Compiled libraries v1.2.1/1.2.0

Precompiled CPU and GPU libraries necessary for the NIRFASTerFF package.

Please download the file(s) appropriate to your system (os and Python version) and extract all the files into the nirfasterff/lib folder.
If you have a CUDA-enabled device, it is necessary to download both the CPU and GPU libraries. If not, it suffices to download only the CPU library.

v1.2.1 fixed a bug in the Python code, but the binaries remain the same. Please make sure to clone the latest Python code from the repo and double-check the version number

v1.2.0 fixed a few bugs in the Python codes, and the GPU codes are recompiled with CUDA 12.9 for a wider range of GPU support. -arch=all-major option was used in compilation (hence the very large files), so the codes are expected to be also future compatible. A compiled version of Jonathan Shewchuk's Triangle is included in the CPU library package for creating 2D meshes. Triangle was also included in both NIRFAST and NIRFASTer, but v1.2.0 is the first NIRFASTerFF to make use of it.

Changes in v1.2.0

• Support for Python 3.13 and Python 3.8 (NeuroDOT compatibility)
• Support for new GPUs (e.g. the 50xx series)
• A few bugs fixed
• 2D meshing added