Journal
Recent Posts
Notes on GPU programming, local AI, and distributed systems.
hpcgoptimization
HPCG: 3. Sparse Matrix Formats and Memory Access Optimization
Understanding sparse matrix formats with simple python code. By using ELLPACK format for representing sparse matrices, we can optimize the memory access pattern of the HPCG benchmark.
hpcgoptimization
HPCG: 2. Multi-coloring for Parallelization
In the reference implementation of HPCG, SYMGS computation cannot be fully parallelized. However, we can use multi-coloring technique to parallelize the computation at the cost of relaxation of the algorithm.
hpcgoptimization
HPCG: 1. The Specification
HPCG is a benchmark like LINPACK but with a different focus. Conceptually, it emphasizes sparse linear algebra, and practically, it is aimed at optimizing memory access patterns, communication, and parallelization.