Any VLSI engineers pls answer...
I have a strange suspicion that computer memory degrades over time. There are millions of transistors, each with a Poisson dist. probability of failure at any given time... obviously when one transistor goes the whole circuit does not fail. But when a ptg of transistors fail say over 3 years, does the memory actually get slower?????
I format my computer about every 1-2 years, but it always seems to be slower even running the same software. Anybody know if this is correct and how you know?
2007-06-17
02:15:38
·
5 answers
·
asked by
gt5364e
3
in
Computers & Internet
➔ Other - Computers
Q, your answer gave me a good laugh. RAM is solely composed of transisors (millions of them). They are built into the IC package on a microscopic level.
I know about software engineers. They are pigs and will waste all the resources they can. (Smoother fonts, dumb animated graphics, etc)
And, yes I know about Moore's Law and the theoretical doubling of memory speed at a rate of approx. seven years. Moore's law will most likely be unattainable in a couple years as it is physically impossible to make Silicon oscillate any faster than a few GHz. The atoms may be tiny, but they do have mass and Newton still applies somewhat even in a quantum world.
Balk, your answer makes sense. I deal with bit error rate and probablilties all the time in communications. However, a DSL modem has an adaptive algorithm that will slow the device down if there are too many errors rec'd. The memory cache probably has something like that too, but what you said also would make sense.
2007-06-17
10:32:41 ·
update #1