ASM Fish for Linux and Windows
Moderators: hgm, Rebel, chrisw
-
Eelco de Groot
- Posts: 4565
- Joined: Sun Mar 12, 2006 2:40 am
- Full name:
Re: ASM Fish for Linux and Windows
It seems a pretty much impossible task to me, to develop and then keep an assembler version of Stockfish up to date with the frequent changes, patches, sometimes total rewrites of some function, and keep the functionality exactly the same, bench the same etc. And he is doing it all alone. It is already very difficult to do for a simple fork of Stockfish, where GitHub can automate part of the job.
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
-
Isaac
- Posts: 265
- Joined: Sat Feb 22, 2014 8:37 pm
Re: ASM Fish for Linux and Windows
As far as I know the bench aren't the same. On one hand the command "bench" does something different in asmFish compared to Stockfish (different depth, etc.). On the other hand there are a few functional differences between the 2 engines.Eelco de Groot wrote:It seems a pretty much impossible task to me, to develop and then keep an assembler version of Stockfish up to date with the frequent changes, patches, sometimes total rewrites of some function, and keep the functionality exactly the same, bench the same etc. And he is doing it all alone. It is already very difficult to do for a simple fork of Stockfish, where GitHub can automate part of the job.
As a side note, I use Linux x86_64 popcnt, and one of the compile showed an about 2.49 speed up compared to original SF (go infinite and wait until depth 22 is reached). Whilst 1 compile later, the speedup was only 1.20 or 1.25. Extremely strange. Basically I had 2.1 Mps for one compile and the other had 0.96 Mnps while with the original SF I am about 0.8 Mps more or less.