Subject: Re: FreeBSD Bus DMA
To: None <>
From: Ted Lemon <>
List: tech-kern
Date: 06/11/1998 19:20:48
> > John, I've never seen any honest benchmark numbers for FreeBSD/i386
> > vs. NetBSD/i386, much less for FreeBSD/Alpha vs. NetBSD/Alpha.   I
> > doubt you have either, or they'd be up on your web site and you'd be
> > pointing us to the URL.
> >
> I don't post benchmarks, sorry :-).

So you're willing to make claims about FreeBSD's performance, but
you're not willing to back them up with numbers?   Am I the only
person to whom this seems more than a little bit disingenuous?

> That isn't what I have seen, and that has nothing to do with
> benchmarks but compiler loads, and real user productivity in that
> environment.

Oh, so anecdotally you've concluded, based on the "feel" of the system
(which you yourself don't use in normal work, I presume), that NetBSD
is slower than FreeBSD.

> One of our teams decided that BSD really performed badly, based upon
> their NetBSD experience.

Why are you speaking for them?   Do you know on what they based this

> We did experience a 3X speedup, and paging performance had little to
> do with it.  I didn't run those tests, and they were done by a
> skeptic :-).

Ah.   An unnamed skeptic, in whose impartiality we are supposed to
implicitely believe.

If my ASCII appears to you to be dripping with sarcasm, it's not an
accident.   I don't mind at all if you say here "I prefer FreeBSD,"
but I mind very much when you make wild claims about 3X performance
improvements between FreeBSD and NetBSD and don't back them up with
facts.   I mean, if you're right, this is something that NetBSD needs
to do something about.   A 3X performance difference in an operating
system implementation is a Big Deal.   We know about and have fixed
the paging problem.   If the problem you're describing actually
exists, we need to know about it too.

> Try it yourself, I actually  haven't seen much of a real performance
> improvement in NetBSD, but YMMV, I guess.  Anything is better than
> the MACH stuff, and we fixed acknowleged and fixed the 1st order
> problems a long time ago.  Frankly, the end conditions aren't a
> major reason for FreeBSD's peformance.

I haven't noticed a performance *problem* in NetBSD - the VM fix is an
edge case - if your system is paging during normal use, you should add
memory, which is shamelessly cheap right now.  In any case, now that
you've brought this up, I fully intend to run some comparitive
benchmarks.   I have an easy-to-swap hard drive in my laptop, so I
think I can do a pretty credible job.   Who knows, maybe I'll be so
impressed that I'll switch.