---------- Forwarded message ---------- From: Adam Stylinski <kungfujesu...@gmail.com> Date: Thu, Apr 28, 2011 at 11:21 AM Subject: Re: em0 performance subpar To: Steven Hartland <kill...@multiplay.co.uk>
On Thu, Apr 28, 2011 at 04:08:31PM +0100, Steven Hartland wrote: > You said your testing with iperf, what settings are you using? > > Flow control is not flowtable no, which could still result in > a switch "issue" if linux and freebsd are setting different > values by default, similarly with duplex / speed, I would recommend > autoneg if your not already using. > > Have you tried without tso, rxcsum, txcsum & lro disabled? > > > ----- Original Message ----- > From: "Adam Stylinski" <kungfujesu...@gmail.com> > To: "Steven Hartland" <kill...@multiplay.co.uk> > Cc: <freebsd-net@freebsd.org> > Sent: Thursday, April 28, 2011 3:45 PM > Subject: Re: em0 performance subpar > > I was using the default value for maxsockbuf, doesn't seem to hurt or help. The only value you've listed that I didn't have > tweaked to a comparable value already was the nmbclusters, changing it seems to have no effect, though. vmstat -i should display > my interrupts, which I don't think are particularly high. I am most definitely not CPU bound (utilization shows no hot CPUs), I > have 16GB of memory and plenty allocated to the kernel, and I believe flow control is enabled (should it not be?): > net.inet.flowtable.enable: 1 > > It's a 16 port netgear gigabit prosafe switch, and it's not likely to be the problem (like I said, the Linux hosts have no issues > with the same card). > > > ================================================ > This e.mail is private and confidential between Multiplay (UK) Ltd. and the person or entity to whom it is addressed. In the event of misdirection, the recipient is prohibited from using, copying, printing or otherwise disseminating it or any information contained in it. > > In the event of misdirection, illegible or incomplete transmission please telephone +44 845 868 1337 > or return the E.mail to postmas...@multiplay.co.uk. > Just using the default settings with iperf. Netblast is giving me similar results. Disabling TCP checksum offloading (transmit and receive) and large receive offload have zero effect. -- Adam Stylinski PGP Key: http://pohl.ececs.uc.edu/~adam/publickey.pub Blog: http://technicallyliving.blogspot.com
pgpyTimeMAz4y.pgp
Description: PGP signature
_______________________________________________ freebsd-net@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"