User Tools

Site Tools


benchmarks2
Translations of this page:


"ACCELERATED" BENCHMARKS ON RETROCOMPUTING

The purpose of this article is to extend the results obtained from the previous benchmarks by adding a series of “accelerated” configurations. The test will be the same: the average data transfer speed when reading a file sequentially from a random access storage (disk drive) using the standard C library1). The platform is then tested using specific configurations, set up using emulators. As a result, the set of supported platforms by MIDRES library remains the same.

PREMISE

At the time of publishing the previous article, I received a number of recommendations regarding the adoption of acceleration hardware. Some of these hardware were widespread at the time of the commercialization of the various home computers; others have become; others never but, however, are still very interesting in terms of performance.

For this reason I decided to repeat the same tests by reconfiguring the emulators to support these “accelerated” configurations. Clearly, the test cannot and does not want to be exhaustive, as there are so many other hardware combinations. Also, I would like to remember that the purpose of this article is not to make comparisons between different hardware, but to identify those specific configurations that can replace the memory expansion (RAM) in order to store videoclip in the IAV format.

Also, since I have tested emulators, it is possible that the results obtained may be different if you are using actual hardware. Unfortunately, unlike the previous test, the emulators cannot guarantee that the operation is right as expected.

METHODOLOGY

Please, refer to the previous article for the description of the methodology used.

HARDWARE COMBINATIONS' TESTED

Below is a list of all the combinations:

  1. Commodore 64 (c64);
    1. Disk Drive C=1541 (*.d64) with jiffyDOS 6.01; disk image
    2. Disk Drive C=1541-II (*.d64) with jiffyDOS 6.01; disk image
    3. Disk Drive C=1571 (*.d71) with jiffyDOS 6.01; disk image
    4. Disk Drive C=1581 (*.d81) with jiffyDOS 6.01; disk image
  2. Commodore 128 (c128);
    1. Disk Drive C=1541 (*.d64) with jiffyDOS 6.01; disk image
    2. Disk Drive C=1541-II (*.d64) with jiffyDOS 6.01; disk image
    3. Disk Drive C=1571 (*.d71) with jiffyDOS 6.01; disk image
    4. Disk Drive C=1581 (*.d81) with jiffyDOS 6.01; disk image
  3. Commodore PLUS/4 (plus4);
    1. Disk Drive C=1551 (*.d64) [YAPE 1.1.8]; disk image
    2. Disk Drive C=1551 (*.d64) [YAPE 1.1.8] with IEC emulation; disk image
  4. ATARI 8 bit (atari);
    1. Emulated single disk drive with SIO2SD; disk image
    2. Emulated extended disk drive with SIO2SD; disk image

Each of these disk images were then loaded onto the emulator, thus letting it automatically identify the format and size of the media, and the relative benchmark was launched.

RESULTS

This diagram summarizes, briefly, the test's results:

The result clearly shows the superior performance of the SIO2SD compared to the other solutions. To the point that the other performances are, to some extent, “flattened”. Wanting to represent in a more articulated way the other solutions, which are not based on modern hardware, we exclude the results for the ATARI from the metric.

Put in these terms, we note the superiority of the IEC emulation of the 1551 hardware over the rest. However, it must be said that I am not sure if this performance is realistic, and this is also according to some experts of the emulator.

Also surprising is the poor performance of the JiffyDOS when it is working with a Commodore 64 and a more modern disk drive than 1541. The runner-up is still the Commodore 128 equipped with the “classic” 1541, and slightly slower if equipped with the 1581. At roughly the same speeds as the 1581 we still have the 1541 when used by a Commodore 64, and a bit detached 1571 (again when used by a Commodore 128).

At the end of the performances we have the “out of the box” solution of the Commodore 1551 when connected with a PLUS/4.

SOURCES AND DATA

POWERED BY

1)
The source code was compiled using cc65 compiler version 2.18 (git 0981c02). The results of this test may change if you use other versions and/or other compilers. For the purpose of measurement, MIDRES sources are fully portable to other compilers, being written in pure C
benchmarks2.txt · Last modified: 2021/02/28 13:09 by spotlessmind1975