Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!jwt!john From: john@jwt.UUCP (John Temples) Newsgroups: comp.sys.ibm.pc.hardware Subject: Re: Difference between 386/33 & 486/25 not counting fp Message-ID: <1991Apr17.021437.10968@jwt.UUCP> Date: 17 Apr 91 02:14:37 GMT References: <1991Apr12.093457.4147@kessner.denver.co.us> <1991Apr13.154941.1204@jwt.UUCP> <1991Apr14.215120.12728@kessner.denver.co.us> Organization: Private System -- Orlando, FL Lines: 29 In article <1991Apr14.215120.12728@kessner.denver.co.us> david@kessner.denver.co.us (David Kessner) writes: >In article <1991Apr13.154941.1204@jwt.UUCP> john@jwt.UUCP (John Temples) writes: >>Who's to say that the 486 (either by design or by chance) doesn't >>run the instruction mix that represents the Dhrystone more >>efficiently than it might run some other instruction mix? >Where they could, the same binaries were used. Thus, the effects of a >'better instruction mix' is negated because both machines benifit from >them. I didn't say "better instruction mix." I realize that the binaries were the same here, but we're comparing two different CPUs which happen to be binary compatible. The point I was trying to make is that the benchmark in question may have included an instruction mix which the 486 could execute more efficiently than might be expected in a more "realistic" application mix -- the Norton SI syndrome. Also, given the importance of the Dhrystone in CPU marketing hype, would you really be surprised to find out a CPU had Dhrystone optimizations in hardware? (No, I'm not saying anyone has done this!) >The bottom line is: Is the 486 faster than the 386? My answer is YES. I certainly wasn't questioning this fact. I was just trying to caution against grabbing on to one benchmark as the ultimate comparison authority. I find it interesting that in Personal Workstation's "Great Performers" section, they quote only one integer benchmark (Dhrystone), yet quote as many as six floating poing benchmark results. -- John W. Temples -- john@jwt.UUCP (uunet!jwt!john)