Simulating CUDA/OpenCL on CPU, with Auto-vectorization and Speedup against Serial Scalar Code A small tool for writing various algorithms as if they were CUDA/OpenCL kernels Click here to read the article