DRAM Errors in the Wild: A Large-Scale Field Study

(October 21, 2009) Bianca Schroeder of the University of Toronto Computer Science Department gives an in depth discussion on how common dynamic random access memory errors are, their statistical properties, and how they are affected by external and chip-specific factors.

Stanford University:
http://www.stanford.edu/

Stanford Engineering Everywhere:
http://see.stanford.edu/

Stanford Center for Professional Development:
http://scpd.stanford.edu/

Stanford School of Engineering:
http://soe.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford/

Category:

Education

Tags:

• science
• technology
• engineering
• computer
• dynamic random access memory
• DRAM
• failure
• error
• data
• disk drives
• hard
• soft
• chipset
• CPU
• google
• correctable
• temperature
• DIMM

[grunder20]

Computer can be held responsible for such failures.

[bomidibi]

After listening to the You Tube presentation on DRAM Errors, I noticed that there was no reference of soft error thresholds of the various hardware platform being tested. There was reference of using events logs to accumulate your soft failures. This threshold I speak of is used by many hardware platform to prevent the posting of soft ECC Single-Bit until the threshold level is exceeded. Depending on threshold set by each hardware platform the data collected could be extremely skewed. .

[GentleSavage1]

The high frequency noise / compression artifacts in the audio here are quite unpleasant to the ears (when listened through halfway decent speakers).

[dormantreign]

and so the saying "back in the good old days" means something now. I guess that is why my old guitar amps are still kicking around and working!