|I noticed that, after running for a while, the muon client uses a substantial amount of RAM (easily exceeding 200MB in the Windows Task Manager). Is that normal behaviour or is there something I can do? Is there an upper limit of memory usage (besides the obvious physical RAM limit)? I looked at the results.txt and it was only 100kB, though the client used 130MB RAM.|
|If the number of results has something to do with it, than you should check the size of results.dat, not .txt|
However, I'm pretty sure that it has nothing to do with that, since I've never seen this.
Some time ago another user found that there was a memory leak which Stephen then fixed.
Are you using the latest version of the client?
|Yes, I am using the latest version.. my results.dat is only 22 MB, so nowhere near the RAM usage|
|Oh, I just noticed: After a recent restart, the client has 27 minutes CPU time and is already using 57 MB RAM... I don't know if that means anything...|
|I have noticed the memory usage also. After a few days of running, mine will be using upwards of 1gig+ of my RAM. I just end task on it and restart it if I feel it's grabbed too much.|
|Using v4.41a with a 7.709kB results.dat, my client uses 34.624kB of RAM after 10:15 hours.|
So I guess this does not happen with every client/configuration.
(I have a P3 933Mhz with WinXP Professional and the client running as a service using a tool some other DPAD user contributed.)
|28 hours CPU Time, 22 MB dat file and the client uses 125 MB. That is on XP Pro (SP1), running the background client on a 1.8GHz P4.|
|The memory usage will be a few times the size of the results.dat file, since when it is loaded into memory it is "unpacked" somewhat. However I think some of the figures here (the ones approaching 1G indicate some sort of real leak. I can test for that before I release another version, though that might not be for a while as I'm studying the accuracy of the simulation right now (trying to improve it).|
|I have the same problem - actually, I can watch it rise in the Task Manager by about 5k / second (it seems to exhibit this behavior without stopping). Right now, it's taking up 130 megs.|
|ya, ive noticed the same thing, however, my computer just died so im afraid that i cant give anymore details.|
|Same here, it,s going up and up till all is used.|
When i use my memory manager/cleaner , it wil go down to 14 Mb use .
In 1 minute time it went to 33 Mb and is stil going up.
|If it helps troubleshooting. I just started muon last friday morning. I've run 24/7 for 6 days so far with it except a restart for windows update and another proggie install. My result.dat file is 4,766 kb and my memory usage is 73,112 k and I can watch it creap up 4k about every 5 seconds in the task manager.|
At the finish of typing this post it's using 73,172.
|Oh sorry. XP Pro. SP1a. Dual Barton core processors.|
|Another oopsy client 4.4. I'm in the stats under 'spldart'|
|Heh! I never bothered to add Mem Usage to Task Manager.|
I'm using the v441b client. Like the man said, memory usage rises at about 4KB per 3 second period.
52,788 K with 11:48:13 CPU Time.
|Does the memory usage reset at some point?|
I awoke this morning and memory usage was under 49,000 K. It sent in the results and Task Manager says muon1.exe Mem Usage is 58,688 K, 23:25:50 CPU Time.
|It must, but I don't know when. I enabled the "peak memory usage" column in WinXP task manager and it shows a peak of 47,437 kb but memory usage is only at 27,136 (and going up slowly).|
|Note that it will typically increase during a simulation, but then reset when it's finished, as the number of particles increases as they decay (and the dead ones are stored right now).|
What would be worrying is if there was a leak between successive simulations. I'm not sure if you can check that yourselves - I certainly have debug/tested for such a leak in the past and made sure it didn't happen for the settings I used. That was a couple of minor versions ago.
|hmm I am seeing nothing like this: I have 2 processes running on this machine: |
process CPU time, Mem use Mem peak
1. 401:57:40 34,928k 53,240k
2. 402:07:48 33,920k 52,896k
I had problems with a previous version runing out of memeory after about 5 days, nothing here