Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!news.ecn.uoknor.edu!news.ysu.edu!usenet.ins.cwru.edu!magnus.acs.ohio-state.edu!freenet.columbus.oh.us!pacific.mps.ohio-state.edu!math.ohio-state.edu!uwm.edu!spool.mu.edu!howland.reston.ans.net!swrinde!sgigate.sgi.com!cygnus.com!usenet
From: jtc@hippo.cygnus.com (J.T. Conklin)
Newsgroups: comp.os.linux.networking,comp.unix.bsd.netbsd.misc,comp.unix.bsd.freebsd.misc
Subject: Re: TCP latency
Date: 17 Jul 1996 13:34:37 -0700
Organization: Cygnus Support
Lines: 24
Message-ID: <tjohlemxeq.fsf@hippo.cygnus.com>
References: <4paedl$4bm@engnews2.Eng.Sun.COM>
<31E7C0DD.41C67EA6@dyson.iquest.net> <4s8tcn$jsh@fido.asd.sgi.com>
<31E80ACA.167EB0E7@dyson.iquest.net> <4sadde$qsv@linux.cs.Helsinki.FI>
<31E9E3A7.41C67EA6@dyson.iquest.net> <4sefde$f0l@fido.asd.sgi.com>
NNTP-Posting-Host: hippo.cygnus.com
In-reply-to: lm@neteng.engr.sgi.com's message of 15 Jul 1996 22:06:38 GMT
X-Newsreader: Gnus v5.1
Xref: euryale.cc.adfa.oz.au comp.os.linux.networking:45574 comp.unix.bsd.netbsd.misc:4099 comp.unix.bsd.freebsd.misc:23854
>>>>> "lm" == Larry McVoy <lm@neteng.engr.sgi.com> writes:
lm> Umm, I'd be happy to entertain suggestions for a better
lm> measurement of a null entry into the system. I don't want
lm> something that anyone special cases - that's just worthless. I
lm> want something that is actually measuring all the work you need to
lm> do to get to the point that you can do something in the kernel.
lm> Maybe gettimeofday() would have been a better null syscall, what
lm> do you think?
Maybe not. In my development kernel, gettimeofday is a user-land
wrapper around clock_gettime(). Thus it has additional function call
overhead plus the (relatively expensive) divisions needed to convert
timespecs to timevals.
If the goal is to measure getting in and out of a kernel, both *BSD
and Linux have the advantage that exactly that can be measured by
adding a true null syscall --- perhaps as a LKM.
With those numbers as a baseline, you can then measure the additional
overheads related to timers, file descriptors/systems, process tables,
etc. if those numbers are interesting.
--jtc