
From: ballyalley@groups.io <ballyalley@groups.io> on behalf of Matt Pilz <info@matthewpilz.com>
Sent: Sunday, July 2, 2023 8:23 PM
To: ballyalley@groups.io
Subject: Re: [ballyalley] Creative Computing Benchmark from 1984

On July 2, 2023, Bill Loguidice asked: Has anyone ever tried to run a version of that BASIC program on the Astrocade

I just did! The problem with that particular benchmark and Bally is that out-of-the-box AstroBASIC does not support floating numbers, built-in square root or exponents. I made revisions to the script to conform to Ballys capabilities, including use of a basic square root function from the Sebrees Math Functions. Also Ballys RND deviates from most in that RND(1) will only return 1 instead of 0-1, so instead I used RND(2)-1 to achieving the same objective. The lack of decimals also means the Precision value (which Bally returned as 990) is not meaningful here.

BENCHMARK TIME: 3:18
RANDOMNESS: 10-12 (Average)

To my surprise this puts it speed-wise ahead of the bottom 25 on the list, including the TI-99/4A, Atari 400/800, ZX Spectrum and Sinclair 2068. Some months back I had ran a more stripped down early-80s bubble sorting benchmark on Bally to compare to other 8-bits, and it was considerably slower. Ive been meaning to make a compilation testing the same benchmark on all my various 8-bits here, maybe as a summer project.

For anyone interested, attached is the source and WAV file of this benchmark using my variations.

Matt

