info-mac@utcsrgv.UUCP (info-mac) (06/19/84)
Date: Thu 14 Jun 84 10:49:29-PDT
From: Malcolm B. Brown <uw-beaver!MBBROWN@SU-SCORE.ARPA>
Subject: More comparative benchmarking
To: INFO-MAC@SUMEX-AIM.ARPA
The following are the results of some benchmarking I have done.
Since the only language that is common to all the computers in
question is BASIC, I used BASIC programs. I took them from the April
issue of BYTE, where they were used to test Apple's BASIC for the
Mac. Each program (except for the sieve) repeated some function
5,000 times. The programs included:
Loop (empty for/next loop)
GOSUB (GOing to an empty subroutine)
Midstring search
Integer division
Real division
Eratosthenes sieve for the first 1,899 prime numbers
In cases were I ran the tests myself, I ran the programs twice and
averaged the times.
Times for the Macintosh with Apple's MacBASIC are given according
to the results listed in the BYTE article, page 328.
Here are the results (time given in seconds):
IBM PC HP150 Macintosh Macintosh HP9816
8088 8088 68000 68000 68000
(4.7MHz) (8MHz) (8 MHz) (8MHz) (8MHz)
MS BASIC MS BASIC MacBASIC MS BASIC HP BASIC 2.1
--------------------------------------------------------------------------
LOOP 6 6 1.5 5 .82
GOSUB 12 11.5 3 9 1.5
Midstring 24.5 20 9 22 2.45
Int div 22 17 3 15 3.04
Real div 20 17 18* 32 3.15
Sieve 3:44 3:12 31.5 grrf** 16.5
* N.B. For the MS BASIC programs, I used double precision division
(R#=S#/3#).
grrf*** MS BASIC on the Mac ran out of memory when the sieve program
tried to DIM an array of over 8,000 elements! The largest
array I could construct was 6,770 elements. This enabled the
program to find the first 1,602 primes and accomplished this
with an average run time of 1:44.
--- Malcolm Brown
---------*
-------