While playing with 2.2.0-preX kernels I recently noticed that my dual
PPro 180Mhz Tyan Titan Pro Red Hat Linux 5.2 system has CPU0 with a
stepping of 7 and CPU1 with a stepping of 9. It seems to run SMP
kernels just fine.
However, I also have a dual PPro 200Mhz SuperMicro P6DNF Red Hat Linux
5.2 system with CPU0 with a stepping of 7 and CPU1 with a stepping of 9
and it always crashes within 24 hours with any SMP kernel (2.1.1XX and
2.2.0-preX with MTRR). I've tried different 3c905 NICs, different EDO
RAM, updated the SuperMicro BIOS, and anything else listed in the
Linux-SMP FAQ to no avail.
I have an NT BackOffice 4.0 (ack...I know, I know) system which is a
dual PPro 200Mhz SuperMicro P6DNF with both CPUs having a stepping of 9
and it almost never crashes (well, OK, NT4SP4 does its usual crap from
time to time, but it runs very loaded for weeks at a time without
errors). This system uses the same RAM, motherboard, 3c905 NIC, and
DPT3334 EATA/DMA SCSI as the aforementioned crashing Linux box. The
only real hardware difference is the CPU stepping.
I have a customer with an NT BackOffice 4.0 system which is dual PPro
200Mhz, 3c905 NIC, DPT3334 EATA/DMA SCSI and CPU steppings of 7 and 9
but with a Tyan Titan Pro motherboard. It has not crashed or otherwise
been rebooted for months.
When, if ever, is CPU stepping an issue? Could it be that mixing
steppings on certain motherboards is a bad idea?
Please don't ignore me just because I admitted that I use NT. :-O
Thanks!
--BOFH
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/