My vintage PC only has 128KB RAM, I can’t run the TOPBENCH program. How can I profile my machine and get it into the database? The TOPBENCH “stub” (TOPBSTUB.EXE) should run on any machine with 128K RAM. When it runs, it dumps information into a text file called output.ini which you can then import into TOPBENCH running on a different computer, or you can email me the file to include into the next database release. See Downloads for the stub.
When I run TOPBENCH with -i (prints the score, then exits), I get a lower score than what is in the database. Are you running it from a floppy disk? Many 286-486 systems slow down when the floppy drive is active as a way to be more compatible with copy-protection schemes. Try running TOPBENCH with -l which continuously prints the score; you should see it jump up as soon as the floppy motor stops.
Why is it important to run PC games at the proper emulation speed? The earlier a game was written, the more susceptible it was to machine speed. Many PC games made before 1985 will only run properly on one single speed, that of a 4.77MHz 8088. For maximum enjoyment, speed needs to match what the developer of the game expected when they wrote it.
It’s also important to get speed right when doing historical research. It’s not fair to judge a work if that work is not being represented properly.
Will I be able to run TOPBENCH on any machine, even current ones? Yes, but doing so is somewhat pointless since the performance metric was designed for a single-core 16-bit processor. But it does work, and should work for decades as long as the systems can boot DOS. The resulting scores can be comically high, btw.
Your metric code is too small; it can fit completely inside an L1 cache. A valid complaint, however only 80486 chips and higher have L1 caches, and such machines are the upper bound of TOPBENCH’s intended audience.
Your metric code is/isn’t optimized for pipelining. See previous question 🙂
Your “3-D Game” metric code doesn’t execute instructions in the same order as actual 3-D games, making it less than an optimal comparison to actual 3-D games. Yes, this was the compromise I had to make. If I ripped exact 3-D game inner loops out and used them, it would be 100% accurate as well as 100% illegal. So the resulting instruction mix is from 8 different games. See the source code for more explanation on the rationale for that particular metric.
When I run TOPBENCH on my LCD/plasma/monochrome CGA display, the colors make everything hard to read. Run “MODE BW80” before starting TOPBENCH and it will pick up the fact that you want a B&W display and adjust accordingly.