Tomasz Sobczyk

GCC fails to keep v2df_t variables in register and writes them back onto the stack often. I got 40% performance improvement by using __m128 directly where possible (i7-920). Clang doesn't...

[WIP] Attempt to optimize the position skipping to fit a distribution.

1

The old algorithm work online and therefore might discard excessive amounts of positions. The new algorithm filters the buffers that are filled in each worker thread (currently each buffer is...

[WIP] Optional local MSYS2 setup.

2

Potentially fixes #189.

Since the file `training_finished` doesn't contain epoch information it will consider the training finished when it's not. Also fools the exit after finished timer.

Detect mismatching trainer/player and exit (unless some explicit flag specified)

[easy_train] Potential issue with msys2 and PATH

13

`...msys2\mingw64\bin` needs to be in PATH, but that might not be setup by default, in which case we need to modify the environment programatically

https://github.com/glinscott/nnue-pytorch/blob/7d47cad764a95d4ad69f1b14a4a0722f1db63911/lib/nnue_training_data_formats.h#L264 revoof on discord suggested: > nthSetBitIndex could be _tzcnt_u64(_pdep_u64(1ULL

[WIP] Visualization for narrow neuron set activation.

1

Run like `python visualize_narrow_preactivation.py --data=d10_10000.bin --count=1000000 nn-63376713ba63.nnue --features= HalfKAv2_hm`

Add a script for finding most important [FT] weights in a net.

15

This script provides a way to find most important (currently only for the feature transformer) weights in a given network under given dataset. The importance is determined by taking a...