1 paper
Aheli Poddar, Sanskar Prasad, Arindam Samanta +3
Deep learning inference and training performance depends critically on GPU kernel efficiency. Modern compilers such as PyTorch Inductor automatically generate GPU kernels from high…