Re: andrea buffer code (2.2.9-C.gz)

Chuck Lever (cel@monkey.org)
Sat, 22 May 1999 14:25:14 -0400 (EDT)

On Fri, 21 May 1999, Manfred Spraul wrote:
> Andrea wrote:
> > >number of spinlocks and choose which to use with some reductive function?
> > I think that if you want to SMP scale better it worth to
> > pay _only_ with memory wastage and _not_ with wasted CPU cycles.
> This is not always true: as Ingo pointed out, more memory
> means more cache misses, and that costs CPU cycles as well.
> The sometimes the reductive function could be as simple as
> (x&0xFF).

seems to me you want to scale the number of spinlocks with the number of
runnable processes that might want to concurrently access your data
objects. having 1K spinlocks is probably overkill if you only have a
dozen processes trying to access all of your mutually exclusive regions.

- Chuck Lever

--
corporate:	<chuckl@netscape.com>
personal:	<chucklever@netscape.net> or <cel@monkey.org>


The Linux Scalability project:
	http://www.citi.umich.edu/projects/linux-scalability/




-
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/