It seems to me that if there were some back scatter communication between Thunderbird mail and Google, it could use the data of ((SPAM || JUNK ) && DELETED) to help them to remove spam and junk mail from the mailbox. The combination of all people would give mail origins a rank, which Google is great at implementing. I assume they do a "cache call", which is different from a "cash call", that can be selected to use cache in (n) cycles, which would determine the energy usage per element analyzed , per unit time. It is a way of controlling the heat generated by cache, which can burn a hole through the floor if you let it run wild. It is such an effective method that every CPU should have a cache area that could be used by a program to make simultaneous comparisons, hashes, unsorted lists, and many other things faster by many powers. It has to have some kind of entropy limit so that people don't do what they like to do, which is jump directly to the most powerful option and slam it. That would make so much heat that it would destroy itself if it were not temperature limited, or a better way is to have the software be aware of how many cycles they are pushing out a comparison from a large list to conserve energy. If the software is aware of the latency, it is easier to keep the user informed in a normal manner.
Any new CPU should have this in it and parallel CPUs is nice, but it is very difficult to untangle the sequence of logic trees without using a technique such as this. I don't think that multiple CPUs has a wonderful future, I would rather see much more user accessible cache.
0 comments:
Post a Comment