Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: Tony Viola who wrote (134455)	5/8/2001 8:01:08 PM
From: Saturn V	Respond to of 186894

Ref <So, the "software patch which would do "cache-scrubbing" sounds like it was made up by someone. I remember the mirrored cache fix that Ravi's article referenced. There used to be a thing called Mainstore Patrole in some large machines that went around in the background and corrected soft errors, but I think it used the ECC hardware itself. > I agree that that without error detection and correction circuitry, soft errors cannot be corrected. However the implementation and timing of the error correction may differ from vendor to vendor. It sounds that some systems will detect and correct the 'bad data' coming from the memory,and feed the CPU the corrected data, but will not rewrite the correct data back in the memory until a later "scrub cycle". This keeps the processor thru put high, but takes the risk that if one more soft error afflicts the same word before the scrub, the error cannot be corrected. I would prefer the approach of correcting the error in the memory as soon as the error is detected, and taking the tiny degradation in system response.