Subject: Re: Not so fun discovery...
To: Anders Magnusson <ragge@ludd.luth.se>
From: Luke Mewburn <lukem@cs.rmit.edu.au>
List: tech-net
Date: 06/11/2000 13:33:09
Anders Magnusson writes:
> I have recently fired up an old VAX 11/750. It had NetBSD 1.2
> on its disks, so and therefore I did some ethernet benchmarks while here.
> I also installed NetBSD 1.4ZA on another disk and checked the same.
> 
> Py putting/getting /netbsd from a machine on the local network I 
> got the following numbers:

Is this just using ftp(1), or other tests as well?

I discovered that even on a fast machine that my rate limiting
code had an observable impact on non-rate limited connections,
which is why I rolled-out the code for the latter case, but maybe
I missed something.

OTOH, it could be the IP cksum stuff mentioned elsewhere on this
thread.

(I'm always a bit concerned when people's benchmarks using ftp are
different, because I don't know if it's my changes or the kernel's :)