CPU Time as tracked by BOINC should be the actual time it has exclusive use of the CPU as tracked by the OS (which uses the time measured for things like task switching). If nothing else, or little else, runs on the CPU then CPU Time will track wall clock time one to one. Otherwise, CPU Time is going to be accumulated over a much longer interval of wall clock time.
As a complementary example, CPDN models take about 20-30 days of CPU Time to run and take several months to run as they compete with the other projects. When I first started my new Xeon-64 I started it with 4 CPDN models, they have accumulated so far:
10 Days 04:53:16
07 Days 04:14:10
04 Days 09:09:22
01 Day- 23:51:00
Which is from the computers initial start-up about a month ago ...
I am using wine in SuSE9.2. The most annoying problem I have encountered is that I sometimes find that boinc/einstein is idle with two units at 100% but not sent and two other units at 0% & not running. [This is a dual-cpu computer.] I think this problem gets started when communications are interrupted and are not available when boinc wants a new unit. Maybe my ISP is down for a while or maybe einstein is down for a while. Boinc does not seem to recover on its own after communications start again. It just sits there until I shut it down myself & restart it. The einstein web site sometimes makes matters worse by refusing to supply new units for 24 hours.
I think this happens with both the commeand line & the gui boinc. I use the gui boinc because it at least easier to see what is wrong.
ADDMP
I don't know whether other wine users have encountered this.
I've never wine-ed the command line, can't see why I should when I can just as well wine the gui. Anyway, the gui is just a nicer way of showing, but does not change the core, so the fact it happens with both the command line and the gui is quite normal. When you start up the gui, and look at the column titled 'status' what does it say for the two units at 100%? Usually, they get sent right after finishing (unless you're on dial-up or you specified it can only connect between x and y time). Then, it waits until it connects the the scheduler again to find out if it's been 'reported' right, and then disappears. Maybe those two are waiting for that confirmation. Why there's no other running, I have no idea. SuSe 9.2: yes, dual-cpu: no.
I am using CentOS with wine. I can't get 4.45 or 5.2.13 to connect to localhost. I can't get it to connect to any comps on my network. BOINC MGR just says it can't connect. What is wrong? How do I fix it?
I am using CentOS with wine. I can't get 4.45 or 5.2.13 to connect to localhost. I can't get it to connect to any comps on my network. BOINC MGR just says it can't connect. What is wrong? How do I fix it?
Same Problem here; but it works with BOINC 4.19.
However, the hosts have a strange behaviour though. Sometimes they claim 0 credit, but all results have validated so far.
Ok, I finally said screw the GUI and went command line in wine. I got it up and running! Sort of...now I just get a can't get time error that scrolls by endlessly forever. Basically locking the system with this endless loop until I hit Ctrl C. Why can't this just work?
Ok, I finally said screw the GUI and went command line in wine. I got it up and running! Sort of...now I just get a can't get time error that scrolls by endlessly forever. Basically locking the system with this endless loop until I hit Ctrl C. Why can't this just work?
This old message gives a hint ... set your WINEDEBUG environment variable according to the current WINE documentation.
...
However, the hosts have a strange behaviour though. Sometimes they claim 0 credit, but all results have validated so far.
...
This could be a benchmarking problem. If benchmarking fails, or if you disable benchmarking, or if for some reason your benchmarks have expired and not been re-run, then the client claims zero credit.
As the Albert WU have a quorum of three, this will not matter much, as everyone gets the middle credit value. It does mean that statistically speaking you (and your unfortunate co-crunchers on that WU) might get a low credit if one of the other hosts is generally low scoring.
Credits can be odd when the result validates, and vice versa - these are two different forms of quality check: "validation" is a quality check on the science, whereas the benchmarks and quorum-credit are qualityu control on the process of measuring the work done. On a complex system like this it is entirely possible that the work can be of high quality but that the means of mesuring the amount of effort is uncertain.
