Rust implementation of dense binary matrices and vectors.
Includes a SIMD implementation of a binary matrix.
- Arithmetic:
- Implement the rest of the basic arithmetic between matrices and vectors
- Faster matrix–vector multiplication using bits directly
- Faster matrix–matrix multiplication using bits directly
- Basic determinant calculation
- Extract out basic reduced row echelon form into own method
- Right multiplication of matrix by vector
- Kernel:
- Use Lanczos algorithm
- Transpose:
- Use rotates
- Switch to SIMD
- SIMD: Use
portable_simd
- Investigate using aarch64 assembly
- Investigate using x86-64 assembly
- Implement row-centric matrix as well
- Sparse matrix support?