Grzegorz Nieweglowski <[EMAIL PROTECTED]> writes:
> Dominik Vogt wrote:
> > Any decent kernel is clever enough not to load multiple copies of the
> > same write-only data.
>
> I'm using the linux kernel, 2.4.21.
>
> Results of the "free" command with 100 instances of FvwmPager running:
>
> total used free shared buffers cached
> Mem: 224728 131148 93580 0 0 18396
> -/+ buffers/cache: 112752 111976
> Swap: 51192 1372 49820
>
> And the same after "killall FvwmPager":
>
> total used free shared buffers cached
> Mem: 224728 70896 153832 0 0 18416
> -/+ buffers/cache: 52480 172248
> Swap: 51192 1372 49820
>
> I guess that with 'a few hundreds' per module te amount of memory freed
> after killing all those pagers should change by a couple of megabytes in
> worst case , not by ~60MBs. And I believe these readings are correct
> - starting more modules (200 and up) chewed up all of RAM and swap, and
> the infamous OOM killer started killing processes.
>
> Either it's my config, or running two pagers uses more ram than just "(1
> FvwmPager) + (less than 1KB)"
Here is my calculation of what the pager uses. This is on Solaris 8
using the "pmap" command. This is the output, slightly reformattted:
3490: /opt/public/src//fvwm/2.5.7/FvwmPager 10 7 none 4800003 1 0 1
00052000 16K read/write/exec /opt/public/src/fvwm/2.5.7/FvwmPager
00056000 72K read/write/exec [ heap ]
FF036000 24K read/write/exec /usr/lib/libc.so.1
FF03C000 8K read/write/exec /usr/lib/libc.so.1
FF072000 8K read/write/exec /usr/openwin/lib/libdga.so.1
FF11A000 32K read/write/exec /usr/lib/libnsl.so.1
FF122000 32K read/write/exec /usr/lib/libnsl.so.1
FF154000 8K read/write/exec /usr/lib/libmp.so.2
FF170000 8K read/write/exec [ anon ]
FF21C000 24K read/write/exec /usr/openwin/lib/libX11.so.4
FF254000 8K read/write/exec /usr/lib/libm.so.1
FF27A000 8K read/write/exec /usr/lib/libsocket.so.1
FF2A8000 8K read/write/exec /usr/openwin/lib/libXext.so.0
FF2D4000 8K read/write/exec /usr/openwin/lib/libICE.so.6
FF2D6000 8K read/write/exec /usr/openwin/lib/libICE.so.6
FF2FA000 8K read/write/exec /usr/openwin/lib/libSM.so.6
FF31C000 8K read/write/exec /usr/lib/libz.so.1
FF366000 8K read/write/exec /opt/public/lib/libpng.so.3.1.2.5
FF38C000 16K read/write/exec /opt/public/lib/libXpm.so.4.10
FF3B0000 8K read/write/exec [ anon ]
FF3E0000 8K read/write/exec /usr/lib/ld.so.1
FFBE8000 32K read/write/exec [ stack ]
360K
00010000 208K read/exec /opt/public/src/fvwm/2.5.7/FvwmPager
FEF70000 16K read/exec /usr/platform/sun4u/lib/libc_psr.so.1
FEF80000 664K read/exec /usr/lib/libc.so.1
FF050000 72K read/exec /usr/openwin/lib/libdga.so.1
FF080000 552K read/exec /usr/lib/libnsl.so.1
FF140000 16K read/exec /usr/lib/libmp.so.2
FF180000 560K read/exec /usr/openwin/lib/libX11.so.4
FF230000 88K read/exec /usr/lib/libm.so.1
FF260000 40K read/exec /usr/lib/libsocket.so.1
FF280000 96K read/exec /usr/openwin/lib/libXext.so.0
FF2B0000 80K read/exec /usr/openwin/lib/libICE.so.6
FF2E0000 40K read/exec /usr/openwin/lib/libSM.so.6
FF300000 48K read/exec /usr/lib/libz.so.1
FF330000 160K read/exec /opt/public/lib/libpng.so.3.1.2.5
FF370000 56K read/exec /opt/public/lib/libXpm.so.4.10
FF3A0000 8K read/exec /usr/lib/libdl.so.1
FF3C0000 128K read/exec /usr/lib/ld.so.1
2832K
total 3192K
The total memory is 3meg.
I separated that into read/write and read/exec.
The read/exec is common memory (its only used once no matter how
often the pager is loaded). Of the 3meg, thats 2.8 meg thats common.
The balance is 360K.
So, I'd say that the each instance of the pager takes 360K.
Probably what is most important is how much of this actually resides
in memory. That could be a lot less than 360K.
Another issue is how many libaries got mapped into memory that
I wasn't already using. I this case, I think I have all of them
already mapped.
One of the nice things about Fvwm is how much memory the pager
consumes for me. (Zero, I don't use it.)
--
Dan Espen E-mail: [EMAIL PROTECTED]
--
Visit the official FVWM web page at <URL:http://www.fvwm.org/>.
To unsubscribe from the list, send "unsubscribe fvwm-workers" in the
body of a message to [EMAIL PROTECTED]
To report problems, send mail to [EMAIL PROTECTED]
