Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/5/84; site terak.UUCP
Path: utzoo!linus!philabs!cmcl2!seismo!hao!noao!terak!doug
From: doug@terak.UUCP (Doug Pardee)
Newsgroups: net.works
Subject: Re: Not again!?  Assembler vs High-Level languages
Message-ID: <492@terak.UUCP>
Date: Mon, 8-Apr-85 13:03:41 EST
Article-I.D.: terak.492
Posted: Mon Apr  8 13:03:41 1985
Date-Received: Thu, 11-Apr-85 04:25:30 EST
References: <483@terak.UUCP> <2251@wateng.UUCP>
Organization: Terak Corporation, Scottsdale, AZ, USA
Lines: 21

My original posting attacked the popular notion that there exists in
some Nirvana an optimizing compiler which produces code as good as
expertly coded assembler.  The response was on a different subject:

> You drop into assembler only when you absolutely HAVE to.
> ...
> Before you drop into assembler though I hope you have
> a) examined your algorithm, and b) profiled your code.

The Unix community apparently disapproves of such practices, since there
is no profiling program in the "toolbox" that is part of 4.2BSD.  I
dunno about System V.

> Doug spent 5 times as much time coding the algorithm so
> it was fast.  He could very easily made a small error.

Heresy!  I'll have you know that I live in the desert so that no one
notices that my shoes never touch the sidewalk when I take a walk in
the rain   :-)
-- 
Doug Pardee -- Terak Corp. -- !{hao,ihnp4,decvax}!noao!terak!doug