View Full Version : generic benachmark for cp/m systems

June 1st, 2017, 05:15 AM
Hello, there are some guys in forum64.de who want to compare the performance of some machines running CP/M to each other. Are there somewhere some common benchmarks which can run on a big variety of CP/M 2.2 machines? Do you have download link?

June 1st, 2017, 06:33 AM
One of the more common sets of benchmarks were done by Kilobaud Magazine. These are very short BASIC programs though I have seen conversions of the code to Pascal and Algol.
See http://www.geocities.ws/peterochocki/computers/pcwbm.html for some examples.

June 1st, 2017, 08:50 AM
Unfortunately, this is really a benchmark of the BASIC and not of the physical system. If the BASIC implementation were different across all the benchmarked systems, the results would be useless.

June 1st, 2017, 09:19 AM
It was a good enough set of benchmarks for IBM to rely on during development of the PC. Between the British computer magazines and the Australian academic report, there were results from hundreds of systems ranging from mainframes down to programmable calculators. It was about as close to a universal comparison as existed before 1985.

June 1st, 2017, 09:28 AM
The problem with the Kilobaud set is that for systems using the same processor, it could only quantify the difference between BASIC interpreter/compilers. None of the benchmarks incorporates any I/O testing. Otherwise, a 4 MHz Z80A system, other than for slight differences in memory speed, is pretty much like any other Z80A system.

So a system using disk caching is pretty much indistinguishable from another system with the same speed CPU that doesn't employ I/O speedups.