On 04/18/2014 08:24 PM, Lars Kruczynski wrote:
It doesn't matter if the activity is stopped or not, it doesn't free the ram. I can get plasma-desktop to take 500 MB of ram even without switching activities just by opening up a host of plasmoids and after days of the system being up. That combined with kwin, kded4, and Xorg really start eating into available memory, and since "free -m" shows all my ram just about taken upon boot, I am guessing the problem I am writing about has nothing to do with how the kernel does disk caching like I earlier wrote about but is related to pixmaps and leaks. Like I said, if it were by design, devs would have closed those bugs out and said something about how KDE functions that way, but they didn't. I can't be the only one here who's having this issue besides me and a friend who has the same problem...
You seem to have a problem with the difference between what (virtual) memory is allocated to a process and what it is actually using. As I mentioned in another post, code pages that aren't being used are LRU'd out. The way the memory-file allocation works if a code page is needed and isn't resident it gets paged in -- "on demand". That is what virtual memory is about. Pages that aren't being used can be re-used. As has already been said, Linux considers unused memory to be wasted. So of course 'free' _should_ show all your memory to be taken. If it didn't them something would be wrong! Perhaps the developers said nothing because they assumed you knew or could google for how virtual memory works. Its not as if it is new; it was developed at Manchester University back in the 1950s when memory, even 'core' memory, was fantastically expensive. -- Never tell people how to do things. Tell them what you want them to achieve, and they will surprise you with their ingenuity. --George Patton -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org