recommend a stable dual proc AMD mobo

tdp05 · 09-22-2003, 06:58 PM

Hi - At my work we use mostly dual proc AMD's running redhat 7.3. We had been using the Asus A7M266D, but it's been end of lifed by our supplier, plus we're looking for boards that will be able to run at a higher bus speed and support the Bartons etc, not going to AMD64, just yet so please limit recommendations to the 32bit realm.

We received a shipment of Gigabyte 7DPXDW-P and they pretty much blow. Their biggest fault is that they will occassionally only see one cpu of the two that are on there. A reboot will fix it, almost every time.

Anyone have a good replacement board that they have personal experience with? We're open to just about any chipset, please let me know how it has performed.

the machines get taxed heavily, pretty much all day. If they aren't being used to do medium to really heavy 3d applications, then they get do distributed rendering overnight.

I included a small /var/log/messages for those that are interested in the cpu dropping.

kernel: CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64 bytes/line)
kernel: CPU: L2 Cache: 256K (64 bytes/line)
kernel: CPU0: AMD Athlon(tm) MP 2600+ stepping 01
kernel: per-CPU timeslice cutoff: 731.42 usecs.
kernel: enabled ExtINT on CPU#0
kernel: ESR value before enabling vector: 00000000
kernel: ESR value after enabling vector: 00000000
kernel: Booting processor 1/1 eip 2000
kernel: Not responding.
kernel: CPU #1 not responding - cannot use it.
kernel: Error: only one processor found.

Thanks for any suggestions

german · 09-22-2003, 11:11 PM

Asus makes kick-ass mobo's. I can't remember the model name of their dual AMD board, but I've used their dual P4 boards, and their single-processor ones, and they have always delivered.

HTH

B.

-- single processor AMD boards that is.

tdp05 · 09-23-2003, 01:08 PM

anyone ?

We had problems with the Asus A7N8X Deluxe for some of the single proc machines we use. Its based on the nforce2 chipset. The machines weren't stable at the highest bus speed, we had to clock them down to 133 and they worked better, but were not completely stable. This wasn't just 1 machine, it was on a group of 15 or so, all exhibiting this behavior.