accelerating-dnn-RISCV-Gemmini

Quantifying performance and efficiency gains of hardware acceleration for representative DNN workloads (matrix multiplication and 2D convolution) and to analyze the system level interactions