Short: RawSpeed Controller Performance Test Author: SpeedGeek & Robert Miranda Uploader: thebajaguy gmail com Type: util/moni Version: 1.12 Architecture: m68k-amigaos Distribution: Aminet RawSpeed Controller Performance Test v1.12 (c) November 1, 2017. Original source code v1.10 July 1992 provided by Russel Miranda. Code additions and optimized by SpeedGeek for SAS C 6.58 compiler. Documentation revisions and code adjustments by Robert Miranda. Permission obtained in October 2017 by Robert Miranda and SpeedGeek to revise the RSCP source to the most recent v6.5x SAS C compiler, to clean up / groom the code for more stable values under 68040/68060 CPUs, and to add additional features. RSCP is freely redistributable, but may not be modified in any way without written permission from the author. It may be made part of any collection of software, as long as no fee is charged other than a reasonable duplication fee. WHY RSCP? There are several benchmarks already. Why write another one? Well... other benchmarks measure how fast your drive and controller are, but don't judge how much of a performance hit your multitasking is taking while disk I/O is being performed. This benchmark shows how much time is stolen from your Amiga's other duties during controller transfers. WHAT'S NEW in V1.12? - The bar graph has been revised to display up to 4000K/sec. - The Dhrystone benchmark memory allocations were adjusted to provide more consistent values with 68040/68060 CPUs. - The code was compiled with SAS C v6.58, and a 68020+ compiled version is provided (just shy of 2K smaller code). - The buffer selector now cycles in descending order, 512K-128K-16K - An Enforcer hit (illegal read) in the About page was squashed. This version is a cleanup / freshening of a tool that has found use in many a performance tester's toolbox. The Dhrystone benchmark is not comparable to values seen on other systems, or other Dhrystone benchmarks. It is used here to measure the available system CPU resources your system has available to perform processing tasks while multitasking enabled. It is not intended to be a measure of clock speed, or MIPS. Because multitasking is not disabled, the background tools and enhancement features you prefer running on your system will have a small impact even while 'idle'. There is additional information on this version of RSCP v1.12 in the in-program About screen. THE FUTURE OF RSCP? The original RSCP was designed when expansion peripherals were pushing the limits of the 68000-based Amiga systems' bus. The 7Mhz, 16-bit data path of the Zorro II bus meant a maximum of ~3.5MB/sec could pass over it, although hardware developers had some tricks still up their sleeve. We are looking to leave RSCP 1.xx as a useful tool for evaluating KS 1.3/68K machines' performance. It can certainly be used on newer systems, too, and it's results still have value. SpeedGeek and I (Robert Miranda) are looking into additional features and measurements that will be useful on the fully 32-bit, and faster main data-bus, of the A1200/3000/4000 series systems. A number of optimizations are known to be possible depending on the OS (2.0 or 3.1+), CPU (68020+), memory design, 68040/68060 (if equipped) software library, and SCSI/IDE Disk I/O options present in the system. If this research effort bears fruit, and time allows, there may be something new again before another 25 years passes... ;-)