version of Einstein certainly seems to run faster, which is better for my 450Mhz G4!
The only problem I have with it is that I'm still having to click on 'Show graphics' if I want the screensaver to work but I assume that's a BOINC problem rather than anything to do with Einstein, as I have to do the same for SETI.
MacStef: You are correct. I did just what the download instructions specified - I put the 0.05 version into the same folder that had the 0.02 version, which was left in there also. When I restarted, the older WU continued to execute under 0.02, and when it finished, a new WU began under 0.05. Pretty slick!!
Now for the other news...
Good: My PowerBook has completed 2 WU since last night, both using the 0.05, and both were valid and received credit already. This combination is BOINC 4.44 Superbench with Einstein 0.05 running as CLI in Terminal. I haven't caught it with the pausing problem yet...
Bad: My iBook still has the pausing problem with BOINC 4.44 SB and E@H 0.05 running in MenuBar. I just switched back to BOINC 4.43 in MenuBar, but left the 0.05 in place, so I'll see in a couple of hours if it continues to have the pausing problem.
My experience has been the same as Shaktai's - by itself, Einstein 0.05 runs clean and fast with the 4.44 SuperBench, both in MenuBar and as CLI; when it has to share resources with other apps, it starts acting like a junk yard dog .
Ok, I'll have another look at the "suspend" problem. I already have a clue what's going on, but this needs a bit more investigation.
Curently we received a bit more than a hundred results from the 0.05 Beta, quite some of them with a 4.44 client, no client errors and none being invalid.
Thanks, C! Am I understanding that you upgraded to Einstein .05 while you still had Einstein .02 WUs crunching and they didn't get fracked?!
This is correct. If you leave the previous .02 executable in place AND use our new app_info.xml file, the jobs that were 'married' to the .02 app will run with that app. Look in app_info.xml and you'll see why: it's because the latest app_info.xml file lists BOTH the .02 and the .05 app.
This is correct. If you leave the previous .02 executable in place AND use our new app_info.xml file, the jobs that were 'married' to the .02 app will run with that app. Look in app_info.xml and you'll see why: it's because the latest app_info.xml file lists BOTH the .02 and the .05 app.
Bruce
And dummy me didn't read the instructions right. Removed the .02 app and sure enough, all the .02 work units errored. Of course by then it was too late to go back. All the .05 are running fine, just as long as they don't run with predictor. Did a little more observation and Einstein won't suspend when the other project kicks in, but when it switches back to Einstein, then it runs only Einstein.
Thanks for all the work. Thinking maybe I'll try the CLI to see if the problem exists there. Definitely had occured with both the Advanced GUI and the menubar GUI.
Mac OS X beta 0.05 running on a 2x1gig G4 running 10.3.9
Beta 0.05 alone runs without a problem.
Beta, as others have noted, does not pause appropriately when run under BOINC Manager with other clients; even when it says that it has.
No problems with installation, graphics, or client communication.
I took BOINC 4.44 off and reinstalled 4.43 in Menubar on my 1.33 GHz iBook G4. Left the new E@H 0.05 in place. Einstein runs fine and I have received valid credits. The problem is it still won't pause when it says it does. SETI does pause ok, and then E@H runs 100%, but as soon as E@H says it's pausing, SETI starts up at 50% and E@H drops to 50%.
This WU: 6768138 has an interesting note. It says it is running 4.44 and E@H app .02, but my MenuBar log shows 4.43 and E@H .05. When I look in the Projects folder, I see only E@H .05...
One other item: I was running the 4.44 SuperBench and claiming around 117 credits for a completed WU. That hasn't changed since I switched back to 4.43.
[edit] I just checked my 1.0 GHz Powerbook G4 running 4.44 and E@H 0.05 as CLI and see in Activity Monitor that E@H is also not pausing properly for SETI and CPDN on that machine.
[/edit]
The problem is it still won't pause when it says it does. SETI does pause ok, and then E@H runs 100%, but as soon as E@H says it's pausing, SETI starts up at 50% and E@H drops to 50%.
C
It's odd that's happening for so many people but not on my G4. I keep checking in Activity Monitor and Einstein is definitely not running at the same time as SETI. The only difference is that I've never had BOINC 4.44 on this machine.
I upgraded the client to version 0.05 BETA from 0.02 BETA on a G4 Mac with OS 10.3.9 and BOINC 4.44 from MacTNN. I have finished just one WU since, but it seems a bit slow compared to the time taken by the 0.02 client. I haven't uploaded the result yet, but I figure the times are ~20 hours/WU for the unoptimized client, ~11 hours for 0.02 client and ~12 - 13 hours for the 0.05 client. Is this by design?
Beta 0.05 installed on a Mac Mini (1,42 Ghz G4 - OS X 10.4.2), running BOINC Advanced GUI v4.43
First WU successfully completed, CPU shared with four other Applications.
