Date: Thu, 1 Jun 2006 08:22:41 -0700 (PDT) From: Nash Nipples <trashy_bumper@yahoo.com> To: Mike Jakubik <mikej@rogers.com> Cc: freebsd-performance@freebsd.org Subject: Re: Poor Samba throughput on 6.0 Message-ID: <20060601152241.45831.qmail@web36304.mail.mud.yahoo.com> In-Reply-To: <1083.172.16.0.199.1148950758.squirrel@172.16.0.1>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi Guys! I just cant sleep till i make this thing clear. We have a 10/100 Mb/s NIC which transmits 33 000 000 Hz x 32 Bytes width = 132 MB/s over PCI 2.2 But how do you guys count 12.5 MB/s in the cable when the NIC has lets say realtek 8139 25 MHz external clock. and 4 cables to transmit bi-directional data that is only TX+ TX- RX+ RX- which is only 2 bits long at a time so 25 MHz x 2 = 6.25 MB/s for 4 wires and 25 MHz x 4 = 12.5 MB/s for 8 wires. and how do you get 25 MB/s on a 1G link when the clock is still 25 MHz? Thank you, Nash Mike Jakubik <mikej@rogers.com> wrote: On Mon, May 29, 2006 8:10 pm, Joao Barros wrote: > On 5/27/06, Mike Jakubik wrote: >> I am using -CURRENT here, disabling net.inet.tcp.inflight improves the download rate by 2MB/s! >> > How old is that CURRENT? I believe that shouldn't happen after Andre's commit back in March. > I can't see any difference toggling inflight on or off. FreeBSD 7.0-CURRENT #0: Tue May 16 13:46:05 EDT 2006. _______________________________________________ freebsd-performance@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-performance To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org" --------------------------------- Yahoo! Messenger with Voice. PC-to-Phone calls for ridiculously low rates.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060601152241.45831.qmail>