[Beowulf] Re: Beowulf Digest, Vol 10, Issue 16

[David Mathog]

> On Mon, 13 Dec 2004, Mark Hahn wrote:
> 
> > > >> I believe the socket 754 and 939 Athlon 64s do not support ECC,
while
> > > >> the socket 940 Athlon 64 and Opteron do.
> > > >
> > > It's hard to tell what the 754 boards support since many say they
> > > will _accept_ ECC memory but they don't say that they can actually
> > > _use_ it.
> > 
> > all K8's support ECC - AMD definitely does not disable ECC on some
chips.
> > I just checked the s754 functional spec, and it certainly supports ECC.
> > 
> > it's conceivable that MB vendors are such twits that they'd fail to 
> > detect ECC dimms and enable ECC in the bios.  it's even possible they'd
> > be such idiots as to fail to connect memcheck pins to the dimm slots.
> 
> FWIW, I now am installing my ASUS K8NE, and it caims in its user guide
> to support unbuffered ECC or non-ECC SDRAM (up to PC3200).
> 

Back to the original thought, which was recycling the PC2100
memory from the Tyan S2466N boards.  The consensus seems to be that
registered ECC recycled from the S2466N systems will not work in
boards designed for unbuffered ECC, which means that both
the 754 and 939 are out.  The 940 boards require registered
ECC but they cost more than the other two and the CPUs that
go in them also cost more.

Very unclear to me what, if anything is gained by using 
registered ECC vs. unbuffered ECC on a smallish system.  On a big
system, with lots of memory, I can see where registered ECC would
be a plus.  But if the system only holds 1Gb of RAM in one
or two memory slots the unbuffered memory should be slightly
faster, and with the ECC enabled, just as reliable.  Correct?

Thanks,


David Mathog
mathog at caltech.edu
Manager, Sequence Analysis Facility, Biology Division, Caltech