OK I should now be unhidden/updated. Ref ID 381946. Comments I hope.
You have two machines, the celeron 2.0GHz and a P4 2.66GHz. I would guess that the P4 should take around 14-20 hours per task and the celeron around 22-30 hours. There is quite a bit of variability between different tasks so it's hard to be precise.
Neither machine currently has a completed task to look at but do you recall approximately how long the P4 has been taking? I'm interested to know if perhaps both machines are performing well below average.
Einstein P4 appr 15 hours (but on this SETI is slow...)
I asked previously and didn't get an answer - is the increment in CPU time (elapsed time) for a crunching task as shown in BOINC Manager pretty much in accord with the increment in wall clock time?
4-6 minutes slow (from memory)
I checked the benchmarks of both your machines - the celeron shows 1048/1890 and the P4 1302/2499. These values are pretty much in line with those of similar machines that I run and so there seems to be no problem with your hosts performing as they should, at least as far as benchmarks are concerned.
At first, I was tempted to think that perhaps you had severely throttled your celeron through your preferences. Can you please check your computing preferences on the website to see if you have "use at most 100% of CPU time" set and not some drastically lower value. I can't see that it can be this because in that case the elapsed CPU time would still be normal even though the wall clock time would be very much larger. The figures you have previously listed show abnormally long CPU time which is not consistent with throttling.
Use at most 60%, P4 70%
There are two other things to check that I can think of - and hopefully more that others might come up with.
Firstly, is the celeron running in a hot environment or does it have a heat sink that is clogged with fluff. Is it possible that the machine is throttling itself due to thermal conditions?
Variable temp (home office w/o air condition)I am no technician but the SMART readings are OK.
Secondly, does the machine have a keyboard attached?
Yes
I'm currently running a very large number of machines, most of which are running Linux and a few run Windows. Virtually all machines are headless and have no keyboard and mouse attached. Those that run Windows are afflicted by an issue that appears after a month or two of continuous running where the hard disk shows continuous activity and network activity becomes just about impossible. The CPU is still crunching but slows to a crawl. The fix is to plug in a keyboard and instantly things return to normal. The linux machines don't seem to suffer from this problem.
One final point. I notice over at Seti about 2.5 years ago you wrote a message about having to investigate why Seti crunching had slowed to a crawl. Did you ever work out why that was happening?
As far as I remember the problem righted itself..
Is your current Seti crunching performing properly or is it vastly slowed down as well? How long is it taking for your celeron to crunch a Seti task?
OK, this is obviously a "feature" of recent versions of BOINC. The most recent version I use is 6.2.15 where it is still CPU time.
Yes, it started around 6.4.7 that the CPU time column was changed to Elapsed Time.
I didn't like it as much and have therefore built my own BOINC Manager which shows both the CPU Time and Elapsed Time columns ticking away. It is backwards compatible to 6.4.5 for sure, perhaps earlier (the lowest anyone tested was 5.10.45 on which it didn't work).
It is available in 32bit .zip format at my Skydrive (also available as .rar and .7z, just change the extension on the URL, if you want to).
Now, I did add the boinc.dll into the archive, but this one is only needed if you have problems with the original boinc.dll... if no problem, then only use the boincmgr26.exe in the file. It's called boincmgr26 as I needed to differ from the then available 6.6 series, which went up to 6.6.20 at the time I put those files on my Skydrive. My latest builds are 6.6.76... ;-)
Einstein P4 appr 15 hours (but on this SETI is slow...)
Is this CPU time or elapsed time? If it's CPU time, it's pretty much as expected. If it's elapsed time that would make the CPU time around 10.5 hours (because of the 70% throttling) and that is pretty slick for a 2.66GHz P4.
Quote:
4-6 minutes slow (from memory)
I presume you are saying that the CPU time is around 4-6 minutes less than the wall clock time over a particular period. If you don't specify the period over which the measurement was taken, it's totally useless information. For example was the difference of 4-6 minutes measured over a 10min period, or perhaps a 1 hour period or perhaps a 100 hour period!! Let me have a guess - over 10 minutes of true time, you noticed that the CPU time only incremented by 6 minutes so that 4 minutes was "lost". Yep, that would be pretty consistent with 60% throttling.
Quote:
Use at most 60%, P4 70%
It's a pity you didn't mention right at the start that you have enabled throttling in your preferences. It explains a part of the reason why your celeron is having problems completing work within the deadline. It cannot be the whole reason however. At 60% throttling, a 30hr CPU time task would take around 50 hours of real time which is still way less than the 176 hour estimate that the figures you have previously supplied suggest. By the way, are the figures you have previously supplied, CPU time or elapsed time? Because of the throttling there should be a big difference.
