Date: Mon, 28 Aug 2000 02:09:02 -0700 From: Mike Smith <msmith@freebsd.org> To: Jin Guojun (DSD staff) <jin@george.lbl.gov> Cc: hardware@freebsd.org, smp@freebsd.org Subject: Re: 370DLE motherboard Message-ID: <200008280909.CAA00739@mass.osd.bsdi.com> In-Reply-To: Your message of "Thu, 24 Aug 2000 16:01:45 PDT." <200008242301.e7ON1jX13791@portnoy.lbl.gov>
next in thread | previous in thread | raw e-mail | index | archive | help
> > Just tested on Supermicro 370DLE motherboard and had problem with dual CPU. These aren't actually "problems" as such; they're fairly consistent with expected results. > The fxp0 is the primary network link, and the sk0 is a back-to-back connection > to the other 370DLE (both are identical). > > With one CPU installed, the sk0 throughput is 900 Mbps. > With two CPU installed, the sk0 throughput drops to 837 Mbps (non_SMP kernel). > With two CPU installed + SMP kernel, the sk0 throughput drops to 770 Mbps. > > Case 1 and 3 are expected, however, the case 2 is NOT expected (non_SMP kernel). Actually, it's not entirely unexpected. I'd accept that even if the second CPU is idle, there is extra latency somewhere in the chipset due to its presence. > The dmesg shows some information about timer (8254) interrupt issue. > > APIC_IO: Testing 8254 interrupt delivery > APIC_IO: Broken MP table detected: 8254 is not connected to IOAPIC #0 intpin 2 > APIC_IO: routing 8254 via 8259 and IOAPIC #0 intpin 0 > Waiting 7 seconds for SCSI devices to settle > SMP: AP CPU #1 Launched! > > Both machines have the same message. I wonder if the 4.1 does not recognize > the board well, or there is something wrong with "ServerWork LE" chipset. No, this is a harmless bug. It's common for systems to lie about the 8254 interrupt routing, and we should probably complain less bitterly about it. -- ... every activity meets with opposition, everyone who acts has his rivals and unfortunately opponents also. But not because people want to be opponents, rather because the tasks and relationships force people to take different points of view. [Dr. Fritz Todt] To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200008280909.CAA00739>