Search found 628 matches

by smatovic
Tue Jul 02, 2019 10:27 am
Forum: Computer Chess Club: General Topics
Topic: How many Qubits need to slove chess?
Replies: 21
Views: 1009

Re: How many Qubits need to slove chess?

smatovic wrote:
Tue Jul 02, 2019 9:37 am
...
But referring to your original question, I estimate it will be about 1 million Qubits.
Okay, maybe I was a bit too quick in my estimation :)

It should be at least 1 million Qubits, considering that you will need some further
logic for legal moves and to determine the game state.

--
Srdja
by smatovic
Tue Jul 02, 2019 9:37 am
Forum: Computer Chess Club: General Topics
Topic: How many Qubits need to slove chess?
Replies: 21
Views: 1009

Re: How many Qubits need to slove chess?

How many Qubits need to slove chess? ...or how many QuDits? https://en.wikipedia.org/wiki/Qubit#Variations_of_the_qubit https://www.bbvaopenmind.com/en/technology/digital-world/towards-the-quantum-computer-qubits-and-qudits/ https://www.cio.com/article/3099631/russian-scientists-just-found-an-easie...
by smatovic
Tue Jul 02, 2019 6:09 am
Forum: Computer Chess Club: Tournaments and Matches
Topic: CCRL 40/40, 40/4 and FRC lists updated (15th June 2019)
Replies: 22
Views: 1790

Re: CCRL 40/40, 40/4 and FRC lists updated (15th June 2019)

Modern Times wrote:
Thu Jun 27, 2019 10:01 pm
Working fine so far.zeta.png
Hi again, not sure if you are already testing Zeta, just want to let you know,
that I will probably do another release next weekend, to fix some issues
with parallel search scaling above 32 workers....

--
Srdja
by smatovic
Sun Jun 30, 2019 6:52 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: ABDADA+ suggestion
Replies: 17
Views: 1088

Re: ABDADA+ suggestion

Yes, but nobody claimed that it is an perfect scheme... -- Srdja Ah yes, but we're trying to improve the scheme in this topic, yes? Touche. <Between Phase 2 and 3> If all other sons are already analyzed by other processors, then return to parent node, continue with next move, and set flag for secon...
by smatovic
Sun Jun 30, 2019 8:54 am
Forum: Computer Chess Club: General Topics
Topic: Summer time - how does your PC cooling solution perform?
Replies: 11
Views: 686

Re: Summer time - how does your PC cooling solution perform?

Ah, I did not thought about people using AC in their houses, I know them only from data-centers.

--
Srdja
by smatovic
Sun Jun 30, 2019 7:43 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: ABDADA+ suggestion
Replies: 17
Views: 1088

Re: ABDADA+ suggestion

@ Srdja, here's the ABDADA paper I'm using: https://dl.acm.org/citation.cfm?id=228345 Hmm, maybe I don't get you, but if you consider move ordering for phase #2 and #3, then P.35.2* has an lower priority than P.2.2.*. From my reading of the paper, "iteration = 0" will send a thread to P.35.*. This ...
by smatovic
Sat Jun 29, 2019 7:41 pm
Forum: Computer Chess Club: General Topics
Topic: Summer time - how does your PC cooling solution perform?
Replies: 11
Views: 686

Summer time - how does your PC cooling solution perform?

As the title says, it is summer time in the northern sphere, so can you still run your gear under full load with temperatures >30°C? Wonder how dual GPU setups are doing... -- Srdja PS: srdja@ubuntu:~$ top top - 21:44:48 up 14:13, 1 user, load average: 0,45, 0,41, 0,22 Tasks: 248 total, 1 running, 1...
by smatovic
Sat Jun 29, 2019 6:31 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: ABDADA+ suggestion
Replies: 17
Views: 1088

Re: ABDADA+ suggestion

ABDADA (and YBW) are both speculative algorithms: they do work hoping that the work was not wasted. It would seem to me that the goal of speculative execution is to do two things: 1. Select nodes (and subtrees) that have a high-chance of being necessary work. 2. Avoid nodes (and subtrees) that have...
by smatovic
Sat Jun 29, 2019 4:49 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: ABDADA+ suggestion
Replies: 17
Views: 1088

Re: ABDADA+ suggestion

Between phase 2. and 3. there could be a clause added: If all other sons are already analyzed by other processors, then return to parent node, continue with next move, and set flag for second iteration (phase #3). Maybe this kind of clause is already implied in 3., but as I undestand it, in 3. all ...
by smatovic
Sat Jun 29, 2019 4:47 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: ABDADA+ suggestion
Replies: 17
Views: 1088

Re: ABDADA+ suggestion

So, if I were to create speculative algorithm, how would it work? 1. Oldest brother must be evaluated to completion (P.1.1.1.1.1, P.2.1.1.1.1, P.35.1.1.1.1). Remember that I'm assuming Depth 5. 2. Instead of spawning all brothers, only spawn 2 or 3 brothers (P.1.2.* through P.1.4.*. "Block" P.5.* t...