Windows 3x on DOSBOx Chess

[spacious_mind]

OK I just played BT2630 Test Position 15 with Classic Champ.

At the WIN-DOSBox Pentium 90 setting of 69,300 cpu_cycles it took 54 seconds to find the correct move 1. e5

At the Win DOSBox max (speed) setting it found the correct move in 13 seconds.

So you can approximately estimate that my I7-7700HQ 2.8 GHz laptop plays the emulator at full speed at approximately the equivalent of a Pentium with 374 MHz,

A Pentium 100 had 188 D/MIPS. My emulator plays Windows games by the look of it (or at least Classic Champ) at approximately 703 D/MIPS.

Millennium Chess Genius Exclusive is a ARM Cortex M7 which runs at 2.14 D/MIPS per MHz. Therefore MCGE's 300 MHz = 642 D/MIPS.

So therefore my I7 Win-DOSBox plays about 10% faster than MCGE with 300 MHz.

However what is weird is that it took MCGE 82 seconds to find the move 1. e5!

Other listed Lang's could not find it at all including TM London. So it is pretty clear that Classic Champ from 1995 was an improved Lang and ditto for MCGE.

Best regards

[spacious_mind]

I complete the spacious_mind test with Champ Classic playing at full speed in DOSBox-Windows 3x.

Test Result DB-WIN 3x I7 Max speed - Champ Classic - Richard Lang - 1995



It finished with a test rating of 2443 ELO which is a 160 ELO improvement over playing at Pentium 90 speed. It also wins the game simulation over against the other Lang programs shows above.

At speed doubling the 374 MHz has a doubling faktor of 2.02. Which means if you divide 160 ELO by 2.02 you get a speed doubling of 79.2 ELO.

Overall Champ Classic really benefits from the increased speed. Especially in test game 4. For test game 5 the increased speed really did not do anything. In fact it lost a couple of points.

Now what I think might be an interesting match is to play Champ Classic against Millennium Chess Genius Exclusive.

Best regards

[lexman]

Yes Champ Classic against MCGE would look to be a great match, its got my vote!

[spacious_mind]

I tested both DOS Box and DOS Box Win 3.x using Rebel 10's 3 test positions. BTW you can play Rebel 10 under Windows or DOS.

Rebel 10 DB-Win 3x Max - 2566
Test 1: Key move found at 8 seconds
Test 2: Key move found at 41 seconds
Test 3: Key move found at 59 seconds
Total Time 108 seconds

Rebel 10 DB-Win 3x P90 69,300 - 2415
Test 1: Key move found at 43 seconds
Test 2: Key move found at 192 seconds
Test 3: Key move found at 323 seconds
Total Time 558 seconds

Rebel 10 DB Max - 2567
Test 1: Key move found at 7 seconds
Test 2: Key move found at 37 seconds
Test 3: Key move found at 57 seconds
Total Time 101 seconds

Rebel 10 DB P90 69,300 - 2418
Test 1: Key move found at 41 seconds
Test 2: Key move found at 189 seconds
Test 3: Key move found at 318 seconds
Total Time 548 seconds

The difference using DOSBox or DOSBox Win 3x on my I7 is marginal with DOS Box performing just very slightly better.

Regards

[spacious_mind]

Here are some more programs I tested with the rating test. The are all highlighted in yellow:



Tascbase's King 2.55 running under DOSBox on my I7 at full speed did remarkably well scoring grand masterly ELO 2669.

Chessmaster 4000 running under DB-Windows 3x did also fantastically. Hiarcs 7.01 scored slightly better than Rebel 10. Frans Morsch's Chessica could have scored higher but did quite badly in test games 2 and 5.

Christophe Theron's Chess Tiger 10 after doing OK in games 1-3 collapsed in the last two games which was disappointing.

Anyway I added the non highlighted to the list for comparison and in order to do a quick tournament simulation to see how it would look.



Well just like in real live tournaments the final tournament ranking looks different to the rating ranking. Except for 1st place and last place.

Best regards