lock contention - what can we do about it?

[Andreas Pakulat]

On 06.03.11 22:58:18, Milian Wolff wrote:
> Alexander Shaduri, 06.03.2011:
> > Hello,
> > 
> > On Sat, 5 Mar 2011 20:17:35 +0100
> > 
> > Milian Wolff wrote:
> > > David Nolden, 05.03.2011:
> > > > If the CPU time is below 100%, then that's for sure not due to lock
> > > > contention, but simply due to I/O. Either repeat the test with the
> > > > disk-cache better filled, or increase the number of threads.
> > > 
> > > If that would be the case, wouldn't iotop show lots of disk access? And
> > > top's %wa would be high, but it's rarely more than 2% for me (wa is
> > > iowait time).
> > 
> > I may be totally wrong here, but could you not use tmpfs or something
> > like that for the files in question to rule out I/O bottleneck?
> > 
> > Cheers,
> > Alexander
> 
> good idea, but I'm not sure how this is done, esp. with the system includes... 
> can I remount /usr/include on a tmpfs?

Sure:

- create tmpfs
- mount tmpfs to /tmp/foobar/
- cp -a /usr/include/* /tmp/foobar/
- mount -o bind /tmp/foobar /usr/include

Now all io to /usr/include goes into the tmpfs.

Andreas