3 papers
hep-lat2022
Performance optimizations for porting the openQD package to GPUs
Roman Gruber, Anton Kozhevnikov, Marina Krstić Marinković +2
OpenQD code has been used by the RC collaboration for the generation of fully dynamical QCD+QED gauge configurations with C boundary conditions. In this tal…
cs.DC2021
On the Parallel I/O Optimality of Linear Algebra Kernels: Near-Optimal Matrix Factorizations
Grzegorz Kwasniewski, Marko Kabić, Tal Ben-Nun +8
Matrix factorizations are among the most important building blocks of scientific computing. State-of-the-art libraries, however, are not communication-optimal, underutilizing curre…
cs.DC2021
COSTA: Communication-Optimal Shuffle and Transpose Algorithm with Process Relabeling
Marko Kabić, Simon Pintarelli, Anton Kozhevnikov +1
Communication-avoiding algorithms for Linear Algebra have become increasingly popular, in particular for distributed memory architectures. In practice, these algorithms assume that…