Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site umcp-cs.UUCP
Path: utzoo!watmath!clyde!bonnie!akgua!whuxlm!harpo!decvax!linus!philabs!cmcl2!seismo!umcp-cs!chris
From: chris@umcp-cs.UUCP (Chris Torek)
Newsgroups: net.lang.c
Subject: Re: more questions about efficient C code
Message-ID: <397@umcp-cs.UUCP>
Date: Fri, 21-Jun-85 20:23:45 EDT
Article-I.D.: umcp-cs.397
Posted: Fri Jun 21 20:23:45 1985
Date-Received: Mon, 24-Jun-85 04:47:52 EDT
References: <474@crystal.UUCP>
Distribution: net
Organization: U of Maryland, Computer Science Dept., College Park, MD
Lines: 43

All of these depend on your compiler.  I'm going to answer them for
4BSD Vax PCC.

> 1) Is the "var = (bool expr) ? : expr1 : expr2" construct faster
>    than just a good old if statement? 

Depends.  PCC generates moves into r0 for expr1 and expr2, followed
by a move from r0 to var.  If var is already a register, this is
clearly slower.  If it's a complex pointer dereference, it could be
faster.

> 2) Are for loops faster than while loops? 

No.

> 3) Is there really much to be gained by using assignment within
>    a boolean expression, for example, is
>
>	if ((fp = fopen("foo", "r") == NULL) {
>
>    really that much faster than
>
>	fp = fopen("foo", "r");
>	if (fp == NULL) {

Again, depends: if "fp" is not a register, the assignment+test is faster,
since you end up branching after the assignment from r0 to "fp".  If it
is a register, it comes out the same.  (At least after optimization, usually).
The first form is never slower, anyway.

> 4) Are tests for zero more efficently written as
>
>	if (var) {
>
>    rather than as
>
>	if (var != 0) {

No.
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 4251)
UUCP:	seismo!umcp-cs!chris
CSNet:	chris@umcp-cs		ARPA:	chris@maryland