stockfish 10 vs. Mephisto III S Glasgow

[mhull]

What is MC69030 @20 Mhz ??

And what is your point. Mephisto III uses an 8 Bit cpu with 3.5 or 6 MHz.

Not what you said at the beginning. You said at the top, "Mephisto III S was running on a 68000 with 12 Mhz and made 3-5 NPS."

I owned a Fidelity Mach II LA and the Spracklen program did 1000-2000 NPS on identical hardware that you describe.

The SF 8 test on 68030 at 20 Mhz did 150-600 NPS depending on game phase and board position. So a few seconds at 600 NPS is a lot more than 1000 nodes. And this hardware is only 30-40% faster than your 68K @12 Mhz. At the time control you described, 1000 nodes/move is arbitrary and cripples the program in some game phases, IMHO.

I don't think your test is valid. If there is a cfish port for the newer SF, I can test it on 68030.

[bob]

One other note. The engine evolved along with the hardware. IE we can do things today, thanks to the faster hardware, than we could do back in the sub 1000 NPS days. So the comparison is not exactly fair...

[mclane]

How do you let modern programs run on 68030 hardware ?

[Uri Blass]

I have no idea what you talk about.

I sum up:

Mephisto III S Glasgow 68000 (5-10 NPS) wins against stockfish 10.
Even worse
Mephisto III 8 bit 1802 1-3 NPS with 6.1 MHz wins against stockfish 10.
Mephisto III 8 bit 1802 1 NPS 3.5 MHz on 4 AA batteries wins against stockfish 10 too.

Ic0 plays better chess then Stockfish.

Stockfish has only a chance if it has a hardware advantage of 1000-1500 times better hardware and 1000 times better ram and rom and core advantage of factor 4-64.

So the question is: is there any software progress at all ??

If a 36 year old engine beats it having the same NPS, has computerchess made any progress at all ??

There is no doubt hardware made progress.

Stockfish is not designed for the conditions that you test.
I am not sure if stockfish10 is stronger than old versions of stockfish or even other free source programs in the conditions that you test.

If you want to test if there is progress in software in your conditions(same low numbers of nodes) then the first thing is to find the best software of today for your conditions.

Maybe there is some free source code program that can beat Mephisto III in your conditions.

Note also that I am almost sure that even with the same number of nodes stockfish is going to win against Mephisto if you give both of them 80000 nodes per move(and it is possible to try it because even with 1 NPS you get 80000 nodes in less than 24 hours).

[jp]

Stockfish is not designed for the conditions that you test.
I am not sure if stockfish10 is stronger than old versions of stockfish or even other free source programs in the conditions that you test.

If you want to test if there is progress in software in your conditions(same low numbers of nodes) then the first thing is to find the best software of today for your conditions.

Maybe there is some free source code program that can beat Mephisto III in your conditions.

I'd also like to see how SF does in his conditions if you go into the code and 'turn off' the aggressive pruning but leave the rest the same.

[mclane]

I have no idea what you talk about.

I sum up:

Mephisto III S Glasgow 68000 (5-10 NPS) wins against stockfish 10.
Even worse
Mephisto III 8 bit 1802 1-3 NPS with 6.1 MHz wins against stockfish 10.
Mephisto III 8 bit 1802 1 NPS 3.5 MHz on 4 AA batteries wins against stockfish 10 too.

Ic0 plays better chess then Stockfish.

Stockfish has only a chance if it has a hardware advantage of 1000-1500 times better hardware and 1000 times better ram and rom and core advantage of factor 4-64.

So the question is: is there any software progress at all ??

If a 36 year old engine beats it having the same NPS, has computerchess made any progress at all ??

There is no doubt hardware made progress.

Stockfish is not designed for the conditions that you test.
I am not sure if stockfish10 is stronger than old versions of stockfish or even other free source programs in the conditions that you test.

If you want to test if there is progress in software in your conditions(same low numbers of nodes) then the first thing is to find the best software of today for your conditions.

Maybe there is some free source code program that can beat Mephisto III in your conditions.

Note also that I am almost sure that even with the same number of nodes stockfish is going to win against Mephisto if you give both of them 80000 nodes per move(and it is possible to try it because even with 1 NPS you get 80000 nodes in less than 24 hours).

To be honest this is not a scientific thesis I want to verify.
I am disappointed that computerchess programmers use the enormous hardware progress we have today to do a race against others or increase ELO instead of concentrating to make the programs better.

I would call this the lazy chess programmer approach,

I could do the same by over clocking the machine instead of programming at all.

It would also benefit the ELO and maybe be enough to beat the opponent. But the program would not be any different.

IMO programmers should consider to do something different.
Ask themselves where they want to go.


In this Forum we Daily see new engines pop out of nothing and they all go the same path.

[corres]

There is no any obstacle to built in the elements of the evaluation of Mephistos engine into the source of Stockfish. But if there is somebody who think such a "Stockfish" would be stronger than the up to date Stockfish is that person is very wrong. Obviously running on the same machine this "Stockfish" will be much more stronger than the Mephistos but much more weaker than the current Stockfish is.
Playing with number of nodes is a very unfair thing because the very low number of nodes gives benefit to those engines what use mainly static position evaluation (like Mephistos and other engines from the old times) and causes big drawback for the engines using mainly dynamical evaluation - like Stockfish.
In my opinion this kind of comparison is absolute superfluous because every engines should exploit those possibilities what the technical and programming development give at the given time.

[mclane]

The question is where do you want to go.

IMO this race about ELO and massive usage of resources is not the way computerchess should go.

NOT making the software faster is the way. But making the software more intelligent.


As mephisto III shows you can generate ELO even out of few NPS.

[mhull]

How do you let modern programs run on 68030 hardware ?

I have made a cross-compiler for 68k architecture. I took Cfish 8 and compiled it for 68030 linux target using gcc.

I wanted to explore a thought experiment, that if programmers circa 1990 had known about software advances 25 years into the future, how much stronger would their programs be on their 1990 hardware against period top programs. The only problem is I do not own any 1990 top programs (or top dedicated machines) to test. I sold all my Fidelity period machines but even they were not in the top league like Mephisto. I could not even test against Chessmaster 2100 because I only had one 1990 machine (Mac IIsi) and Cfish 8 ran under a linux initramfs while Chessmaster needed Mac OS 7. I couldn't run both at the same time.

The solution is to play a skype (or ICQ) match or something with someone who has a 1990 top chess machine, and who is willing to waste the time playing some "standard" games. Or, get the machine in a state where it can play on FICS. That would mean getting a more full Linux or NetBSD to run on the Mac IIsi which is a lot more work.

But as Bob suggests, knowledge of future software advances would more likely have been optimized, not just compiled in C. Maybe they would have used assembler to be more efficient. SF is written in C++ which would be bloatware in 1990, even when translated into C. But I rationalized as follows, "Even if Cfish is bloatware in 1990, and not optimized, it will be the same as if it were running on even older hardware, a kind of handicap. So in reality, SF8 running on 68030 @20 Mhz might be the same as an ASM Fish running on 68000, 12 Mhz. All the better.

The project is on the back burner at the moment, but I intend to pick it up again.

[Laskos]

Lol

This mclane is something. The software progress is tremendous compared to early 1990s. At fixed nodes above 10k or fixed time above 5s/move, if the port of the new SF to the old hardware is reasonable.