The future of computer chess

[Vinvin]

With the new 68000 cpu richard lang was capable to program a very strong engine in assembler.
And with the use of 68020 or 68030 and the increase of the mhz from 12 to 33 or 66 or even 133 mhz
the software was made faster.

The next big step was that the pcs made progress from a few mhz into much more mhz.
We remember 8086, 286, 386, 486-33, 486-66 and more. Then came pentiums much more speed.
And one day we had cpus with 2 cores, then with 4 cores and the race continues,

Today we have very fast cpus with lot memory, high speed and several cores,

So how big was the software progress that was not due to big hardware progress ?

Someone posted a study around a cpouple of years ago about progress : I remember : Chess computers improved by around 1800 Elo in 30 years : 1200 Elo due to software and 600 Elo due to hardware, so in average +40 elo/year and + 20 elo/year.
forum3/viewtopic.php?p=851253#p851253
Here's the study : forum3/viewtopic.php?p=703278#p703278

[mclane]

Yes but the 2 programs have nothing to do with each other.

[chrisw]

With the new 68000 cpu richard lang was capable to program a very strong engine in assembler.
And with the use of 68020 or 68030 and the increase of the mhz from 12 to 33 or 66 or even 133 mhz
the software was made faster.

The next big step was that the pcs made progress from a few mhz into much more mhz.
We remember 8086, 286, 386, 486-33, 486-66 and more. Then came pentiums much more speed.
And one day we had cpus with 2 cores, then with 4 cores and the race continues,

Today we have very fast cpus with lot memory, high speed and several cores,

So how big was the software progress that was not due to big hardware progress ?

Someone posted a study around a cpouple of years ago about progress : I remember : Chess computers improved by around 1800 Elo in 30 years : 1200 Elo due to software and 600 Elo due to hardware, so in average +40 elo/year and + 20 elo/year.
forum3/viewtopic.php?p=851253#p851253
Here's the study : forum3/viewtopic.php?p=703278#p703278

Prefer a three way split, improvements due to software algorithms, improvements due to hardware speed, improvements due to software algorithms made possible by memory increases (which is a form of hardware improvement).

[Rebel]

A fourth one, volume testing. It was invented by Christophe Theron (Chess Tiger), he had 2 poor Pentium 90's, he could do 2 bullet test runs of 800 games a day which put him on top for about ~2 years. The idea was adopted by Vasik Rajlich (Rybka) who had much better hardware, the idea leaked by phishing Vasik Rajlich and is now in use by everybody.

[mclane]

I wonder if there really was a SOFTWARE progress in computerchess that was NOT caused by hardware progress.

If there really WAS a software progress it should be possible to best martin bryant, richard lang or ed schroeder on those 8 bit hardware. This means 6502 or z80a and maximum of 64 kb rom and 8 kb ram.

IF , and yes IF there was ever a progress concerning algorithms, that is not based on hardware progress, it should be possible to best martin bryant, richard lang or ed schroeder on their 8 bit machines.

Or ??

Maybe there is no software progress but instead hardware progress in computer chess.

[Uri Blass]

I wonder if there really was a SOFTWARE progress in computerchess that was NOT caused by hardware progress.

If there really WAS a software progress it should be possible to best martin bryant, richard lang or ed schroeder on those 8 bit hardware. This means 6502 or z80a and maximum of 64 kb rom and 8 kb ram.

IF , and yes IF there was ever a progress concerning algorithms, that is not based on hardware progress, it should be possible to best martin bryant, richard lang or ed schroeder on their 8 bit machines.

Or ??

Maybe there is no software progress but instead hardware progress in computer chess.

There is certainly software progress but nobody is interested to change their software to run in very old hardware that nobody is using.

From the ssdf rating list in dos format
145 Rybka 1 Beta K6-2 450 MHz 2774 64 -70 115 38% 2857
232 Genius 6.5 K6-2 450 MHz 2481 28 -28 605 48% 2494

253 Chess Tiger 11.8 Pentium 90 MHz 2389 42 -42 261 50% 2390
262 Genius 5 DOS Pentium 90 MHz 2336 17 -17 1558 47% 2355

[lucario6607]

A leela net the size of an llm would be cool. Currently the biggest llm has about 10000x more parameters than one of the biggest leela nets

[Jouni]

There has not been any significant progress in the last 3 years. The future is already here ?

[syzygy]

I wonder if there really was a SOFTWARE progress in computerchess that was NOT caused by hardware progress.

If there really WAS a software progress it should be possible to best martin bryant, richard lang or ed schroeder on those 8 bit hardware. This means 6502 or z80a and maximum of 64 kb rom and 8 kb ram.

IF , and yes IF there was ever a progress concerning algorithms, that is not based on hardware progress, it should be possible to best martin bryant, richard lang or ed schroeder on their 8 bit machines.

Or ??

Maybe there is no software progress but instead hardware progress in computer chess.

I think you have posted this argument over a dozen times, but did you ever think of letting Fruit or Rybka from 2005 or so play the latest Stockfish on current hardware?

[mclane]

Current hardware would mean 1 cpu since the older chess engines cannot run more then one core.
In a uniform platform tournament, so same hardware, one can try to find out.