FC4 kernel performance

Colin Walters walters at redhat.com
Tue Jun 21 15:54:05 UTC 2005


On Tue, 2005-06-21 at 11:43 -0400, Malita, Florin wrote:

> Regarding the configuration component, I can understand why certain
> features and the overhead associated with them are preferred vs raw
> kernel performance. OTOH, leaving 62% on the table makes me feel uneasy.
> Do I really need high mem, SE Linux or a debug-enabled kernel on my
> desktop? Don't think so. 

Ah, but is Unixbench a relevant benchmark for a desktop?  For example
making process creation 15% faster isn't very useful if process creation
accounts for .001% of application startup time.

A server's a different story, but there e.g. highmem and SELinux are
very useful.

> But I do want the kernel preemption enabled...

The kernel team has that disabled because it's unsafe, IIRC.

> My point is: with so many kernel features, "one size fits all" doesn't
> hold anymore and maybe we should have a much broader array of kernels to
> choose from at install time (not just architecture/SMP variants). This
> should be fairly easy to support as it's just a matter of adding new
> build configurations in the kernel SRPM/spec. 

I'm skeptical that (apart from preemption) having e.g. a "desktop"
kernel would be useful.  We have a lot of optimization we could do (and
are doing) in userspace.





More information about the fedora-devel-list mailing list