Joe Cooper wrote:
> > It would also be nice if you could do a benchmark of hno-asyncio from
> > sourceforge. The hno-asyncio branch includes a different I/O sheduler
> > which should make a lot higher utilization of the threads while
> > hopefully wasting less CPU time.
>
> Per your other mail, I assume you'd like me to wait a day or two on this
> one? If not, I'll fetch it tonight and run it on a second machine.
No, please go ahead if you can and if it does not hinder your benches of
2.4-HEAD or other important tasks. It's only that I am not too
optimistic that the changes actually improves things.
It is not the end of the road however. There is lots of opportunity at
making a more efficient async-ufs implementation. One important
optimization from 2.2-HNO not yet in 2.4.DEVEL is to split unlink to a
separate queue from the normal storage uperations (2.2-HNO uses unlinkd,
but it should be implemented using separate threads).
/Henrik
Received on Mon Jun 12 2000 - 15:08:49 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:12:29 MST