On Fri, Dec 24, 2010 at 2:34 AM, Von Fugal wrote:
>
> swap not quite ram place to put the compcache. I suppose you could code
> compcache to keep a certain margin open by itself initiating evictions
> when the threshold is passed. I wonder though how it would convince the
> kernel to store the ev
> On Thu, Dec 23, 2010 at 2:52 AM, Von Fugal wrote:
> > It really should sit as an in-between swap and memory place. Recently
> > "swapped" pages should go straight to compcache, evicting pages if
> > necessary *from* compcache to disk. That would be ideal. I don't know if
> > the kernel devs wou
On Thu, Dec 23, 2010 at 2:52 AM, Von Fugal wrote:
> It really should sit as an in-between swap and memory place. Recently
> "swapped" pages should go straight to compcache, evicting pages if
> necessary *from* compcache to disk. That would be ideal. I don't know if
> the kernel devs would ever wan
> There has been research into heuristics that allow one to get the best
> of compressed ram and disk based swap device, providing performance
> that is never worse than disk swap (minus 2% overhead or so). Also
> simply compressing the pages being swapped out could be useful for
> reducing the nu
On Thu, Dec 23, 2010 at 1:26 AM, Von Fugal wrote:
>> One thing I have found quite useful is compcache. This allows you to
>> use compressed ram a swap-device. It doesn't need a physical disk and
>> has zero latency so it avoids many of the disadvantages of physical
>> swap devices. It tends to giv
> On Wed, Dec 22, 2010 at 1:35 AM, AJ ONeal wrote:
> > The bugs I've already been bitten by is that if you have no swap the kernel
> > will default to sending the oom (out-of-memory) killer on any random process
>
> One thing I have found quite useful is compcache. This allows you to
> use compr
Are there any other ideas for what I might do with sysctl, the kernel
config, or things I can change it /etc?
AJ ONeal
On Tue, Dec 21, 2010 at 10:35 AM, AJ ONeal wrote:
> The bugs I've already been bitten by is that if you have no swap the kernel
> will default to sending the oom (out-of-memor
Thank you very much!
This looks very interesting. I'll keep it in mind.
AJ ONeal
On Wed, Dec 22, 2010 at 5:28 AM, John McCabe-Dansted wrote:
> On Wed, Dec 22, 2010 at 1:35 AM, AJ ONeal wrote:
> > The bugs I've already been bitten by is that if you have no swap the
> kernel
> > will default to
On Wed, Dec 22, 2010 at 1:35 AM, AJ ONeal wrote:
> The bugs I've already been bitten by is that if you have no swap the kernel
> will default to sending the oom (out-of-memory) killer on any random process
One thing I have found quite useful is compcache. This allows you to
use compressed ram a s
AJ,
> I'm looking for a list of other settings sysctl and otherwise - that will
> prevent me from banging my head against the wall.
It's interesting to see you on the gumstix list. Are you working on a
hobby project? It would be fun to do a group project.
It is sad that gumstix doesn't have more
The bugs I've already been bitten by is that if you have no swap the kernel
will default to sending the oom (out-of-memory) killer on any random process
and kill it in a fashion that to the casual observer appears to be a
segfault (a segfault that nor gdb nor the powers of above will tell you
isn't
11 matches
Mail list logo
