From owner-freebsd-hackers  Wed Dec  4 15:54:52 1996
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.3/8.7.3) id PAA23064
          for hackers-outgoing; Wed, 4 Dec 1996 15:54:52 -0800 (PST)
Received: from parkplace.cet.co.jp (parkplace.cet.co.jp [202.32.64.1])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id PAA23058
          for <hackers@FreeBSD.org>; Wed, 4 Dec 1996 15:54:49 -0800 (PST)
Received: from localhost (michaelh@localhost) by parkplace.cet.co.jp (8.8.3/CET-v2.1) with SMTP id XAA22366; Wed, 4 Dec 1996 23:54:28 GMT
Date: Thu, 5 Dec 1996 08:54:28 +0900 (JST)
From: Michael Hancock <michaelh@cet.co.jp>
Reply-To: Michael Hancock <michaelh@cet.co.jp>
To: Jake Hamby <jehamby@lightside.com>
cc: hackers@FreeBSD.org
Subject: Re: Speaking of performance...
In-Reply-To: <Pine.BSF.3.95.961130032519.229B-100000@hamby1>
Message-ID: <Pine.SV4.3.95.961204162142.17926C-100000@parkplace.cet.co.jp>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-hackers@FreeBSD.org
X-Loop: FreeBSD.org
Precedence: bulk

Are you confusing streams with the STREAM benchmark?

Mike Hancock

On Sat, 30 Nov 1996, Jake Hamby wrote:

> That TCP bandwidth thread was obnoxious, but I was interested in one
> comment from Mr. Miller (and Terry Lambert's response) on STREAM
> performance, and Solaris performance tuning.  Sun has a new IPC mechanism
> called "doors" which came from their Spring OS project that they claimed
> in a recent developer newsletter is faster than any other form of IPC in
> Solaris.  Right now the door system call is officially UNdocumented (an
> excerpt from "man door" is included below for your amusement), and the
> only thing that uses it is nscd (name service cache daemon, which caches
> passwd, group, and host data), but I'd expect it to take a bigger role in
> Solaris 2.6.  Terry, or anyone, is this a technology worth investigating,
> or just some tidbit that Sun threw in because STREAMs were so slow?
> 
> -- Jake
> 
> door(2)                   System Calls                    door(2)
> 
> WARNING
>      Please do not attempt to reverse-engineer the interface  and
>      program to it. If you do, your program will almost certainly
>      fail to run on future versions of Solaris, and may  even  be
>      broken  by  a  patch.   This document does not constitute an
>      API. Doors may not exist or may have a completely  different
>      set of semantics in a future release.
> 
> NOTES
>      This manual page is here solely for the  benefit  of  anyone
>      who  noticed  door_call()  in  truss(1)  output and thought,
>      "Gee, I wonder what that does..."
> 
> SunOS 5.5.1         Last change: 29 Aug 1995                    1