Could you please do a simple observation over an accurate period of say 1 hour. Note the CPU time at the start and end of this one hour period and see if you get around a 36 minute difference. If so this would be entirely consistent with the 60% throttling. If something else is stealing CPU cycles then you will get way less than 36 minutes and this would then explain why your tasks are taking 176 hours to complete.
Quote:
No, 15-17 hours.
Considering the 60% throttling, 15-17 hours (if that is elapsed time) is probably not too bad for a seti task. I guess it depends if you are using the stock app or the 3rd party optimised ones.
Note: Disabled BOINC Screensaver-last 50% of task finished in less than 10 hours. (I never run this on P4.)
Shame, the one screensaver I really like...may I guess that throttling up will allow me to run it?
+See below
Quote:
Quote:
Einstein P4 appr 15 hours (but on this SETI is slow...)
Is this CPU time or elapsed time? If it's CPU time, it's pretty much as expected. If it's elapsed time that would make the CPU time around 10.5 hours (because of the 70% throttling) and that is pretty slick for a 2.66GHz P4.
Elapsed
Quote:
4-6 minutes slow (from memory)
I presume you are saying that the CPU time is around 4-6 minutes less than the wall clock time over a particular period. If you don't specify the period over which the measurement was taken, it's totally useless information. For example was the difference of 4-6 minutes measured over a 10min period, or perhaps a 1 hour period or perhaps a 100 hour period!! Let me have a guess - over 10 minutes of true time, you noticed that the CPU time only incremented by 6 minutes so that 4 minutes was "lost". Yep, that would be pretty consistent with 60% throttling.
*Correction: One Hour by my clock=37 Elapsed CPU minutes (Of course I may be confused again)
Quote:
Use at most 60%, P4 70%
It's a pity you didn't mention right at the start that you have enabled throttling in your preferences.
I agree, with hindsight. Concept Confusion, I am only digging into BOINC every 2nd to 3rd year..
It explains a part of the reason why your celeron is having problems completing work within the deadline. It cannot be the whole reason however. At 60% throttling, a 30hr CPU time task would take around 50 hours of real time which is still way less than the 176 hour estimate that the figures you have previously supplied suggest. By the way, are the figures you have previously supplied, CPU time or elapsed time? Because of the throttling there should be a big difference.
Elapsed (Again the Elapsed in Manager)
Could you please do a simple observation over an accurate period of say 1 hour. Note the CPU time at the start and end of this one hour period and see if you get around a 36 minute difference. If so this would be entirely consistent with the 60% throttling. If something else is stealing CPU cycles then you will get way less than 36 minutes and this would then explain why your tasks are taking 176 hours to complete.
It's beginning to dawn on me, see above*
Quote:
No, 15-17 hours.
Considering the 60% throttling, 15-17 hours (if that is elapsed time) is probably not too bad for a seti task. I guess it depends if you are using the stock app or the 3rd party optimised ones.
Actually I don't know what you mean by 3rd party-again my partial level of ignorance.
At least I have (re)learnt something here, and know where to find this conversation next time I get confused.
Thanks for your patience!
Note: Disabled BOINC Screensaver-last 50% of task finished in less than 10 hours. (I never run this on P4.)
So, everything is now explained. The task just completed on your celeron took 23 hours of actual CPU time. At 60% throttling, this would have translated to around 38 hours of elapsed time. By running the graphics continuously, the elapsed time blows out to around 176 hours, so it seems. I have never run the graphics previously because it is a well known resource hog. I honestly didn't realise just how bad it is, particularly on a weak processor like a celeron. I must apologise for being particularly dense. I had discounted the possibility that you might have been running the graphics. OK, you might run it for 5 minutes and say, "ahhh ... isn't that nice ....", but then you turn it off, right, and get back to more pressing things. I didn't imagine you would be allowing the graphics to run 24/7. As I said, my apologies for being so dense and not asking you about that.
Quote:
Shame, the one screensaver I really like...may I guess that throttling up will allow me to run it?
If it were my machine I'd disable the throttling completely. It shouldn't be needed on a desktop machine, particularly one with sufficient RAM like yours. There's no reason why you can't run the graphics whenever you want a "distraction". Just turn it off again whenever you are finished. It's not like you would sit staring at the screen for hours on end is it? :-).
