On Sun, Mar 10, 2013 at 02:57:16PM +0100, Seth wrote: > I'd just look at the working set/vmsize for that (usually, just pressing > 'M' vs. 'P' in top)
You wouldn't know how much of it would be written back. > Do you run with swap? I haven't configured any of my machines with swap > for at least .. many years now[*]. That might be a key difference in our > setups. Yes but that's not the point, I only use swap as a backing store for unused anonymous memory. Let's simply say my guest systems don't support memory ballooning. > Whenever I want to make stuff faster, I just give it tmpfs to use. Which > is a lot of the time. I usually work with everything (source code trees, > even virtual machine images on tmpfs). Then whenever I created something > I want to keep I "commit" it (either to a repository or by copying to > persistent storage). My current build trees are too large to fit in memory. I use ccache, and tuning virtual memory helps a lot when rebuilding parts of a build tree (which is my common working set, and not easily put in a tmpfs without the rest of the tree). This is actually what made me think about patching htop. > I was a little sketchy on the caching things I mentioned. I should have > said I'm used to seeing cache/buffer distinctively. It's probably as > simple as that. That's very understandable. -- Richard Braun ------------------------------------------------------------------------------ Symantec Endpoint Protection 12 positioned as A LEADER in The Forrester Wave(TM): Endpoint Security, Q1 2013 and "remains a good choice" in the endpoint security space. For insight on selecting the right partner to tackle endpoint security challenges, access the full report. http://p.sf.net/sfu/symantec-dev2dev _______________________________________________ htop-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/htop-general
