TalkChess.com

OK, so I need to do some analysis on powerful computer. I do not own one, but I feel ready to pay for renting it. So far I have found two options:

1) Rybka cluster - the price is ridiculous and the engine is obsolete so that's not an option.
2) nextchessmove.com - 20 CPU is fine but they offer max only 15 seconds for analysis which is a joke.

Are there any other options? Or perhaps there is someone ready to rent some computing power for a little fee?

Best
Karol

Karol Majewski wrote: ↑Tue Jul 24, 2018 9:35 pm OK, so I need to do some analysis on powerful computer. I do not own one, but I feel ready to pay for renting it. So far I have found two options:

1) Rybka cluster - the price is ridiculous and the engine is obsolete so that's not an option.
2) nextchessmove.com - 20 CPU is fine but they offer max only 15 seconds for analysis which is a joke.

Are there any other options? Or perhaps there is someone ready to rent some computing power for a little fee?

Best
Karol

Just post your position here. You'll get all sorts of free analysis.

If you want to buy a machine with a lot of power, you can get very powerful used server machines for a cheap price on e-bay.

For instance:
https://www.ebay.com/itm/Dell-PowerEdge ... 3278057747

However, commercial server machines are REALLY loud because the squirrel cage blowers are meant for a air-conditioned computer room and not for a house.

I ended up buying soundproof boxes that are about $800 each.

OK, so here's the position. Black to move. After moving pieces and looking at evaluations I came to the conclusion that 10...Bd7 is objectively stronger than 10...Nd7. Now I want to - kind of - prove it and check if the engine (probably latest Stocky) will find 10...Bd7 in a single run. This requires very deep analysis and very powerful machine, I think. Best line should start like: 10...Bd7 11.Nxd7 Nxd7 and now probably 12.Be2 or 12.Rd1 or even 12.g3...

[d]r1b1kb1r/1pq2ppp/p3pn2/2p1N3/Q2P4/2P1B3/PP3PPP/R3KB1R b KQkq - 3 10

r1b1kb1r/1pq2ppp/p3pn2/2p1N3/Q2P4/2P1B3/PP3PPP/R3KB1R b KQkq - acd 36; acs 632; bm Bd7; cce -72; ce -52; pm Nd7 {138} Bd7 {60}; pv Bd7 Nxd7 Nxd7 O-O-O O-O-O Kb1 Nb6 Qb3 Kb8 g3 c4 Qc2 Qc6 Rg1 Qb5 Bd2 Bd6 Ka1 h6 Bg2 Nd5 Qe4 Nf6 Qf3 Rd7 Rde1 Qa5 h4 Qb5 Re2 Qa5 Kb1 Nd5 Ree1 Nf6 g4 Qb5 Bf4 Qb6;

Thanks Dann, what was the final depth and final eval? I'm pretty stunned...

Only 36 plies. Any position that has been played a lot, I have already analyzed.
36 is my rock bottom depth. If it is a position I am interested in, I usually go to 40 or more now.

Then engine (probably stockfish but it could be Komodo or Houdini) favored Bd7 also, but gave an advantage of about 1/2 pawn to white.

Here's what Cfish-dev has now. The search is ongoing:

+0.35 10... Nd7 11. Bb5 cxd4 12. Bxd7+ Bxd7 13. Qxd4 Bb5 14. a4 Bd6 15. Nxf7 Kxf7 16. axb5 axb5 17. Ke2 Be7 18. Qe4 Qc6 19. Qxc6 bxc6 20. h3 Rhb8 21. Kd3 h5 22. Rhd1 Kf6 23. Ke4 e5 24. Rdb1 Ke6 25. Rc1 Ra4+ 26. Rxa4 bxa4 27. Ra1 Rxb2 28. Rxa4 Rc2 29. Kd3 Rb2 30. Ra6 Kd6 31. Ra7 Bf6 32. g3 Ke6 33. Ra6 Kd6 34. Ke4 Rb3 35. Kf5 Rxc3 36. g4 hxg4 37. hxg4 Rc2 (depth 45, 0:10:24)

Dann Corbit wrote: ↑Tue Jul 24, 2018 10:52 pm Only 36 plies. Any position that has been played a lot, I have already analyzed.
36 is my rock bottom depth. If it is a position I am interested in, I usually go to 40 or more now.

Then engine (probably stockfish but it could be Komodo or Houdini) favored Bd7 also, but gave an advantage of about 1/2 pawn to white.

Hmmm... 1/2 pawn is to much to call it equality. I'm searching for an equlizer. I belive that 10...Bd7 in a long run should give something around 0.2. So hard to prove it.

zullil wrote: ↑Tue Jul 24, 2018 10:59 pm Here's what Cfish-dev has now. The search is ongoing:

+0.35 10... Nd7 11. Bb5 cxd4 12. Bxd7+ Bxd7 13. Qxd4 Bb5 14. a4 Bd6 15. Nxf7 Kxf7 16. axb5 axb5 17. Ke2 Be7 18. Qe4 Qc6 19. Qxc6 bxc6 20. h3 Rhb8 21. Kd3 h5 22. Rhd1 Kf6 23. Ke4 e5 24. Rdb1 Ke6 25. Rc1 Ra4+ 26. Rxa4 bxa4 27. Ra1 Rxb2 28. Rxa4 Rc2 29. Kd3 Rb2 30. Ra6 Kd6 31. Ra7 Bf6 32. g3 Ke6 33. Ra6 Kd6 34. Ke4 Rb3 35. Kf5 Rxc3 36. g4 hxg4 37. hxg4 Rc2 (depth 45, 0:10:24)

Yeah... that's Georg Meier's line (up to 17.Ke2). I've look at this extremely deeply and White is always better here. Hope Cfish will change his mind.