Need some results for comparison ... would really appreciate a feedback post

[Peter z80.eu]

For a sample series of x86 Assembler learning, I wrote also a small benchmark program.
I've tested it with DOSBOX and with PCem, but the results were ... at least ... weird.
The values are depending also if BIOS ROM Shadowing is used (or not).
I would really appreciate some values for comparison (so I can change the displayed program text to real world values).
See also my blog entry ( http://www.z80.eu/blog/index.php?entry=entry210208-095857 ) and the attachment here (a very small DOS program, with source code).
Thank you in advance for every result you will reach with your vintage computer.

 

[Trixter]

Your benchmark is more a test of the video card RAM speed and video BIOS efficiency than it is of the system. Swapping the video card in a system for a faster one will affect the results. Shadow RAM affects the results because it relocates the BIOS from slower ROM to faster RAM. DOSBox provides unrealistic results because DOSBox emulates every instruction in one cycle.

My point is that I'm not entirely sure what you're benchmarking.

https://dosbenchmark.wordpress.com/ might be interesting for you to investigate.

 

[Jove]

Sergey 8088
Oak 077 VGA card
V20 at 9.55 Mhz
3 loops measured 91 ticks.

IBM PS/2 Model 25-286
onboard VGA adapter
286 at 10MHz
9 loops measured in 91 ticks

 

[Peter z80.eu]

Your benchmark is more a test of the video card RAM speed and video BIOS efficiency than it is of the system. Swapping the video card in a system for a faster one will affect the results. Shadow RAM affects the results because it relocates the BIOS from slower ROM to faster RAM. DOSBox provides unrealistic results because DOSBox emulates every instruction in one cycle.

My point is that I'm not entirely sure what you're benchmarking.

https://dosbenchmark.wordpress.com/ might be interesting for you to investigate.

It's not a CPU benchmark at all, I know. But it does not matter what instructions and I/O is used in detail, running on different systems will give different results. And that's what I just wanted to get. Yes, if the BIOS works more efficient, the results will be better. But this can be an advantage also, a better BIOS, a better graphics card, better (faster) Video RAM. So it was NOT my goal to do it in a scientific way. It just reflects the performance in text mode.
Getting more results will help not only to change the text displayed at the end, but will improve my understanding of the BIOS implementation of different machines, too.
If you do not like to try it out, it does not matter. May be someone else just give it a try, despite the lack of a "scientific approach". Every additional result I will get (from real machines) will make me more happy and I thank everyone (else) for posting results in advance.