Date: Fri, 26 Jan 2007 08:07:09 +0100 From: Hans Petter Selasky <hselasky@c2i.net> To: freebsd-current@freebsd.org Cc: "M. Warner Losh" <imp@bsdimp.com>, current@freebsd.org, Colin Percival <cperciva@freebsd.org> Subject: Re: Interesting speed benchmarks Message-ID: <200701260807.09055.hselasky@c2i.net> In-Reply-To: <cb5206420701252243l12e871dfhbc99886d884052b7@mail.gmail.com> References: <20070125.192448.-432840241.imp@bsdimp.com> <45B9895B.9020709@freebsd.org> <cb5206420701252243l12e871dfhbc99886d884052b7@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Friday 26 January 2007 07:43, Andrew Pantyukhin wrote: > On 1/26/07, Colin Percival <cperciva@freebsd.org> wrote: > > M. Warner Losh wrote: > > > On a lark, I just got a combo USB/Firewire external disk drive. I ran > > > some crude benchmarks, and I was surprised by what I found. This is > > > on a fairly stock -current kernel. > > > > > > Firewire does around 40MB/s, while USB 2.0 maxes out at about 12MB/s. > > > > Either amd64 is slower than i386, there has been a performance regression > > between 6.2-RELEASE and whatever version of FreeBSD you're running, or > > you have a rather slow external drive. I get 25MB/s from my Vantec > > Nexstar3 USB 2.0 enclosure: > > > > http://www.daemonology.net/blog/2006-01-28-vantex-nexstar3.html > > I've always (since 5.x I think) got 25Mb/s on i386 > and half that on amd64, no matter what equipment I > try. 1) Maybe this has something to do with the interrupt threshold. Does changing: EHCI_CMD_ITC_2 into EHCI_CMD_ITC_1 increase performance, in "ehci.c" ? 2) You could try setting the transfer multiplier, "EHCI_QH_SET_MULT()" to 3 for HIGH-speed bulk/control transfers, but not interrupt transfers. The EHCI manual states that the USB host controller can choose to ignore this field, but what if it doesn't? 3) How about the new USB stack? Any difference? --HPS
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200701260807.09055.hselasky>