In short: you can be reassured that the science you are turning in is valid, even if your scores are a little below what you strictly deserve.
...
However, the hosts have a strange behaviour though. Sometimes they claim 0 credit, but all results have validated so far.
...
This could be a benchmarking problem. If benchmarking fails, or if you disable benchmarking, or if for some reason your benchmarks have expired and not been re-run, then the client claims zero credit.
Pretty sure this is not the case, as I have the same problem too, but it doesn't happen all the time, just most of the time. Also, my clients have gone through benchmarks several times since updating the binaries in wine.
such things just should not be writ so please destroy this if you wish to live 'tis better in ignorance to dwell than to go screaming into the abyss worse than hell
CPU Time as tracked by BOINC
)
CPU Time as tracked by BOINC should be the actual time it has exclusive use of the CPU as tracked by the OS (which uses the time measured for things like task switching). If nothing else, or little else, runs on the CPU then CPU Time will track wall clock time one to one. Otherwise, CPU Time is going to be accumulated over a much longer interval of wall clock time.
As a complementary example, CPDN models take about 20-30 days of CPU Time to run and take several months to run as they compete with the other projects. When I first started my new Xeon-64 I started it with 4 CPDN models, they have accumulated so far:
10 Days 04:53:16
07 Days 04:14:10
04 Days 09:09:22
01 Day- 23:51:00
Which is from the computers initial start-up about a month ago ...
I am using wine in SuSE9.2.
)
I am using wine in SuSE9.2. The most annoying problem I have encountered is that I sometimes find that boinc/einstein is idle with two units at 100% but not sent and two other units at 0% & not running. [This is a dual-cpu computer.] I think this problem gets started when communications are interrupted and are not available when boinc wants a new unit. Maybe my ISP is down for a while or maybe einstein is down for a while. Boinc does not seem to recover on its own after communications start again. It just sits there until I shut it down myself & restart it. The einstein web site sometimes makes matters worse by refusing to supply new units for 24 hours.
I think this happens with both the commeand line & the gui boinc. I use the gui boinc because it at least easier to see what is wrong.
ADDMP
I don't know whether other wine users have encountered this.
I've never wine-ed the
)
I've never wine-ed the command line, can't see why I should when I can just as well wine the gui. Anyway, the gui is just a nicer way of showing, but does not change the core, so the fact it happens with both the command line and the gui is quite normal. When you start up the gui, and look at the column titled 'status' what does it say for the two units at 100%? Usually, they get sent right after finishing (unless you're on dial-up or you specified it can only connect between x and y time). Then, it waits until it connects the the scheduler again to find out if it's been 'reported' right, and then disappears. Maybe those two are waiting for that confirmation. Why there's no other running, I have no idea. SuSe 9.2: yes, dual-cpu: no.
I am using CentOS with wine.
)
I am using CentOS with wine. I can't get 4.45 or 5.2.13 to connect to localhost. I can't get it to connect to any comps on my network. BOINC MGR just says it can't connect. What is wrong? How do I fix it?
RE: I am using CentOS with
)
Same Problem here; but it works with BOINC 4.19.
However, the hosts have a strange behaviour though. Sometimes they claim 0 credit, but all results have validated so far.
# wine --version
Wine 0.9.9
Ok, I finally said screw the
)
Ok, I finally said screw the GUI and went command line in wine. I got it up and running! Sort of...now I just get a can't get time error that scrolls by endlessly forever. Basically locking the system with this endless loop until I hit Ctrl C. Why can't this just work?
RE: Ok, I finally said
)
This old message gives a hint ... set your WINEDEBUG environment variable according to the current WINE documentation.
Regards.
-rg-
RE: RE: ... However, the
)
~~gravywavy
RE: RE: RE: ... However
)
such things just should not be writ so please destroy this if you wish to live 'tis better in ignorance to dwell than to go screaming into the abyss worse than hell
Sooo...basically my only
)
Sooo...basically my only recourse is to use BOINC 4.19?