Guenther wrote: ↑Thu Oct 08, 2020 11:57 am
Dorky website is gone!
It was saved in wayback though (changed the links in my chronology).
Links for Dorky versions above 4.3 have not changed. They were not on Matts site.
Guenther wrote: ↑Thu Oct 08, 2020 11:57 am
Dorky website is gone!
It was saved in wayback though (changed the links in my chronology).
Links for Dorky versions above 4.3 have not changed. They were not on Matts site.
Koivisto wrote: ↑Thu Oct 08, 2020 1:28 pm
I don't think I've mentioned it here on talkchess before, but this is kind of a predecessor to Koivisto, https://github.com/altarchess/Altar. Its not very strong but it is definetly original
This is probably one of its best games (crushing Koivisto when it was around 2800 elo cclr strength): https://lichess.org/51Bn6b2j
I know it and informed Gabor too, but I could never compile it for my hardware.
(also it was an anonymous repo, which isn't my interest, so I had no motivation for fiddling too much with it)
The available binary obviously has popcount and/or instructions above SSE 4.1.
The binary works for me but I couldn't compile it either.
Static exchange evaluation has sadly made Weiss a bit slower but on the plus side compiling is now easier - just 'make' will likely be optimal for any machine. Also no longer requires 'isready' (or anything else) before 'go'
I've spent a good chunk of my free time this summer writing a chess engine in C++ which can be found here: https://github.com/connormcmonigle/seer ... s/tag/v1.0. It should, hopefully, play at around 2900-3000 elo if AVX2 extensions are available (around 300-400 knps should be expected with AVX2). It uses a custom NNUE implementation with training code written in PyTorch and inference code relying on OpenMP SIMD for auto vectorization.
I'm not providing any compiles at this time, though compiling Seer should, with any luck, be straight forward (compilation instructions can be found in the readme). I'd love to see my engine appear on various rating lists if that's a possibility.
Thanks in advance and be sure to inform me if you encounter any issues,
Connor
I've spent a good chunk of my free time this summer writing a chess engine in C++ which can be found here: https://github.com/connormcmonigle/seer ... s/tag/v1.0. It should, hopefully, play at around 2900-3000 elo if AVX2 extensions are available (around 300-400 knps should be expected with AVX2). It uses a custom NNUE implementation with training code written in PyTorch and inference code relying on OpenMP SIMD for auto vectorization.
I'm not providing any compiles at this time, though compiling Seer should, with any luck, be straight forward (compilation instructions can be found in the readme). I'd love to see my engine appear on various rating lists if that's a possibility.
Thanks in advance and be sure to inform me if you encounter any issues,
Connor
Congratulations on the release!
It's awesome that you also released your training code and I like that you have separate testing code in your project. Looks great and I'm looking forward to playing around with it!
Static exchange evaluation has sadly made Weiss a bit slower but on the plus side compiling is now easier - just 'make' will likely be optimal for any machine. Also no longer requires 'isready' (or anything else) before 'go'
The SRL rating list of : Sun Oct 18 10:23:08 2020
EPD : epd\lc1.epd
Time : 1000ms
Max Time Hash
Engine Points Used Time Found Pos Elo Score Score ms Mb Cpu Errors
35 Weiss 1.2 293517 11:27:33.1 20047 40000 2935 400000 73.38% 1000 128 1 0
54 Weiss 1.0 284693 11:23:14.3 19241 40000 2846 400000 71.17% 1000 128 1 0
Cool....
90% of coding is debugging, the other 10% is writing bugs.