TalkChess.com

I have included in the executable the 4men syzygy tables.

I have just released a new version of my engine. This are the changes: Tuning Probcut Removed razoring Speed optimizations Store in the hash a move also in allnodes Fixed a bug with multipv Don't reduce a low depths LMP tuning SEE pruning tuning Use history ordering also in qsearch Clear killer at p...

With gcc you can use also -flto when compiling and when linking.

The elo increase given by singular extension is different from engine to engine. In my engine it gives about 10-20 elo after a lot of tries to find the best formula for the depth of the search to find if the move is singular. Stockfish uses depth/2 but in my engine was not optimal.

I've found the problem! In init_table there is this code: for (int t = 0; t < num; t++) { data = (uint8_t *)(((uintptr_t)data + 0x3f) & ~0x3f); ei[t].precomp->data = data; data += size[t][0][2]; if (split) { data = (uint8_t *)(((uintptr_t)data + 0x3f) & ~0x3f); ei[num + t].precomp->data = data; data...

Please, read carefully the code.

I've found the point, this is the original map_tb code: static void *map_tb(const char *name, const char *suffix, map_t *mapping) { FD fd = open_tb(name, suffix); if (fd == FD_ERR) return NULL; void *data = map_file(fd, mapping); if (data == NULL) { fprintf(stderr, "Could not map %s%s into memory.\n...

There are a lot of engines that uses internal tb, my engine would not be the only one. Stockfish uses a bitbase too.

I've found a difference in decompress_pairs, the initialization of the variable "code" is different, the other variables are all correct but I can't fully understand the code.

These are the changes that I've made: This is the array of the tb files and the file data typedef struct { char FileName[32]; int FileSize; const void *FilePointer; } TInternalTB; // Array containing all the tb files #define DIMTBDATA 543332 const uint64_t TBData[DIMTBDATA]={ ... ... }; #define TBFI...