On-line engine blitz tourney December

[Joost Buijs]

The monthly on-line engine blitz tourney for December will take place on:

Saturday December 12, 2:00 PM EST (Boston time), 20:00 CET (Amsterdam time)

To connect:

winboard -zp -ics -icshost nightmare-chess.nl -icshelper timeseal -fcp ENGINE.exe -fd ENGINEFOLDER -autoKibitz

(for UCI engines, add -fUCI)

People without a server account can submit a request for a (new) account to admin@nightmare-chess.nl

Our tournament manager will be: Eendje (Lisebeth) from ICC

The format will be 9 rounds Swiss 5/1.

Because it is impossible to compile the mamer tourney manager with a recent version of G++ we have to stick with 9 rounds.
Although it compiles without errors both 32 bit and 64 bit builds produce a partly broken executable.
Recently I started to rewrite parts of mamer to remove a lot of probematic stuff like using long for pointers which of course breaks a 64 bit build.
It contains a lot of C++ violations too that were probably fine 20 years ago.
Until the rewrite is finished we have to stick with the old format, max. 9 rounds and no McMahon system.

[hgm]

Is it not possible to force a 32-bit compile? I think I had to do that for the ICS itself too, adding a -m32 compiler flag.

Would it be possible to run a 32-bit executable compiled on a different Linux version? I have one of those.

[Pi4Chess]

Oh that's tomorrow ! Nice !
I will be there with my little pi4

It may stay connected since the morning. So extra games will be possible.

[Joost Buijs]

Is it not possible to force a 32-bit compile? I think I had to do that for the ICS itself too, adding a -m32 compiler flag.

Would it be possible to run a 32-bit executable compiled on a different Linux version? I have one of those.

Of course I tried to compile it with -m32, it compiles without errors but the 32 bit executable has problems too. It logs on fine but some functions behave not like they should. I could start debugging it, but it seems more logical to change al the things that could give problems like using parenthesis to force operator precedence, removing these word-size pointers, and fixing a lot of other things the compiler warns about.

Most of the 32 bit dependencies in the ICS software are gone (if not all), I run a 64 bit executable, it runs fine but it is still vulnerable to very long strings even with your time-seal fix in place. To avoid most problems I increased the buffer size(s) to 4K, this is not a real fix, but I don't want to go through all the code to find the spots where it actually could go wrong.

I use a mamer executable from several years back that is compiled either by you or by Daniel Dugovic, so I suppose that a 32 bit executable compiled on a different (older) Linux version would work.

[hgm]

I have put the version of mamer that I always was using at http://winboard.nl/mamer .

[Joost Buijs]

I have put the version of mamer that I always was using at http://winboard.nl/mamer .

Thanks! I presume this is the one with McMahon and the 10 rounds limitation?

[hgm]

Indeed. I never looked at lifting the rounds limitation; I suppose it must be easy.

[Joost Buijs]

Indeed. I never looked at lifting the rounds limitation; I suppose it must be easy.

It's just one constant in config.h, MAX_ROUNDS, first it needs to compile properly. What I don't understand is that DEFAULT_MAX_PLAYERS == 16, I assume you increased this number in your build, maybe this can be overruled in mamer.config, I have to look at the source, but I didn't see anything like that in your mamer.config.

My intention is to make a mamer version with a CMake build system that can run natively under Windows too without needing the MinGW-w64 runtime.

[Henk]

Most efficient is not to play.

First statement in my alphabeta search is create a key(position)
Second is tryTerminateTheSearch(key)

Only in it for the money.

[Joost Buijs]

I'm only in it for fun. Computer-chess makes you poor, you don't want to know how much money I spent on it during my lifetime.
Still better than powdering my nose.