Re: Hurd Latencies -- Questions...
On Tue, Jun 08, 1999 at 03:19:55PM -0700, Brent Fulgham wrote:
> Anyone using the hurd for non-trivial tasks (such as compiling
> large packages, etc.) can attest to the fact that there is a
> noticable difference in speed between the hurd and Linux
> running on the same hardware.
Indeed :)
> I was wondering if any profiling had been done to determine
> where the time is being spent. Is it due to the overhead
> of RPC? Is it because of poor swap memory allocation?
> Is it something else entirely?
AFAIK, there has not been done any profiling. "Making sure that profiling
works" is still an item on the task list.
Furthermore, a lot of the speed penalty is due to disk access, even
sequential access requires one RPC per page or so. It would help a big deal
(so is assumed) if there would be another RPC that could collect ranges of
pages and send them as a packet. If this description is inaccurate that's
because I only have a naive understanding of how things work.
You can see this when you "dd if=/dev/hdc of=/dev/null" for example. Watch
the blinking of the disk led or the clicking noise.
Well, there is of course more, but that's what I assume.
Thanks,
Marcus
--
`Rhubarb is no Egyptian god.' Debian http://www.debian.org finger brinkmd@
Marcus Brinkmann GNU http://www.gnu.org master.debian.org
Marcus.Brinkmann@ruhr-uni-bochum.de for public PGP Key
http://homepage.ruhr-uni-bochum.de/Marcus.Brinkmann/ PGP Key ID 36E7CD09
Reply to: