memory usage on Scyld slave nodes (fwd)
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
Rick Niles niles at scyld.comFri Feb 16 19:34:46 PST 2001
- Previous message: memory usage on Scyld slave nodes (fwd)
- Next message: memory usage on Scyld slave nodes (fwd)
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
> > I'm adding up all the memory that's in use on the entire machine, include > > buffer cache for the filesystem. Linux tries to use any spare memory up > > to about 90% of available as a cache of the filesystem. > > When you run the 'free' command, the first few numbers it shows seems to > be similar to what beostatus shows. What would probablly be a lot more > useful would be the numbers from the seconds line, with the buffer/cache > usage subtracted from to total used. I haven't yet looked to see how > expensive that computation is, but I think this number would be a lot > more useful to show as I know the kernel will generally make total usage > near 90% Yeah that's just subtracting off the buffers and cache numbers. Problem is, at least it use to be true, that buffers and cache numbers on not mutually exclusive. So the used number is smaller than reality since you're double bookkeeping some memory that's in both the buffers and cached values. I remember it was actually possible to get a negative number for "used" memory using this technique. I'm not sure with modern kernels if this is still the case. Perhaps, this value would still be better than what I have. Rick Niles. Scyld Computing Corporation
- Previous message: memory usage on Scyld slave nodes (fwd)
- Next message: memory usage on Scyld slave nodes (fwd)
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
