TalkChess.com

Do you mean maybe MVV/LVA in one and LVV/MVA in the other???

Yes, thats what i mean.
Thank you for correcting my wording.

I achieve the best nodes/seconds while using ascending MVV-LVA in AlhbaBeta and descending MVV-LVA in QSearch.

Is this common?

--
Srdja

I tried to port the MicroMax 0x88 movegeneration to opencl. on different GPUs. It performed very slow compared to the available computing power. I guess the SIMD-like GPU-architecture and the loop-iterations from 0x88 move generation exclude each other. A Magic BitBoard approach couldnt be completed...

Do I have to change the entire code to have stockfish optimized for parallel computation or will the code structure allow me to insert the new code in the legacy one without changing the original one. Imo you have to redesign the complete code and architecture to perform well on a GPU. So it could ...

...ran some tests...packing and unpacking lowers performance significantly.

hmm, maybe you should init the result array

What is result[] doing ? Assumed the score for an position is stored in sq[] then you got the total score of a file in result[]. Where result[0] is file a and result[7] is file h. for (i =0; i<64; i++) { result[i-(i&56)]+=sq[i]; } for (i =0; i<8; i++) { cout << result...

Ok, i got a clue how this works....will think about that.

would be a reasonably efficient way to generate moves.

I dont get it. What is the content of nextMoveTable ?

not sure if this is what you look for: