5 ms·
The author’s argument is that a highly irradiated landscape would be too energetic for silicon based memory to work as it’s now designed. I’m not an EE but I’ve
by Dangeranger 8y ago
The author’s argument is that a highly irradiated landscape would be too energetic for silicon based memory to work as it’s now designed. I’m not an EE but I’ve spoken with some engineers of early silicon systems which had to account for high energy cosmic rays flipping the bits within RAM in a random fashion. My guess is that using a solution like Hamming codes would not work when retrieving arbitrary data at a large scale.
- FPGAhacker 8y agoEcc does work for radiation hardening of storage. You can still get logic upsets in transit, between storage elements. You can either design your cells around it (high power ) or you can use a variety of redundant logic and monitors.
- amaterasu 8y agoOff topic, but with smaller lithography, ANY subatomic particle of sufficient energy becomes a problem: https://www.xilinx.com/support/documentation/white_papers/wp208.pdf https://www.xilinx.com/support/documentation/white_papers/wp... I've spent months chasing random crashes on another vendor's device which turned out to be solder ball alpha emission issues.