NetNews Usenet Archive 1993 #1/NN_1993_1.iso/spool/comp/lang/c/19318

Internet Message Format | 1993-01-06 | 1.6 KB

Path: sparky!uunet!psinntp!heimdall!thor!scjones
From: [e-mail] (Larry Jones)
Newsgroups: comp.lang.c
Subject: Re: Problem with string processing.
Message-ID: <[e-mail]>
Date: 6 Jan 93 17:43:26 GMT
References: <[e-mail]> <[e-mail]> <[e-mail]>
Sender: [e-mail]
Distribution: usa
Lines: 23

In article <[e-mail]>, [e-mail] (Richard M. Hartman) writes:
> If there were a decent way of getting the string length w/o
> counting all the bytes then direct support for strcpy() could
> be done easily.  Are there any CPUs that have a single-instruction
> way of saying "find the next 0 byte & tell me how far it was from here"?

Of course -- everyone's favorite example of a CISC machine, the good
old DEC Vax has the Locate Character (LOCC) instruction that locates
the next occurrence of any single character (as well as a whole bunch
of other character manipulation instructions); the Intel x86 family has
the Scan String (SCAS) instruction which can do the same thing; and
even the venerable IBM 360/370 family of CPUs have a Translate and Test
(TRT) instruction that locates the next occurrence of an character from
a set.

Of course, it's not clear that any of these instructions are actually
faster than the equivalent loop of primitive instructions and a good bit
of historical precedent indicates that they are probably slower.
----
Larry Jones, SDRC, 2000 Eastman Dr., Milford, OH  45150-2789  513-576-2070
[e-mail]  or  ...uunet!sdrc!larry.jones
Why can't I ever build character in a Miami condo or a casino somewhere?
-- Calvin