https://hgpu.org/?p=24200
Exploring FPGA Optimizations to Compute Sparse Numerical Linear Algebra Kernels