> >Also, I was wondering if anyone knew how much faster of slower
> >(I hope not!)
> >Linux runs compared to NT/Win2K on dual processor machines. I
> >heard someone tell
> >me between 60-80% of the combined speeds of the two chips.
>
> 2.4 has support for the A7V built in. To install on the ATA-100 with 2.2,
> just get the linux-ide patches and install with the drive connected to the
> ATA-66 interface. Then move the drive to the ATA-100 and it should just
> work (haven't tried this myself yet).
Just grab the patches, you might need to turn on the experiment stuff to
allow you to get at them.
> Performance on both Linux and NT/Win2k on SMP boxes depends heavily on the
> application. I haven't had much experience with SMP Linux, but 2 CPUs on
> NT/Win2k goes anywhere from no improvement to over 99% improvement over 1
> CPU. I've heard similar results for Linux.
Depends how threaded the task is, if your running more than one single
threaded task youll see a performance increase. Seti (for example) is
single threaded (when i was into it) so if you wanted smp benefits you
ran two
copies. Threads are divided between cpus. However 2xcpu <> 2xperformance
> Note that 2.4 is a lot better with SMP than 2.2 is. There are many more
> locks removed from the kernel.
More locks where added i believe, removing them would make things worse,
as only one cpu can access one subsystem (eg. the network or the file
system)
by spliting them into more bits you allow the other cpu to access them
instead of waiting.
Its worth noting that some hardware (and/or drivers) disagree with
smp performance. ISA stuff for example. Stick with new pci stuff.
Dean
--
BONG: http://www.bong.com.au
EMAIL...
[EMAIL PROTECTED] [EMAIL PROTECTED]
[EMAIL PROTECTED] [EMAIL PROTECTED]
ICQ: 16867613
--
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug
