High-Performance Low-Memory Lowering: GEMM-based Algorithms for DNN Convolution | IEEE Conference Publication | IEEE Xplore