Date: Wed, 22 Sep 2010 21:50:59 -0700 From: Jeremy Chadwick <freebsd@jdc.parodius.com> To: Bryce <bryce@bryce.net> Cc: freebsd-stable@freebsd.org Subject: Re: SuperMicro i7 (UP) - very slow performance Message-ID: <20100923045059.GA29396@icarus.home.lan> In-Reply-To: <da189ecd-c9ba-4b49-9f46-6ee4e3dfea56@g18g2000vbn.googlegroups.com> References: <AANLkTimRpatbjTcpdBT6EV1Jb2U3ake4me7YPxHhwO7K@mail.gmail.com> <mailpost.1285145047.2294223.42059.mailing.freebsd.stable@FreeBSD.cs.nctu.edu.tw> <da189ecd-c9ba-4b49-9f46-6ee4e3dfea56@g18g2000vbn.googlegroups.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Sep 22, 2010 at 05:54:46PM -0700, Bryce wrote: > On Sep 22, 3:43 am, free...@jdc.parodius.com (Jeremy Chadwick) wrote: > > On Wed, Sep 22, 2010 at 03:30:33AM -0500, Adam Vande More wrote: > > > On Wed, Sep 22, 2010 at 3:12 AM, Peter Jeremy <peterjer...@acm.org> wrote: > > > > > > I think something is badly wrong here. That's less than 1/2 the speed > > > > of my Athlon 4850e (2.5GHz) and only 60% more than my Atom N270. None > > > > of the other figures you posted look anomolous. Are you sure the CPU > > > > is actually running at full speed and you haven't done something like > > > > disable the caches in BIOS? > > > > > FWIW: > > > > > FreeBSD galacticdominator.com 8.1-PRERELEASE FreeBSD 8.1-PRERELEASE #1: Sun > > > Jun 20 21:05:37 CDT 2010 > > > a...@galacticdominator.com:/usr/obj/usr/src/sys/GENERIC > > > amd64 > > > CPU: Intel(R) Core(TM) i7 CPU 870 @ 2.93GHz (2940.64-MHz K8-class > > > CPU) > > > > > MD5 time trial. Digesting 100000 10000-byte blocks ... done > > > Digest = 766a2bb5d24bddae466c572bcabca3ee > > > Time = 2.012719 seconds > > > Speed = 496840352.000000 bytes/second > > > > > vmstat > > > -i > > > > > interrupt total > > > rate > > > > > irq16: vgapci0+ 10720642 54 > > > irq18: fwohci0 2 0 > > > irq23: ehci1 623712 3 > > > cpu0: timer 393496151 1996 > > > irq256: hdac0 8063581 40 > > > irq257: re0 4136265 20 > > > irq259: ahci1 1925783 9 > > > cpu1: timer 393494902 1996 > > > cpu6: timer 393494606 1996 > > > cpu5: timer 393494653 1996 > > > cpu7: timer 393494701 1996 > > > cpu4: timer 393494785 1996 > > > cpu3: timer 393494732 1996 > > > cpu2: timer 393494404 1996 > > > Total 3173428919 16102 > > > > > His interrupts seem high compared to this setup, but I don't what expected > > > values should be. > > > > How are his interrupt rates "higher" than yours? If you're focused on > > the cpuX entries, don't be. > > > > To the OP: > > > > 1) I don't see how/why USB Legacy support would have anything to do with > > your problem (meaning: you stated that things "improved a little" if you > > disabled USB Legacy support in the BIOS, which makes no sense given what > > that option does). > > The machine runs *MUCH* slower if Legacy BIOS is not disabled. Just > booting up is excruciating, and takes forever. I'm not sure what "Legacy BIOS" is (you mention it twice in your mail), but I think it's just a typo for "Legacy USB" or "USB Legacy". Regarding "the machine runs much slower", can you provide something as simple as "md5 -t" both with the option enabled and with the option disabled? System boot time increasing due to this option wouldn't surprise me; I've battled Supermicro in the past over USB-related issues during or after BIOS POST (though they were specific to booting from USB flash drives): http://koitsu.wordpress.com/2008/04/05/supermicro-pdsmi-bios-bugs/ http://koitsu.wordpress.com/2008/07/10/supermicro-pdsmi-bios-bugs-finale/ I've run into other BIOS bugs as well: http://koitsu.wordpress.com/2009/01/06/supermicro-x7sba-and-ecc-ram-part-1/ http://koitsu.wordpress.com/2009/01/22/supermicro-x7sba-and-ecc-ram-finale/ I hate defaulting to this kind of response, but I assume you're running the latest available BIOS for your mainboard? > > 2) There's been a discussion on -stable about FreeBSD incorrectly > > determining different kinds of CPU characteristics on newer processors > > like the i7, with HTT in use. I can dig up the thread if you'd like. > > It does include a patch. > > Yes, I'd like that. In case this helps: > > kern.sched.topology_spec: <groups> > <group level="1" cache-level="0"> > <cpu count="8" mask="0xff">0, 1, 2, 3, 4, 5, 6, 7</cpu> > <children> > <group level="3" cache-level="2"> > <cpu count="8" mask="0xff">0, 1, 2, 3, 4, 5, 6, 7</cpu> > <children> > <group level="5" cache-level="1"> > <cpu count="2" mask="0x3">0, 1</cpu> > <flags><flag name="THREAD">THREAD group</flag><flag > name="SMT">SMT group</flag></flags> > </group> > <group level="5" cache-level="1"> > <cpu count="2" mask="0xc">2, 3</cpu> > <flags><flag name="THREAD">THREAD group</flag><flag > name="SMT">SMT group</flag></flags> > </group> > <group level="5" cache-level="1"> > <cpu count="2" mask="0x30">4, 5</cpu> > <flags><flag name="THREAD">THREAD group</flag><flag > name="SMT">SMT group</flag></flags> > </group> > <group level="5" cache-level="1"> > <cpu count="2" mask="0xc0">6, 7</cpu> > <flags><flag name="THREAD">THREAD group</flag><flag > name="SMT">SMT group</flag></flags> > </group> > </children> > </group> > </children> > </group> > </groups> Worth reading -- apparently you're not the only one. Not sure if your boards or setups are identical though: http://lists.freebsd.org/pipermail/freebsd-stable/2010-May/056591.html http://lists.freebsd.org/pipermail/freebsd-stable/2010-May/056607.html http://lists.freebsd.org/pipermail/freebsd-stable/2010-May/thread.html#56591 As for the Intel processor topology patch: http://lists.freebsd.org/pipermail/freebsd-stable/2010-September/058668.html http://lists.freebsd.org/pipermail/freebsd-stable/2010-September/thread.html#58668 > > 3) Reset the BIOS settings to Factory Defaults ("Load Setup Defaults" > > or the like), and then write down whatever you change, then post the > > changes here. > > After resetting BIOS, I make the following changes: > > - Disable Legacy BIOS > - Change drive controller from enhanced mode to AHCI > > and for your #4, the temp is almost always between 49-60 degrees C > and the cpu freq is always 2801. > > I am not running powerd as I saw odd behavior and wanted to get a know > good before introducing other variables. I think your core temperatures are fine -- they look about right for Core i7 CPUs (a few Celsius on the high side, but that's going to vary from environment to environment). -- | Jeremy Chadwick jdc@parodius.com | | Parodius Networking http://www.parodius.com/ | | UNIX Systems Administrator Mountain View, CA, USA | | Making life hard for others since 1977. PGP: 4BD6C0CB |
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20100923045059.GA29396>