Exploring high bandwidth memory for PET Image Reconstruction
Authors: Dai Yang, Tilman Küstner, Rami Al-Rihawi, Martin Schulz
Memory bandwidth plays an important role in high performance computing. Its impact on system performance is evident when running applications with a low arithmetic intensity. In this paper, we present our optimizations for the Maximum Likelihood Expectation Maximization (MLEM) algorithm, a method for positron emission tomography (PET) image reconstruction, with a sparse matrix-vector (SpMV) kernel. The results show significant improvement in performance when executing the code on an Intel Xeon Phi processor with multi-channel DRAM. The latency of the MCDRAM becomes the limiting factor. After implementing cache-blocking optimization, we achieved a total memory bandwidth of up to 180 GB/s for the SpMV operation.