Date: Fri, 03 Nov 2006 07:54:21 +0100 From: Joerg Sonnenberger <joerg@britannica.bec.de> To: freebsd-hackers@freebsd.org, hackers@freebsd.org Subject: Re: vr(4) performance Message-ID: <20061103065421.GA829@britannica.bec.de> In-Reply-To: <9ab217670611021511l3120d58bhd0b61bf44f8ecc87@mail.gmail.com> References: <9ab217670611021511l3120d58bhd0b61bf44f8ecc87@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Nov 02, 2006 at 06:11:00PM -0500, Devon H. O'Dell wrote: > So, vr(4) kind of sucks, and it seems like this is mostly due to the > fact that we call m_defrag() on every mbuf that we send through it. > This seems to really screw performance on outgoing packets (something > like 33% the output efficiency of fxp(4), if I'm understanding this > all correctly). What hardware are you running at that it is a problem? Seriously -- copying 100Mbit/s in memory shouldn't fully busy any post-2k non-embedded CPU. You could just allocate a static output queue and copy into that, dropping at least the allocations of the mbuf clusters, but I don't think it should b worth it. And yes, if you can sustain Fast Ethernet speed with scp to localhost, m_defrag is absolutely not the main problem. Joerg
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20061103065421.GA829>