[linux-mm-cc] First experience of compressed cache

Nai Xia nai.xia at gmail.com
Tue Apr 1 08:03:30 EDT 2008


On Tue, Apr 1, 2008 at 4:53 PM, Nitin Gupta <nitingupta910 at gmail.com> wrote:
> On Mon, Mar 31, 2008 at 3:37 PM,  <nai.xia at gmail.com> wrote:
>  > Hi, I found this project both from kernelnewbies.org and linux.org. I think it's
>  >  interesting and so I checked out the code HEAD and tried today.
>
>  I often leave SVN in uncompilable state or some half-cooked/untested
>  code which can cause strange bugs.
>  So, I would suggest you to use released version.
>
>
>  >  My first experience is based on a very simple RAM memory overloading:
>  >
>  >  1. Concurrently booted up two virutal machines which ate up about 75% of my total
>  >  virutal memory(RAM + hard-disk swap) without compcache.
>  >  2. Booted them with default loaded compcache (that means 25% of RAM was used as ramswap).
>  >
>  >  I saw a significant performance loss in the 2nd case. ( Actuall I killed one of
>  >  the virutal machines before it could eat all my hard-disk swap).
>  >
>  >  Does problem lie in the current "static cache" implementation?
>  >  Or was I just using it at the wrong place or with wrong module parameters?
>  >
>
>  I seems you are running compcache on host only and not inside VMs.
>  Try running compcache inside VMs (assuming these are Linux too). This
>  way you can run VMs with less
>  than usual RAM assigned. I am not very sure if running compcache on
>  host in addition to in VMs will have
>  any negative performance impact.

Ok, I'll try that in VM.

>
>
>
>  >  I get from the paper "Adaptive Compressed Caching: Design and Implementation"
>  >  that this method should suit to both hard swap backed cache and file cache,
>  >  while it seems that the features are still not in your code.
>  >
>  >  will all that features planned on the roadmap of the version based on kernel
>  >  2.6? (sorry, I did not find roadmap info in the project homepage)
>  >
>
>  Compressing page-cache(filesystem backed) pages, is highly desirable
>  but I don't yet have
>  a way of cleanly implementing this without intrusive kernel patching.
>  If you want an implementation for pagecache+swapcache compression, you
>  can refer patches
>  for 2.6.x at:
>
>  http://linuxcompressed.sourceforge.net/

   Thanks for the information. I'll try that patch too.

>
>
>
>  >  I am very much interested in this project, however I am still a kernel newbie.
>  >  I will be very glad to help, if someone could kindly be my mentor. :-)
>  >  (And I am currently reading related papers and code by myself.)
>  >
>  >
>  >
>
>  I would be glad to help you out :)

       Thanks for your kindness :)
>
>
>  Cheers,
>  Nitin
>


More information about the linux-mm-cc mailing list