Quote:
*Correction: One Hour by my clock=37 Elapsed CPU minutes (Of course I may be confused again)
Nope, you're not confused. That sounds pretty good. Of course you couldn't have been running the graphics during this test, otherwise the CPU time would have probably been around 10 mins rather than 37.
Quote:
Actually I don't know what you mean by 3rd party-again my partial level of ignorance.
As a quick translation, 3rd party = unofficial, ie applications developed by a third party and not the project staff. At Einstein, all apps are "official" although some very gifted "third parties" did assist with the optimisation. At seti, some very gifted "third parties" developed quite separate (and hence unofficial) apps that you can download and run under the anonymous platform mechanism of BOINC. These "unofficial" apps do give quite an increase in performance.
Quote:
At least I have (re)learnt something here, and know where to find this conversation next time I get confused.
Thanks for your patience!
You're most welcome!
The preferences you choose have quite an impact on the science output of your machine. At least you can make a much more informed choice now.
With regard to your CPU-Z screenshot, everything looks quite normal there. Good luck!
It is available in 32bit .zip format at my Skydrive ....
Very impressive!! However, I very rarely run BOINC Manager so extra features would be lost on me. I have around 200 machines on two separate LANs and I use a single BOINC Manager on each, occasionally, if I want to monitor any particular host. I've written a number of scripts which use boinccmd to interact with the running boinc client if I want to do any serious clent management. I have scripts that can update all or selected groups of hosts, suspend or resume particular projects on selected hosts, allow or deny new tasks on selected hosts, or manipulate network or run mode preferences, etc, so I find I don't need BOINC Manager much these days. Some scripts run automatically every day and create a log of the "health" of each host. As long as I force myself to review the log each day, I get to see straight away if any of the 200 hosts happens to be "playing up" at the moment.
In the recent E@H outage, I was able (once I realised there was an outage on) to use these scripts to prevent further loss of completed results to the dreaded bug affecting the 5.10.21 to 5.10.45 clients when trying to contact a non-existent file upload handler. Since then I've completed the transition away from this problematic range. All my linux hosts now run 6.2.15 and the majority of windows hosts run 5.10.20 so no more possibility of results loss, whatever happens to the file upload handler. The next time I'm likely to change anything to do with BOINC version will probably be when E@H releases a CUDA app and I decide to acquire a capable graphics card. At Richard's suggestion, I've been lurking on the alpha list for some time now and the messages there have confirmed my desire to stay well clear of 6.6.x for a while yet. :-).
RE: RE: OK I should now
)
I don't know what I don't know
RE: RE: Is your current
)
You answered an open question ("or") with "no" ;-)
And you should surround your answers with
Computer sind nicht alles im Leben. (Kleiner Scherz)
RE: OK, this is obviously a
)
Yes, it started around 6.4.7 that the CPU time column was changed to Elapsed Time.
I didn't like it as much and have therefore built my own BOINC Manager which shows both the CPU Time and Elapsed Time columns ticking away. It is backwards compatible to 6.4.5 for sure, perhaps earlier (the lowest anyone tested was 5.10.45 on which it didn't work).
It is available in 32bit .zip format at my Skydrive (also available as .rar and .7z, just change the extension on the URL, if you want to).
Now, I did add the boinc.dll into the archive, but this one is only needed if you have problems with the original boinc.dll... if no problem, then only use the boincmgr26.exe in the file. It's called boincmgr26 as I needed to differ from the then available 6.6 series, which went up to 6.6.20 at the time I put those files on my Skydrive. My latest builds are 6.6.76... ;-)
RE: Einstein P4 appr 15
)
Is this CPU time or elapsed time? If it's CPU time, it's pretty much as expected. If it's elapsed time that would make the CPU time around 10.5 hours (because of the 70% throttling) and that is pretty slick for a 2.66GHz P4.
I presume you are saying that the CPU time is around 4-6 minutes less than the wall clock time over a particular period. If you don't specify the period over which the measurement was taken, it's totally useless information. For example was the difference of 4-6 minutes measured over a 10min period, or perhaps a 1 hour period or perhaps a 100 hour period!! Let me have a guess - over 10 minutes of true time, you noticed that the CPU time only incremented by 6 minutes so that 4 minutes was "lost". Yep, that would be pretty consistent with 60% throttling.
