Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!cs.utexas.edu!tut.cis.ohio-state.edu!ucbvax!ucdavis!iris!schell
From: schell@iris.ucdavis.edu (Stephan Schell)
Newsgroups: comp.sys.hp
Subject: 9000/825 integer performance
Message-ID: <5097@ucdavis.ucdavis.edu>
Date: 11 Aug 89 19:25:33 GMT
Sender: uucp@ucdavis.ucdavis.edu
Reply-To: schell@iris.ucdavis.edu (Stephan Schell)
Organization: U.C. Davis - Department of Electrical Engineering and Computer Science
Lines: 11

Why do integer multiplies take longer than double precision floating point
multiplies (at least, in C) ?

I timed 180K integer multiply-accumulates and 206K double precision floating
point multiply accumulates (including integer-to-double conversion) using
the "times()" function.  This seems kind of hokey that integer multiplies run
slower than double precision floats, doesn't it?

Steve
Dept of EE & CS
Univ. Calif., Davis