Multi-core doesn't help on CPU-intensive applications much if at all, unless the process is amenable to parallelization (some things are, like media encoding (that the process is in principle parallelizable, of course, doesn't necessarily imply that the software implementing that process is parallelizable (though given a sufficiently smart compiler...)), but there's a fair number of CPU-intensive applications that are so dependent on intermediate state that they're not that parallelizable (in those cases, the popular interpretation of Moore's Law basically stopped applying about a decade ago). 64-bit doesn't help that many CPU-bound applications either. Where multicore does deliver a benefit is when there are multiple threads (be it in the same application, multitasking, or system services), and 64-bit's much more about the increased address space (though 1 "bit" of that is effectively given back by some platforms by making most things take twice as much space in memory...), especially per-process (hacks like PAE allow more than 4GB to be used, but no more than 4GB per process). On Thu, May 15, 2014 at 11:43 AM, Robert Heller <heller at deepsoft.com> wrote: > Right. At this point transistor size (which determines base processor > cycle > time) is about as small as can be. For most things done on personal > computer, > having multiple cores or even 64-bit processors actually does little to > increase perceived speed. Multiple cores and/or 64-bit processors are much > more useful in servers and/or computationally intensive processing. The > only > 'home' users that would see much improvement with multiple cores or 64-bit > processors would be gamers. > -- Levi Ramsey leviable at gmail.com lramsey at umass.edu -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.hidden-tech.net/pipermail/hidden-discuss/attachments/20140517/dafd9a18/attachment.html