Starting from the observation that quite frequently the memory system is the performance bottleneck, I will discuss a typical memory efficient algorithm, namely the classical one for dense matrix multiplication. Then I will briefly sketch how similar ideas are helpful in the setting of Sparse Grids (I’ll explain what that is too), and report on experiments we did with this memory efficient algorithm.
Thursday, Feb 18, 2016, 4:30pm-5:30pm in POST 126
Bio: Dr. Jacob is Associate Professor for Algorithm Engineering in the Theoretical Computer Science Section of IT University of Copenhagen, where he is a member of the Algorithms Group.
In 1997 Dr. Jacob received his Diploma in Computer Science in Warzburg (Germany) and in 2002 his PhD from Aarhus University (Denmark). He has worked as a researcher at Los Alamos National Lab (USA), LMU Munich (Germany), TU Munich (Germany), and ETH Zurich (Switzerland).