lkml.org 
[lkml]   [1998]   [Nov]   [30]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: SMP scalability: 8 -> 32 CPUs
On Sat, 28 Nov 1998, Kurt Garloff wrote:
> On Sat, Nov 28, 1998 at 12:50:03AM +0000, Eloy A. Paris wrote:
> > Guys, excuse my ignorance, but does anyone know why Linus said in his
> > last Linux Journal interview that right now it is fairly unrealistic
> > to expect scaling from 8 to 32 CPUs? What's stopping Linux to scale in
> > terms of CPUs?
>
> Basically there are different mechanisms which limit the speed of operation:
> (1) CPU(s) speed
> (2) Memory speed (and size)
> (3) I/O (Network, HD, VGA, ...)
>
> By increasing the number of CPUs, you only do something about (1).
> So only CPU bound processes may profit.
>
> If you have a large number of independent CPU bound processes, they
> will scale very nicely until they altogether hit the Memory or I/O
> bound.

Number (2) can be tackled with a number of scheduler adjustments:
- longer time slices (for CPU-bound tasks only!)
- make it harder to switch CPUs (depending on how long a process
has slept)
- more intelligent rescheduling in reschedule_idle() -- better
burn a few hundred cycles there than completely mess up the
cache of three CPUs and hogging the memory bus
- in tune with this, move the bottleneck in other pieces of
code from memory to CPU, better cache alignment of task
struct and stuff

If you have 8 CPUs, 8 CPU-bound tasks and one process
taking 10% of one CPU, it might just be better to have
one of the large tasks get 90% and the rest 100%.

Switching around the one non-hog might create such an
amount of cache/memory problems that it will make the
CPU hog suffer more than the 10% in question.

Weird counterintuitive things like this are what make
the difference between a high-performance OS and your
usual Redmond crap...

cheers,

Rik -- who prefers a certain OS with fairly non-intuitive innards.
+-------------------------------------------------------------------+
| Linux memory management tour guide. H.H.vanRiel@phys.uu.nl |
| Scouting Vries cubscout leader. http://www.phys.uu.nl/~riel/ |
+-------------------------------------------------------------------+



-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:45    [W:0.080 / U:0.300 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site