|
|
Get set, Benchmark, Go!
Benchmarks are a popular diversion when using an operating system, and AmigaOS is no exception. Most computer magazines cater for this compulsion and publish at least one of these tests per edition. If the computer in question is Windows based, the famous/infamous Ego-Shooters or MadOnion's 3DMark 2000 are used. But the Amiga, too, can point to quite a few examples of this kind of program, even though the graphics are a bit more moderate. Just think of programs like AIBB, PowerMeter, SysSpeed or P96Speed. An Amiga-user can benchmark until the men with white coats are called.
AmiGOD
A fairly recent representative of this brand of program is AmiGOD [1] by the Czech Lukas Stehlik [2]. The version 1.43 Millennium from December 12, 2000 was put through its paces.
AmiGOD is quite moderate. If there are at least 200 KB RAM available and AmigaOS 3.0 or better is running, the program should run, on a 68k CPU alone or together with a PPC as co-calculator.
AmiGOD refers to the identify.library for information about the Amiga, therefore this should reside in the LIBS: directory. If not, it's no mishap at all, Lukas Stehlik includes an applicable version in his program archive.
Installation is a simple as the system requirements are modest. After uncompressing the LHA archive just move the drawer "AmiGOD" to the required position in the jungles of your HD, and that's it.
You can preset the language in the tooltypes (beside English there are adaptations in German, Czech, French, Italian and Polish languages provided) and a custom font including a variable size for those higher screen resolutions.
The obligatory mouse-doubleclick upon the program icon reveals the main window. In this the user gets a first overview of the hardware and operating system of his Amiga. Here he may accept the reports of the identify.library about among other things CPU, FPU, MMU, graphics chipset, Chip and Fast RAM, Kickstart and Workbench versions. Similar to several other contestants a comment about the supposed performance of the recognized Amiga is related.
A small menu provides access to additional functions of the program. Under "AmiGOD/Database" is quite a collection of benchmark results of very different Amiga configurations allowing a more or less edifying comparison with your own results. The menu option "AmiGOD/Test" opens the actual benchmark.
A total of twelve different categories are ready for choice. This is, compared with other current competitors a more moderate number. Beside selection/deselection of single test steps a button enables the test run itself to be started. Immediately AmiGOD takes over the Amiga and disables multitasking. No matter what screen resolution is set at the Amiga, AmiGOD keeps a test window size of 320x240 pixels.
After a few minutes all test steps are done and AmiGOD presents the results. The gentle user may add the results to the database, but only if he has previously filled the boxes with his personal data.
Whether you add your results to the database or not, you of course can view the existing entries. Two examinees share the display and present their results. Beside the actual test results, clicking the "i" button reveals the hardware configuration of that Amiga including the selected screen resolution and colour depth. Because there are nearly 60 entries, there should be a worthy "opponent" for your own Amiga.
Under "UAE-JIT Ipswich Town", too, AmiGOD was persuaded to cooperate without problems. Host-computer in this case was an AMD Athlon with 600 MHz, 128 MByte RAM, a GeForce-2-GTS graphics board and installed SuSE Linux 7.1.
Only the identify.library was not able to recognize the virtual Amiga properly, something observed with other benchmark programs, and offered the (amusing) advice to speed up the Amiga by purchasing a better CPU board. But the test run itself revealed, that even an Amiga1200-060/50 with CyberVision64/3D and a lower resolution of 1024x768x16 is inferior to the emulated Amiga under UAE in the benchmark.
| And this, even though the Amiga only ran in an X11 window, in DGA mode the results are |
| radically better for UAE. This is similar to | the results of SysSpeed and P96Speed. |
The custom font selection is realised only half-heartedly. The selected font is not taken into consideration in all program windows. Furthermore bigger fonts can have part of the written information spill out of the provided editing area and become unreadable.
Benchmarks should be accepted with a good grain of scepticism, and the whole system will certainly not be characterized exhaustively by the results.
However, especially after the installation of new CPU or graphics hardware it is quite some fun to see the gain of performance in black and white or whatever colour combination.
AmiGOD worked its test program on all available configurations without any problems. Even if it cannot keep up with the test span of SysSpeed and P96Speed, for a little test in between it is always appropriate. :)
[1] on Aminet or http://www.volny.cz/akf/download.html
[2] luky-amiga@volny.cz
Uwe Pannecke <Uwe.Pannecke@t-online.de>