Date: Thu, 18 Mar 2004 20:10:20 -0500 From: Lanny Baron <lnb@FreeBSDsystems.COM> To: Olaf Hoyer <ohoyer@ohoyer.de> Cc: "freebsd-hardware@freebsd.org" <freebsd-hardware@freebsd.org> Subject: Re: Multiprocessor system VS one processor system Message-ID: <1079658619.33813.33.camel@panda> In-Reply-To: <20040318235000.B44078@gaff.hhhr.ision.net> References: <20040316071708.C765C43D39@mx1.FreeBSD.org> <20040318235000.B44078@gaff.hhhr.ision.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Olaf, Our Server Boards (mainboards are those $19.99 boards you can buy at the stores that sell the high-end 299.99 servers) are IA32 and our Server Boards (except one tyan that uses the opteron) are Intel. Sun<tm> uses different architecture and have significantly higher costs. Lanny On Thu, 2004-03-18 at 17:59, Olaf Hoyer wrote: > On Tue, 16 Mar 2004, Simon wrote: > > > > > Sounds like a cool feature. What detects/monitors the CPUs to spot any problems > > and mark them offline at the next reboot? is this a feature of FreeBSD or motherboards > > you use? I have never heard of anything like this on Intel based servers, before. > > > Hi! > > No, its a feature of a real "Server"-Mainboard. > > Standard feature in real iron, meaning non-i386 based. > > I hava @work lots of suns sitting around, that also get beaten and > (ab)used quite a lot, and then, we also have some share of defective > CPU's a year. But... the things an old 450 with 400 MHz CPUs and 20 > HDD's stacked will do > I/O-Wise, is not easily achievable with some modern dual-Xeon Server. > > in Intel world, I rarely experienced flaky CPU, but these i386 boxes are > rarely loaded to the load a Solaris/Sparc box can take, in most cases > the Intel box simply runs out of I/O-possibilities. > > I have some boxes here, that never go below 1000 procs simultaneously. > So a load of 10 or so is normal there. (Ok, they have more than 4 CPU, > though) > > > > PS: then again, I never had a CPU fail after it passed DOA, maybe I haven't gone > > through enough CPUs, yet. > Well, in i386 world, CPUs are mostly only going south when cooling is > bad, or you get some current spikes or so, with real iron, that is being > really beaten up, you have this more often. > But real iron also remains in production use for more that 3 years... > > > > Just my 0.02 Euro on this > Olaf -- =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+= Lanny Baron Proud to be 100% FreeBSD http://www.FreeBSDsystems.COM =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1079658619.33813.33.camel>