https://hgpu.org/?p=28104
Task parallelism-based architectures on FPGA to optimize the energy efficiency of AI at the edge