Hello,
regarding the 30 secs glitch of Matlab when running the Timing test script:
Has anybody checked if the "every 30 secs break" could be caused by the
update - daemon of the operating system?
MacOS - X, like most Unix operating sytems, has a system process called
"update" (see man update). This process is started at system boot time,
wakes up every 30 seconds by default on MacOS (as far as i remember) and
calls the sync() system call in order to flush dirty unwritten in memory
filesystem disc buffers back to the harddisc. The purpose of it is to
loose less data in case of a system crash/unclean shutdown/power failure.
The amount of work done / delay caused by this would depend on memory
configuration, amount of unwritten buffers and basically interactions
between the OS memory management and the memory/filesystem
operations/access pattern of any running process. --> Could be fairly
random.
Someone with the software installed - and administrator rights - could
check this by:
1. Login as root/admin.
2. kill the update process (something like sudo killall update)
3. Check via the top command, if it is really dead and doesn't show up
in the process listing.
4. Issue a few sync - commands in the shell.
5. Run Matlab timing scripts.
6. *restart* the update process (sudo update) and make sure it runs again.
7. Check if the timing glitch pattern goes away or at least changes.
Just a wild guess, but easy to check quickly. I would be interested in
the results but can't do it at our own box at the moment, because we
don't have a Matlab license for OS-X yet.
ciao
-mario
--
*********************************************************************
Mario Kleiner
Max Planck Institute for Biological Cybernetics
Spemannstr. 38
72076 Tuebingen
Germany
e-mail: mario.kleiner@...
office: +49 (0)7071/601-734
fax: +49 (0)7071/601-616
www: http://www.kyb.tuebingen.mpg.de/bu/people/kleinerm/index.html
Phone@Home: (0)7071/942494
*********************************************************************
Akademische Bildung verringert nicht die Neigung zu Vernunftwidrigkeiten,
sondern potenziert sie. (Mannfred Rommel, Ex-OB Stuttgart)
regarding the 30 secs glitch of Matlab when running the Timing test script:
Has anybody checked if the "every 30 secs break" could be caused by the
update - daemon of the operating system?
MacOS - X, like most Unix operating sytems, has a system process called
"update" (see man update). This process is started at system boot time,
wakes up every 30 seconds by default on MacOS (as far as i remember) and
calls the sync() system call in order to flush dirty unwritten in memory
filesystem disc buffers back to the harddisc. The purpose of it is to
loose less data in case of a system crash/unclean shutdown/power failure.
The amount of work done / delay caused by this would depend on memory
configuration, amount of unwritten buffers and basically interactions
between the OS memory management and the memory/filesystem
operations/access pattern of any running process. --> Could be fairly
random.
Someone with the software installed - and administrator rights - could
check this by:
1. Login as root/admin.
2. kill the update process (something like sudo killall update)
3. Check via the top command, if it is really dead and doesn't show up
in the process listing.
4. Issue a few sync - commands in the shell.
5. Run Matlab timing scripts.
6. *restart* the update process (sudo update) and make sure it runs again.
7. Check if the timing glitch pattern goes away or at least changes.
Just a wild guess, but easy to check quickly. I would be interested in
the results but can't do it at our own box at the moment, because we
don't have a Matlab license for OS-X yet.
ciao
-mario
--
*********************************************************************
Mario Kleiner
Max Planck Institute for Biological Cybernetics
Spemannstr. 38
72076 Tuebingen
Germany
e-mail: mario.kleiner@...
office: +49 (0)7071/601-734
fax: +49 (0)7071/601-616
www: http://www.kyb.tuebingen.mpg.de/bu/people/kleinerm/index.html
Phone@Home: (0)7071/942494
*********************************************************************
Akademische Bildung verringert nicht die Neigung zu Vernunftwidrigkeiten,
sondern potenziert sie. (Mannfred Rommel, Ex-OB Stuttgart)