It's a pity you didn't mention right at the start that you have enabled throttling in your preferences. It explains a part of the reason why your celeron is having problems completing work within the deadline. It cannot be the whole reason however. At 60% throttling, a 30hr CPU time task would take around 50 hours of real time which is still way less than the 176 hour estimate that the figures you have previously supplied suggest. By the way, are the figures you have previously supplied, CPU time or elapsed time? Because of the throttling there should be a big difference.
Could you please do a simple observation over an accurate period of say 1 hour. Note the CPU time at the start and end of this one hour period and see if you get around a 36 minute difference. If so this would be entirely consistent with the 60% throttling. If something else is stealing CPU cycles then you will get way less than 36 minutes and this would then explain why your tasks are taking 176 hours to complete.
Considering the 60% throttling, 15-17 hours (if that is elapsed time) is probably not too bad for a seti task. I guess it depends if you are using the stock app or the 3rd party optimised ones.
Cheers,
Gary.
Note: Disabled BOINC
)
Note: Disabled BOINC Screensaver-last 50% of task finished in less than 10 hours. (I never run this on P4.)
Shame, the one screensaver I really like...may I guess that throttling up will allow me to run it?
+See below
At least I have (re)learnt something here, and know where to find this conversation next time I get confused.
Thanks for your patience!
I don't know what I don't know
RE: Note: Disabled BOINC
)
So, everything is now explained. The task just completed on your celeron took 23 hours of actual CPU time. At 60% throttling, this would have translated to around 38 hours of elapsed time. By running the graphics continuously, the elapsed time blows out to around 176 hours, so it seems. I have never run the graphics previously because it is a well known resource hog. I honestly didn't realise just how bad it is, particularly on a weak processor like a celeron. I must apologise for being particularly dense. I had discounted the possibility that you might have been running the graphics. OK, you might run it for 5 minutes and say, "ahhh ... isn't that nice ....", but then you turn it off, right, and get back to more pressing things. I didn't imagine you would be allowing the graphics to run 24/7. As I said, my apologies for being so dense and not asking you about that.
If it were my machine I'd disable the throttling completely. It shouldn't be needed on a desktop machine, particularly one with sufficient RAM like yours. There's no reason why you can't run the graphics whenever you want a "distraction". Just turn it off again whenever you are finished. It's not like you would sit staring at the screen for hours on end is it? :-).
Nope, you're not confused. That sounds pretty good. Of course you couldn't have been running the graphics during this test, otherwise the CPU time would have probably been around 10 mins rather than 37.
As a quick translation, 3rd party = unofficial, ie applications developed by a third party and not the project staff. At Einstein, all apps are "official" although some very gifted "third parties" did assist with the optimisation. At seti, some very gifted "third parties" developed quite separate (and hence unofficial) apps that you can download and run under the anonymous platform mechanism of BOINC. These "unofficial" apps do give quite an increase in performance.
You're most welcome!
The preferences you choose have quite an impact on the science output of your machine. At least you can make a much more informed choice now.
With regard to your CPU-Z screenshot, everything looks quite normal there. Good luck!
Cheers,
Gary.
RE: ... built my own BOINC
)
Very impressive!! However, I very rarely run BOINC Manager so extra features would be lost on me. I have around 200 machines on two separate LANs and I use a single BOINC Manager on each, occasionally, if I want to monitor any particular host. I've written a number of scripts which use boinccmd to interact with the running boinc client if I want to do any serious clent management. I have scripts that can update all or selected groups of hosts, suspend or resume particular projects on selected hosts, allow or deny new tasks on selected hosts, or manipulate network or run mode preferences, etc, so I find I don't need BOINC Manager much these days. Some scripts run automatically every day and create a log of the "health" of each host. As long as I force myself to review the log each day, I get to see straight away if any of the 200 hosts happens to be "playing up" at the moment.
In the recent E@H outage, I was able (once I realised there was an outage on) to use these scripts to prevent further loss of completed results to the dreaded bug affecting the 5.10.21 to 5.10.45 clients when trying to contact a non-existent file upload handler. Since then I've completed the transition away from this problematic range. All my linux hosts now run 6.2.15 and the majority of windows hosts run 5.10.20 so no more possibility of results loss, whatever happens to the file upload handler. The next time I'm likely to change anything to do with BOINC version will probably be when E@H releases a CUDA app and I decide to acquire a capable graphics card. At Richard's suggestion, I've been lurking on the alpha list for some time now and the messages there have confirmed my desire to stay well clear of 6.6.x for a while yet. :-).
Cheers,
Gary.