Viz wrote: ↑Tue Jul 02, 2024 5:45 pm all of them can use hardware almost equally as efficient.
And that is done in the WCSC played alongside the WCCC.
Ah yes, that famous WCSC on 8 core laptop with no GPU, indeed.
Why not? the hardware is allready enough to me and i don't need GPU, my engine works only on the CPU, even i don't need the 8 cores, just single core is enough
its a fair compare of the software between two engines, not the compare of the hardware speeds !
Arbitrary hardware was interesting back when actual hardware skills were involved - like dedicated units, Deep Blue, Hydra. Just buying the fastest off-the-shelf PC components and plugging them together is boring.
However, it could become intriguing again if some energy limit were introduced where exceeding that budget would count as automatic loss. There could even be several energy leagues.
Making an engine run well on a cluster is not completely trivial, and many engines cannot do it at all. So I think hardware cannot be discarded completely.
hgm wrote: ↑Wed Jul 03, 2024 10:15 pm
Making an engine run well on a cluster is not completely trivial, and many engines cannot do it at all. So I think hardware cannot be discarded completely.
yes, of course its not easy to do that, this is very complex to get many cores together and i have fully respect from them that can do it.
But this has nothing to do with the real strength difference between two engines, just only added many cores of hardware to get more speeds and performance that not comes from the software it self.
Even i will say that Deep Blue and Hydra was not very good implemented algorithm, they just profit from the speed of the hardware.
hgm wrote: ↑Sat Jun 29, 2024 7:39 am
The purpose of a computer chess championship is not to measure the strength of the engines
One caveat - it should be, at least as close as you can do in a tournament format.
If you have any other purposes - don't call it a championship. Call it a "friendly beer drinking showmatch" or anything, just don't call it championship.
You don't get it. There are other aspects to compter chess than engines. Opening books count. Hardware counts.
You can argue for opening books but you definitely can't argue for hardware, sorry.
We are not in 1960 where chess engines could run only on specific hardware and wouldn't run on anything else.
With current knowledge any AB engine can run at any sort of modern PC - from phones to supercomputers - and scale as well as any other engine more or less.
So comparing different engines with different hardware is completely pointless because all of them can use hardware almost equally as efficient.
Sorry to come late to the party. Only just seen this debate on the thread...
My tuppence worth is that I've always found the WCCC's no-limit-hardware to be part of the excitement of the event. The underdogs on consumer hardware had nothing but upside in the David vs Goliath matches. It was a privilege to be there in 1992 when Ed Schroder became the first non-mainframe World Champion. Same in 1995 when all in the playing hall were rapt with excitement with Deep Blue Prototype losing its queen for a rook during its loss to Frans Morsch's Fritz. What a day!
I obviously agree with you that 10-round tournaments (only 5 in the "old days" ), won't guarantee an accurate ranking of participants. Why else would Woodpusher's 66MHz 486 finish 3 places ahead of Lachex's Cray X-MP 48 in the 1995 WCCC? Spoiler alert: it was (a lot of) luck! However, they are marvelous events for spectators and programmers alike.
smatovic wrote: ↑Mon Apr 22, 2024 7:55 am
I think this about finishing a journey that started ~50 years ago....have a good time!
pijl wrote: ↑Wed Mar 06, 2024 10:39 pm
* This will be the last WCCC. After 50 years, it's time to declare success. The WCCC started out with ELO 1600 programs and today we have ELO 3600 programs. It's time for a celebration...
--
Srdja
"I think this about finishing a journey that started ~50 years ago", as in "doing the Compostela". haha!
Unfortunately for the potential pilgrims, Compostella is pretty dull, tourists aren't really there for the nightlife, restaurants are unexiting and few in number, likewise tapas bars and the hotels (pilgrims just don't support it). You're in fact expected to go another 20kms to the coast at Finisterre, take of all your clothes and burn them in a fire. Not sure what happens after that. Airport is out of town and mostly budget airlines, so not easy access. Off season Spain would be better, more accessible, more fun if they'ld have chosen one of the Mediterranean cities, hell, I might even have driven there, but Compostella is just way off route.
My understanding is that ICGA is simply colocating the event with the 2024 European Conference on AI (https://www.ecai2024.eu/). If ECAI were being held somewhere else this year, than the WCCC would be there instead too.
), won't guarantee an accurate ranking of participants. Why else would Woodpusher's 66MHz 486 finish 3 places ahead of Lachex's Cray X-MP 48 in the 1995 WCCC? Spoiler alert: it was (a lot of) luck! 
However, they are marvelous events for spectators and programmers alike.
