Vapor Chess v0.02 Release for Linux

[xsadar]

Vapor is a new(er) chess engine I'm writing from scratch but loosely based on my old private chess engine Zilch. It uses the UCI protocol. Unfortunately, since I started writing it, a couple years ago or so, I haven't had a lot of time to work on it, so I haven't gotten very far. The new release adds repetition and 50-move rule draw detection as well as a transposition table. With the default hash-size (256 MB) this gives roughly a 195 Elo improvement over Vapor v0.01, according to my testing.

You can read more and download it from http://vapor.mikeleany.com/.

A Windows build should be coming shortly. I ran into some issues there. For now, please enjoy the Linux build.

[Graham Banks]

Mike - can you give a rough idea of Vapor's strength? Which engines is it on a par with?

[xsadar]

Mike - can you give a rough idea of Vapor's strength? Which engines is it on a par with?

Well, I was testing it against Mizar, TSCP, Micromax and SSEChess (sorry, I don't remember version numbers). Mizar was significantly worse than even my old version, so I'm not sure why I chose it (probably something to do with the difficulty of finding stable engines that would run under wine, or natively, on Linux and wouldn't lose on time). TSCP and Micromax are about on par with the new version with a very small hash. With the default hash size it's about on par with SSEChess. I believe I used default settings for all engines, but did not use any opening books for testing (using a fixed set of openings instead).

Also see how my old version did here: http://wbec-ridderkerk.nl/html/5thdiv.htm

[Graham Banks]

Mike - can you give a rough idea of Vapor's strength? Which engines is it on a par with?

Well, I was testing it against Mizar, TSCP, Micromax and SSEChess (sorry, I don't remember version numbers). Mizar was significantly worse than even my old version, so I'm not sure why I chose it (probably something to do with the difficulty of finding stable engines that would run under wine, or natively, on Linux and wouldn't lose on time). TSCP and Micromax are about on par with the new version with a very small hash. With the default hash size it's about on par with SSEChess. I believe I used default settings for all engines, but did not use any opening books for testing (using a fixed set of openings instead).

Also see how my old version did here: http://wbec-ridderkerk.nl/html/5thdiv.htm

Thanks for the info Mike.