Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!hao!oddjob!mimsy!chris
From: chris@mimsy.UUCP (Chris Torek)
Newsgroups: comp.lang.c
Subject: Re: Printing binary
Message-ID: <9110@mimsy.UUCP>
Date: Mon, 26-Oct-87 19:34:25 EST
Article-I.D.: mimsy.9110
Posted: Mon Oct 26 19:34:25 1987
Date-Received: Thu, 29-Oct-87 23:08:03 EST
References: <35@ateng.UUCP> <3194@sol.ARPA> <2783@xanth.UUCP> <235@snark.UUCP> <1668@spar.SPAR.SLB.COM>
Organization: U of Maryland, Dept. of Computer Science, Coll. Pk., MD 20742
Lines: 21

In article <1668@spar.SPAR.SLB.COM> hunt@spar.SPAR.SLB.COM (Neil Hunt)
writes, in response to a question about efficiency:
>Check out all the stuff that has to happen to get the binary bits
>displayed on the screen, or written to a file, and see if building
>a new stack frame 31 times is a significant penalty.

On a Vax 11/780, yes.  putchar (of other than '\n') expands to 5
instructions n-1 out of n times, where n is the amount of buffering
on a stdio stream; the loop and test expand to 3 more.  A
call-and-return on a 780 takes about 12 times longer than the
`average' instruction.  Hence if you are printing a very large
number of binary numbers, the recursive routine will take more than
twice as long user-time-wise.  As another metric, the `average'
4.3BSD program uses a bit more user time than system time.

Of course, you are probably printing very few binary numbers; and
you are probably not using a machine with as heavy a call+return
penalty as a 780.
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7690)
Domain:	chris@mimsy.umd.edu	Path:	uunet!mimsy!chris