[mythtv-users] Ratpoison on dedicated frontend

Michael T. Dean mtdean at thirdcontact.com
Tue Oct 10 18:34:11 UTC 2006


On 10/10/06 10:02, Steven Adeff wrote:

>On 10/10/06, Nowhere <nowhere at cox.net> wrote:
>  
>
>>Thanks for the info guys. I will play around with it. In the meantime, I
>>simply started the frontend from command line while running ratpoison wm,
>>and it turns out I still get the pauses. As it turns out the pauses don't
>>seem to be some buffering but more like dropped frames. The amount dropped
>>is always less than a second and happens at random intervals.
>>
>>Any ideas what might be the source since running ratpoison hasn't seemed to
>>improve it?
>>    
>>
>I've found that 0.19 and 0.20 run fine without a WM. The newer Xine
>lib properly supports this as does mplayer.
>  
>
Sure, Myth runs fine without a WM.  However, when you use more than one 
window, all bets are off--i.e. if you start up an xterm or an external 
media player or a browser or ...

Please don't run without a WM (or at least don't ask questions when 
things like keypresses/remotes don't work properly in that configuration 
;).  If you're not using a Desktop Environment (i.e. GNOME/KDE) and, 
instead, are just using a simple WM, it's not using much in the way of 
resources.  We're talking about a meg or two of RAM (or significantly 
less) if you choose properly.  You're much better off spending $20 on 
another 256MiB stick of RAM; after all, if that 1MiB of RAM is really 
necessary, you actually need more than just the 1MiB.

Mike

Oh, and if you want to check up on my values, remember that "ps aux" 
reports RSS/VSZ as if the process were the only running process on the 
system--it doesn't take into account memory requirements of shared 
libraries that are required for other processes.  Here's my "ps aux" 
output for my WM--fluxbox, which is not a particularly low-resource 
choice compared to, e.g., RatPoison

mythtv    2917  0.0  0.2  8108 1536 tty1     S    Oct09   0:01 fluxbox

So, ps says I'm using ~8MiB VSZ with ~1.5MiB RSS.

Here's the "pmap -d" output for the process:

