On Mon, Feb 14, 2005 at 10:39:22AM -0600, Neil wrote:
> Brad Templeton writes:
> >
> >On Athlon-3000s people typically see about 80% usage. On P4-3ghz we see
> >more like 55% usage but that number is probably a touch misleading
> >due to hyperthreading.
>
>
> Hi Brad,
>
> Just curious since I bought an Athlon 64 3000+ cpu. I'm also waiting for my
> 6600 GT nvidia graphics adapter. You are saying that when I start playing
> HD, my cpu utilization will hit 80% too? This value is bad! Please confirm.
I have not tested the Athlon 64. The Athlon 64 runs at a slightly lower
clock than the Athlon XP, but has a more efficient core. But it will
probably be similar.
The 6600, which is a way overkill card for this purpose, might gain you
a slight amount over the 5200, but I don't know if it does. I presume you
do a lot of heavy 3D gaming if you bought that card, since as far as I
know that's the only reason to get it.
So far, my experiments show the P4 outdoing the Athlon handily, though
the better memory bandwidth on the P4 may be part of this. Or it may
be the hyperthreading and the siginficantly faster clock speed Intel chips
run at (normally a downside.)
Now 80% usage isn't necessarily bad, unless you have something else you
need the CPU to be doing!
My new benchmark is as follows:
a) Play an HD Video.
b) While playing, run mythcommflag on another video and note the
frames per second. Also note idle times.
My quick tests for a P4-3ghz, hyperthreaded, PC 3200 memory are:
a) Play 720p video, commflag SD video -- 85 frames/second, 35% "idle"
b) Play 1080i video + deinterlace, commflag SD -- 42 frames/second, 20% idle
HD video coming in over ethernet from master backend, SD video located
locally.
_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
