16-input dmodulation_decimation library part now works in non-vectorized and avx2 modes. Other modes cause a kernel panic. models have a new parameter, "vectorization" that can be set to avx512, avx2, sse3 or none (default).
16-input dmodulation_decimation library part now works in non-vectorized and avx2 modes. Other modes cause a kernel panic. models have a new parameter, "vectorization" that can be set to avx512, avx2, sse3 or none (default).