Port-vax archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: NetBSD/vax performance (Was: /etc/disktab on vax)



On 2012-07-03 18:22, Anders Magnusson wrote:
On 07/03/2012 05:50 PM, David Brownlee wrote:
Something that measure the time for system calls? I think I've seen such
things in the past, but I don't remember offhand right now.
lmbench would be the classic for that - sample at
http://www.bitmover.com/lmbench/lmbench-summary, and provides good
microbenchmarks, but I think it would be helpful to have some higher
level metrics as well, as well as raw network throughput.
Many years ago I found that hardclock() had begun to take very much
time, and this is something that affects the overall performance but do
not reveal itself easily.  I assume that the same thing can be said
about softclock().

Good point! I also have eyeballed the various lock primitives, atomic ops and so on...

        Johnny


Home | Main Index | Thread Index | Old Index