[mythtv-users] The Ultimate Myth Setup

Eyal Lebedinsky eyal at eyal.emu.id.au
Sun Oct 24 23:17:21 UTC 2004


Neil Davidson wrote:
> The front ends can be a little as an Epia M6000 i believe, but an
> M10000 would probably be better. Have you looked at MiniMyth
> (http://www.linpvr.org) for front end ideas? even if you don't use MiniMyth
> there is some very good info on the forum there. Diskless, network booting,
> low power, zero noise frontends are what you should go for in my opinion :)

If you plan on watching HDTV then you will find the frontend needing much
more CPU power that is suggested above. I have the back+frontend on the one
machine, so I will present the results accordingly. This is a P4/3GHz.

# vmstat 5
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu-----
  r  b   swpd   free   buff  cache   si   so    bi    bo   in    cs us sy  id wa

Frontend idle:
  0  0 343996 524056   8036 125460    0    0     0     0 1102  1426  0  0 100  0

Frontend tuned to SDTV, but not watched (backend):
  0  0 343996  74064   8916 561920    0    0     0   998 2574  6546 17  3  80  1

Frontend tuned to SDTV and visible (backend+frontend):
  0  0 343996   5260   8916 629580    0    0     0  1003 2546  4762 39  4  56  1

Looks like 20% used by the backend, 23% by the front. I count us+sy.

56% idle - no big deal. Now with HDTV. Tuned, not watched (backend):
  0  0 343996   3484   1124 631296    0    0  1410  1231 2708  6192 41  3  53  4

Watched  (backend+frontend):
  2  1 343996   2752    984 632184    0    0   470  1266 2712  2793 96  4   0  0

Oh boy. Modern physics tells us that watching has an effect on the experiment :-)
My CPU is really busy with HDTV, looks like 44% used by the backend, 56% by
the front. The frontend actually has trouble keeping up.

FYI

-- 
Eyal Lebedinsky	 (eyal at eyal.emu.id.au) <http://samba.org/eyal/>


More information about the mythtv-users mailing list