Date: Thu, 16 Jul 2020 10:50:17 +0200 From: "Kristof Provost" <kp@FreeBSD.org> To: "Patrick Lamaiziere" <patfbsd@davenulle.org> Cc: "Eugene Grosbein" <eugen@grosbein.net>, freebsd-net@freebsd.org Subject: Re: poor performance with Intel X520 card Message-ID: <0B227515-A53A-4354-86C5-EADFEDC65CAD@FreeBSD.org> In-Reply-To: <20200716095732.728c0917@mr185033.univ-rennes1.fr> References: <20200710084530.777ce321@mr185033.univ-rennes1.fr> <7b994967-17cd-8f4f-cb4d-8fcff349f7e9@grosbein.net> <20200716095732.728c0917@mr185033.univ-rennes1.fr>
next in thread | previous in thread | raw e-mail | index | archive | help
On 16 Jul 2020, at 9:57, Patrick Lamaiziere wrote: > On Wed, 15 Jul 2020 17:07:23 +0700 > Eugene Grosbein <eugen@grosbein.net> wrote: > > Hello, > >>> That is mostly for the record but it looks like the intel X520 is >>> not very good and generates a high level of interrupts. >>> >>> On a router / firewall with 500 Kpps in input (dropped by pf) is >>> enough to put the CPUs at 100% busy. >> >> [skip] >> >>> Well, do you think another NIC cards can help to reach a better pps >>> rate ? I think 500 Kpps is quite low for such a machine. >> >> I'm sure pf is the bottle-neck. Try testing such card without any >> packet filter enabled and you'll see great difference definitely. > > That's not a good news as I don't see how to simplify the ruleset :( > But thanks anyway :) > I’d strongly recommend that you look at Olivier’s recommended tools (specifically pmcstat) to try to diagnose the bottleneck. I see no reason to assume this must be pf (it’s possible, but I see no specific evidence for it). On this hardware I’d expect pf to be able to push around 3.5Mpps. Even a single core system (or in a situation where you end up with a lock of lock contention) it should be able to do 1.5-1.8Mpps. Best regards, Kristof
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?0B227515-A53A-4354-86C5-EADFEDC65CAD>