*BSD News Article 59098


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!news.mira.net.au!inquo!hookup!news.kei.com!newsfeed.internetmci.com!panix!not-for-mail
From: tls@panix.com (Thor Lancelot Simon)
Newsgroups: comp.unix.bsd.netbsd.misc,comp.unix.bsd.bsdi.misc,comp.unix.solaris,comp.unix.aix
Subject: Re: ISP hardware/software choices (performance comparison)
Date: 16 Jan 1996 03:18:56 -0500
Organization: Panix
Lines: 33
Distribution: inet
Message-ID: <4dfn1g$hgr@panix2.panix.com>
References: <4cmopu$d35@vixen.cso.uiuc.edu> <4cv8j1$59k@park.uvsc.edu> <4cvjpk$rpf@durban.vector.co.za> <4d3h2s$j3r@helena.MT.net>
Reply-To: tls@rek.tjls.com
NNTP-Posting-Host: panix2.panix.com
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.netbsd.misc:1845 comp.unix.bsd.bsdi.misc:1977 comp.unix.solaris:56652 comp.unix.aix:68207

In article <4d3h2s$j3r@helena.MT.net>,
Nate Williams <nate@sneezy.sri.com> wrote:
[deletia]
>Now, I admit that having more standard include files is good, but I have
>little use for a VM system which is a pig, and all the other niceties that
>Slowlaris bought me.
>
>Many of the 'new' features on Slowlaris could have been ported to the
>BSD kernel, but weren't for political reasons.  Now, in 5 years, Slolaris
>might get as fast as SunOS used to be, but that will be because it finally
>has some time to mature and hardware will be an order of magnitude faster.

Hell, if you look at the Usenix papers of a few years ago, it's pretty
obvious that the 'new' features on Slowlaris were ported *from* the BSD
kernel, and that developers at SUN were given pretty strict marching
orders not to *ever* let any of the 4.1.3 "prototype" versions get outside
the fence, probably because they'd make Slowlaris look, well, like shit.

I'm particularly fond of the paper on the slab memory allocator, where the
performance of the 4.4BSD, 4.1.3, and "slab" kernel allocators is compared
_in a 4.1.3 kernel_, and great advantages are claimed for the new code --
which then got saddled with running in Slowlaris 2.X and became part and
parcel of one of the most notoriously sluggish VM systems ever.

Mind you, the SMP is nice, but I'm not convinced it's any better than SVR4MP,
despite being a totally independent implementation.  If SUN had had a clue,
they would have bought Solbourne's code before they went under and managed
not to lose so badly for years and years just for the "privilege" of running
System V.
-- 
Thor Lancelot Simon	                                           tls@panix.COM
-   What is hardest to accept about the passage of time is that the people who -
- once mattered the most to us are wrapped up in parentheses.    --John Irving -