2917:   fluxbox
Address   Kbytes Mode  Offset           Device    Mapping
08048000    1028 r-x-- 0000000000000000 003:00002 fluxbox
08149000     128 rw--- 0000000000101000 003:00002 fluxbox
08169000     556 rwx-- 0000000000000000 000:00000   [ anon ]
40000000      80 r-x-- 0000000000000000 003:00002 ld-2.3.2.so
40014000       4 rw--- 0000000000014000 003:00002 ld-2.3.2.so
40015000      32 r-x-- 0000000000000000 003:00002 libSM.so.6.0
4001d000       4 rw--- 0000000000007000 003:00002 libSM.so.6.0
4001e000       4 rw--- 0000000000000000 000:00000   [ anon ]
4001f000      80 r-x-- 0000000000000000 003:00002 libICE.so.6.3
40033000       4 rw--- 0000000000014000 003:00002 libICE.so.6.3
40034000       8 rw--- 0000000000000000 000:00000   [ anon ]
40036000      52 r-x-- 0000000000000000 003:00002 libXext.so.6.4
40043000       4 rw--- 000000000000d000 003:00002 libXext.so.6.4
40044000      68 r-x-- 0000000000000000 003:00002 libXft.so.2.1.2
40055000       4 rw--- 0000000000011000 003:00002 libXft.so.2.1.2
40056000       8 r-x-- 0000000000000000 003:00002 xlcDef.so.2
40058000       4 rw--- 0000000000001000 003:00002 xlcDef.so.2
40061000     140 r-x-- 0000000000000000 003:00002 libfontconfig.so.1.0.4
40084000      12 rw--- 0000000000023000 003:00002 libfontconfig.so.1.0.4
40087000       4 rw--- 0000000000000000 000:00000   [ anon ]
40088000     376 r-x-- 0000000000000000 003:00002 libfreetype.so.6.3.5
400e6000      28 rw--- 000000000005e000 003:00002 libfreetype.so.6.3.5
400ed000      68 r-x-- 0000000000000000 003:00002 libz.so.1.2.1
400fe000       4 rw--- 0000000000010000 003:00002 libz.so.1.2.1
400ff000       4 rw--- 0000000000000000 000:00000   [ anon ]
40100000     116 r-x-- 0000000000000000 003:00002 libexpat.so.0.5.0
4011d000      12 rw--- 000000000001c000 003:00002 libexpat.so.0.5.0
40120000      28 r-x-- 0000000000000000 003:00002 libXrender.so.1.2.2
40127000       4 rw--- 0000000000006000 003:00002 libXrender.so.1.2.2
40128000     796 r-x-- 0000000000000000 003:00002 libX11.so.6.2
401ef000      12 rw--- 00000000000c6000 003:00002 libX11.so.6.2
401f2000      56 r-x-- 0000000000000000 003:00002 libXpm.so.4.11
40200000       4 rw--- 000000000000d000 003:00002 libXpm.so.4.11
40201000      12 r-x-- 0000000000000000 003:00002 libXrandr.so.2.0
40204000       4 rw--- 0000000000002000 003:00002 libXrandr.so.2.0
40205000     628 r-x-- 0000000000000000 003:00002 libstdc++.so.5.0.5
402a2000      84 rw--- 000000000009d000 003:00002 libstdc++.so.5.0.5
402b7000      24 rw--- 0000000000000000 000:00000   [ anon ]
402bd000     132 r-x-- 0000000000000000 003:00002 libm-2.3.2.so
402de000       4 rw--- 0000000000020000 003:00002 libm-2.3.2.so
402df000      28 r-x-- 0000000000000000 003:00002 libgcc_s.so.1
402e6000       4 rw--- 0000000000007000 003:00002 libgcc_s.so.1
402e7000    1188 r-x-- 0000000000000000 003:00002 libc-2.3.2.so
40410000      16 rw--- 0000000000128000 003:00002 libc-2.3.2.so
40414000      12 rw--- 0000000000000000 000:00000   [ anon ]
40417000       8 r-x-- 0000000000000000 003:00002 libdl-2.3.2.so
40419000       4 rw--- 0000000000001000 003:00002 libdl-2.3.2.so
4041a000       4 rw--- 0000000000000000 000:00000   [ anon ]
4041b000    2048 r---- 0000000000000000 003:00002 locale-archive
4061b000      36 r-x-- 0000000000000000 003:00002 xomGeneric.so.2
40624000       4 rw--- 0000000000008000 003:00002 xomGeneric.so.2
40625000      32 r-x-- 0000000000000000 003:00002 libXcursor.so.1.0
4062d000       4 rw--- 0000000000007000 003:00002 libXcursor.so.1.0
4062e000      68 r---- 0000000000000000 003:00002 Vera.ttf
bfff8000      32 rwx-- 00000000ffff9000 000:00000   [ stack ]
mapped: 8108K    writeable/private: 1000K    shared: 0K

If you look through the memory map, you'll note that the vast majority 
of that 8MB is in shared libs--which happen to be in use by other 
processes that you/I will run on the system, anyway.  Therefore, the 
true cost of running this WM is 1000KiB (the writable/private size 
reported at the bottom)--not the 8MB or 1.5MB that ps said.  That's less 
than 1MiB!!!

I'm guessing RatPoisong would be even lower values.  Any RatPoison users 
want to post a ps/pmap to help convince people to use a WM?

(Yeah, I know these programs/libraries are ancient versions, but I'm 
working on that issue right now... :)

Q: How can I save a huge amount of memory in my X windows environment?
A: Don't run a WM.  It could save you over a million bytes of memory!!!
Q: Isn't that just /one/ megabyte.
A: Ummm.  Yeah.  But, "one megabyte," doesn't sound like much.  I'm 
trying to sell the idea.  Geeez.


More information about the mythtv-users mailing list