It seems that .05 is a little bit less efficient that 0.02 (29 ksec per WU, instead of 22-25 ksec)
I didn't see any problem in Log file, or Stalled WU.
RE: Okay, backed off to the
)
I'm running both Einstein 0.05 and SETI 4.18 through the standard BOINC Manager 4.43, on a G4 PowerMac with OSX 10.3.9, and with those, Einstein is pausing as it should do, and not running in the background. The beta version of Einstein certainly seems to run faster, which is better for my 450Mhz G4!
The only problem I have with it is that I'm still having to click on 'Show graphics' if I want the screensaver to work but I assume that's a BOINC problem rather than anything to do with Einstein, as I have to do the same for SETI.
MacStef: You are correct.
)
MacStef: You are correct. I did just what the download instructions specified - I put the 0.05 version into the same folder that had the 0.02 version, which was left in there also. When I restarted, the older WU continued to execute under 0.02, and when it finished, a new WU began under 0.05. Pretty slick!!
Now for the other news...
Good: My PowerBook has completed 2 WU since last night, both using the 0.05, and both were valid and received credit already. This combination is BOINC 4.44 Superbench with Einstein 0.05 running as CLI in Terminal. I haven't caught it with the pausing problem yet...
Bad: My iBook still has the pausing problem with BOINC 4.44 SB and E@H 0.05 running in MenuBar. I just switched back to BOINC 4.43 in MenuBar, but left the 0.05 in place, so I'll see in a couple of hours if it continues to have the pausing problem.
My experience has been the same as Shaktai's - by itself, Einstein 0.05 runs clean and fast with the 4.44 SuperBench, both in MenuBar and as CLI; when it has to share resources with other apps, it starts acting like a junk yard dog .
C
Join Team MacNN
Ok, I'll have another look at
)
Ok, I'll have another look at the "suspend" problem. I already have a clue what's going on, but this needs a bit more investigation.
Curently we received a bit more than a hundred results from the 0.05 Beta, quite some of them with a 4.44 client, no client errors and none being invalid.
BM
BM
RE: Thanks, C! Am I
)
This is correct. If you leave the previous .02 executable in place AND use our new app_info.xml file, the jobs that were 'married' to the .02 app will run with that app. Look in app_info.xml and you'll see why: it's because the latest app_info.xml file lists BOTH the .02 and the .05 app.
Bruce
Director, Einstein@Home
RE: RE: This is correct.
)
Team MacNN - The best Macintosh team ever.
Mac OS X beta 0.05 running on
)
Mac OS X beta 0.05 running on a 2x1gig G4 running 10.3.9
Beta 0.05 alone runs without a problem.
Beta, as others have noted, does not pause appropriately when run under BOINC Manager with other clients; even when it says that it has.
No problems with installation, graphics, or client communication.
Thanks!
I took BOINC 4.44 off and
)
I took BOINC 4.44 off and reinstalled 4.43 in Menubar on my 1.33 GHz iBook G4. Left the new E@H 0.05 in place. Einstein runs fine and I have received valid credits. The problem is it still won't pause when it says it does. SETI does pause ok, and then E@H runs 100%, but as soon as E@H says it's pausing, SETI starts up at 50% and E@H drops to 50%.
This WU: 6768138 has an interesting note. It says it is running 4.44 and E@H app .02, but my MenuBar log shows 4.43 and E@H .05. When I look in the Projects folder, I see only E@H .05...
One other item: I was running the 4.44 SuperBench and claiming around 117 credits for a completed WU. That hasn't changed since I switched back to 4.43.
[edit] I just checked my 1.0 GHz Powerbook G4 running 4.44 and E@H 0.05 as CLI and see in Activity Monitor that E@H is also not pausing properly for SETI and CPDN on that machine.
[/edit]
C
Join Team MacNN
RE: The problem is it still
)
It's odd that's happening for so many people but not on my G4. I keep checking in Activity Monitor and Einstein is definitely not running at the same time as SETI. The only difference is that I've never had BOINC 4.44 on this machine.
I upgraded the client to
)
I upgraded the client to version 0.05 BETA from 0.02 BETA on a G4 Mac with OS 10.3.9 and BOINC 4.44 from MacTNN. I have finished just one WU since, but it seems a bit slow compared to the time taken by the 0.02 client. I haven't uploaded the result yet, but I figure the times are ~20 hours/WU for the unoptimized client, ~11 hours for 0.02 client and ~12 - 13 hours for the 0.05 client. Is this by design?
Beta 0.05 installed on a Mac
)
Beta 0.05 installed on a Mac Mini (1,42 Ghz G4 - OS X 10.4.2), running BOINC Advanced GUI v4.43
First WU successfully completed, CPU shared with four other Applications.
It seems that .05 is a little bit less efficient that 0.02 (29 ksec per WU, instead of 22-25 ksec)
I didn't see any problem in Log file, or Stalled WU.