From: wnight@cafe.net (William Night) Newsgroups: comp.emulators.apple2 Subject: Speed Test Results Date: 22 Apr 1995 11:19:31 GMT Organization: Cafe.Net Internet Access System Lines: 63 Message-ID: <3naoo3$9q1@scipio.cyberstore.ca> NNTP-Posting-Host: espresso.cafe.net X-Newsreader: TIN [version 1.2 PL2] These are the latest results from SpeedTest 2.2.2 AplWin2 Sim2e Apl2em-2 Applemu Test CPU-Index 6.472 3.155 8.532 11.47 TEXT-Index 2.629 0.1957 4.470 6.274 VIDEO-Index 1.388 0.1957 3.990 5.271 DISK-Index 3.297 0.4715 5.690 0.9626 I will post the results from all the tests as soon as I get around to typing it in. These highlight the actual strengths and weaknesses of each emulator and are usefull in figuring out how fast an emulator will run a certain class of programs. The tests used will remain the same for each minor version level. All patch-level changes 2.2.x etc will just be to the main program and the results will be the same as all other patch levels with the same major and minor versions. For example the results for 2.2.2 and 2.2.8 will be the same, while the results between 2.2.2 and 2.3.2 are not gauranteed to be. These results were obtained on a 486dx4/100 with 4-Dos and QEMM loaded, the AplWin tests were performed in Windows 3.1 with the appropriate Win32 extensions added. I have a cirrus logic 5426 1MB VLB card installed. My system tests out at 195.7 in Sysinfo from Norton. I don't have any Video benchmarks that are common enough that the results would mean anything (see #1 below) A few comments about the results. 1) Until I get a benchmark program for the machine the emulator is running on that all people involved can agree on, test results do not mean anything between platforms or even between two of the same type of machines running at different speeds. The only way the results should be used is to compare the relative speeds of two emulators on the same machine. 2) The relative speeds compared to a //e do not mean that the emulator will perform that quickly or slowly on a video game or other application, for example, Sim2e appears to be very slow, but plays games quite well on my machine. The tests are very intensive in a specific area, where most programs are more general in their use of system resources. 3) These tests aren't completely fair. The CPU tests are for the most part, but differences in the way an emulator is implemented make certain things automatically faster or slower. For instance, both Apl2em-2 and Applemu use IBM text for the apple text screen, therefore all text operations are handled by the IBM graphics card, both Sim2e and AplWin2 do their own translation of text to graphics (and achieve a much more realistic looking screen). With the video tests, Applemu will only do black-and-white high-res graphics. This gives it an unfair advantage over the others. Any suggestions for the next version of speed-test or comments or questions are welcome. The new version of SpeedTest will be posted in the next message, and will probably make it's way onto an FTP site soon.