OK,  I set up to test this based on Brad's suggestions to first record a
program, then to baseline mythcommflag's fps performance, the to watch a
program while again running mythcommflag against anoither recorded show of the
same resolution.  In this case I used 720p shows recorded off FOX.

Here's the baseline,...but it gets more interesting,...

mythcommflag
2005-02-16 02:46:37.067 mythcommflag version: 0.17.20050130-1 www.mythtv.org
2005-02-16 02:46:37.068 Enabled verbose msgs :important

MythTV Commercial Flagger, started at Wed Feb 16 02:46:37 2005

Flagging commercial breaks for:
ALL Un-flagged programs
ChanID  Start Time      Title                                      Breaks
------  --------------  -----------------------------------------  ------
1031    20050215200000  House

 99%/154fps2005-02-16 03:09:52.080

Finished commercial break flagging at Wed Feb 16 03:09:53 2005

---

While running against another 720p show while watching a 1080i ( 9 news here
locally), I was seeing about 80/fps and a nearly solid on disk light on my
system.  So I ran TOP ( ya, I know it's not too accurate but it's close
enough) and noted that IDLE time was bouncing between 11 and 34% .


top - 23:32:03 up 27 min,  2 users,  load average: 3.06, 3.14, 2.62
Tasks:  87 total,   2 running,  85 sleeping,   0 stopped,  0 zombie
Cpu(s): 46.2% us,  1.7% sy, 38.6% ni, 12.2% id,  1.3% wa,  0.0% hi,  0.0%
Mem:    516036k total,   512492k used,     3544k free,     1288k buffers
Swap:  1092380k total,   161928k used,   930452k free,   328156k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 4622 root      32  17  188m 101m  23m R 39.1 20.2  10:38.12 mythcommflag
 4678 root      15   0  207m  59m 180m S  8.3 11.8   1:44.02 X
 4975 root      15   0  219m  74m  91m S  0.7 14.7   0:02.60 mythfrontend

 This strongly suggests that these processes are more disk bound ( or Bus
Bound) than processor bound, at least in my case.  The disk channel ran out of
steam long before the Processor and Graphics card did.  My guess is that if my
processor wasn't pegged, the 6600GT probably wasn't either.  My basic system
specs are below;

 MSI 865PE Mobo with a 2.8GHZ Proc
 512MB ddr-333 Memory
 ATA/100 Disks ( Seagate 120GB and 160GB)
 nVidia 6600GT Graphics Card
 pcHDTV HD-3000 HD Tuner Card.

 I've been meaning to pick up a couple of SATA/150 drives since this Mobo
supports 2.  That should help the PCI channels breath a bit and relieve some
of the disk bottleneck. We'll see how well MSI's controller mange this I
guess.  At any rate, I'll rerun the test against the faster drives in a few
weeks if folks are interested.

 So far though, I am very impressed with what the 6600 brings to the table.
I'd love to find the ceiling but think that new drivers from nVidia and a
stronger disk channel are going to be minimum to get it there.  I'm avioding
overclocking though I might start jackin with that if the disk channel doen't
result in slamming the proc.

Anyone else have testing idea's?

 -Blair

_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users

Reply via email to