[Beowulf] Re: cheap PCs this christmas

Tony Travis ajt at rri.sari.ac.uk
Tue Nov 15 06:44:36 PST 2005

Andrew Piskorski wrote:
> [ajt]
>> because that slows memory access down
> No, it doesn't - not according to what I've read anyway.  (I'm not an
> expert on this.)  All else being equal, ECC does make memory a bit
> slower.  Parity does not.  Early PCs had parity, but not ECC.  The
> parity was later eliminated solely as a (probably quite small) cost
> savings measure - Parity had NO performance impact.  Driving parity
> RAM out of the market is one of those industry incidents or trends
> that supports the "most PC hardware is crap" point of view...

Hello, Andrew.

I'm not an expert either, but I've done a lot of reading about this in 
the light of the memory problems we've had on our Beowulf cluster. I'm 
sure I read this in a discussion about MCA (Machine Check Architecture) 
and speed of memory access. However, I could easily have misunderstood!

I'll see if I can find the article...

>> and reduces the total memory capacity of a chip.
> I'm not sure about that, although it does sound plausible.

Well, the chip area used to store parity bits is not available for data, 
but there were some old 486 motherboards that let you reserve part of 
the RAM for parity, or not, depending on what was important to you...

Best wishes,

Dr. A.J.Travis,                     |  mailto:ajt at rri.sari.ac.uk
Rowett Research Institute,          |    http://www.rri.sari.ac.uk/~ajt
Greenburn Road, Bucksburn,          |   phone:+44 (0)1224 712751
Aberdeen AB21 9SB, Scotland, UK.    |     fax:+44 (0)1224 716687

More information about the Beowulf mailing list