P.E. Strazdins, A Comparison of Lookahead and Algorithmic Blocking Techniques for Parallel Matrix Factorization, International Journal of Parallel and Distributed Systems and Networks, 4(1), Jun 2001, ACTA Press Calgary, pages 26-35.
Results are given on the Fujitsu AP1000 and AP+ multicomputers, which have relatively high communication to computation to speeds. The results indicate that both methods are superior to storage blocking (without lookahead). They also indicate that for such machines, the hybrid method is optimal for smaller matrices, due to savings in communication startups. For larger matrices, algorithmic blocking gave the best performance, due to its better load balancing properties. An exception was LLT for the AP+, where lookahead alone gave comparable or better performance for larger matrix sizes as well. Performance models, predicting the minimum matrix size where lookahead becomes effective, indicate this trend can be expected for machines with lower communication to computation speeds, but that the range for where lookahead is superior is extended.