artswrapper defanged

[Ryan Cumming]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On July 12, 2002 06:01, Neil Stevens wrote:
> Well, my system isn't the ideal for testing: 1GHz PIII, 512 MB RAM, good
> Matrox Xv support, good disk IO, preemptive kernel.  A system built for
> KDE development.  I'll try, though.

I thought the point was to max out the CPU? Good disk I/O will decrease the 
time mpeglib is waiting for I/O, and increase the time it's actually spending 
decoding. Having lots of memory will decrease the time mpeglib is blocked 
because of swap operations, and increase the time it spends actually 
decoding. And if you think about the nature of realtime processes, lower 
priority process will -never- preempt a realtime process, but a realtime 
process will -always- preempt a lower priority process (if it's runnable). So 
the preempt patch will also amplify the CPU drain. The only part of your 
system that isn't ideal to being ripped apart by realtime processes is the 
CPU.

- -Ryan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQE9L1KILGMzRzbJfbQRAifPAJoCV/MFAkuCtGOrLXkRrh14ye2tKwCfXxKU
pHOmXLkEYrcSfCTLxR6ynys=
=bx8z
-----END PGP SIGNATURE-----