Re: forwarding/ipfw/pf evolution (in pps) on -current

On 25 April 2013 02:28, Andre Oppermann <andre_at_freebsd.org> wrote:

> Again one has to be really careful drawing any firm conclusions from this
> as it was measured on a Pentium4 and UP kernel (GENERIC would add WITNESS
> and INVARIANT overhead as well).
>
> The Pentium4 is about the worst micro-architecture when it comes to locks
> and easily regresses.  At the same time modern Intel Core i[3-7] and AMD64
> may actually improve with these changes.  Unless more recent micro-archs
> have been shown to exhibit the same regression we can't claim this change
> was bad (other than for Pentium4).

Sure, but he's done the heavy lifting.

It'll be interesting to compare these results on a variety of
platforms, not just the modern desktop/server style CPUs.

Eg, if someone has the time, spinning this stuff up on the multi-core
MIPS stuff in the netperf cluster (that's supposed to be a network
forwarding engine) would be nice.

And to be honest - having a set of performance checks for the same SVN
revision but different physical machines is a good comparison point.
It may be that we can start classifying different kinds of platform
silliness from this which could lead to some better coding guidelines.


Adrian