Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!princeton!caip!ut-sally!seismo!brl-adm!brl-smoke!smoke!gwyn@brl.arpa
From: gwyn@brl.arpa (VLD/VMB)
Newsgroups: net.lang.c
Subject: Re:  non-built-in exponentiation (Re: What should be added to C)
Message-ID: <1595@brl-smoke.ARPA>
Date: Mon, 23-Jun-86 15:56:12 EDT
Article-I.D.: brl-smok.1595
Posted: Mon Jun 23 15:56:12 1986
Date-Received: Sat, 28-Jun-86 07:43:35 EDT
Sender: news@brl-smoke.ARPA
Lines: 13

I wrote
	exponent % 2
because that was what I meant.  If your compiler cannot optimize
	expr % 2 == 0
to use a low-order bit test (two's complement only!), then perhaps
you should get somebody to improve your compiler.  (The SVR2 VAX
PCC generates remarkably ugly code for this, by the way.)

I'm a software engineer, not a speed phreak; I prefer code to be
correct regardless of any particular machine architecture.

The algorithm I posted is so much more efficient than the "obvious"
method that concerns for microefficiency are relatively unimportant.