• License: BSD-3-Clause
• Home: https://github.com/AdhocMan/spla
• Development: https://github.com/eth-cscs/spla
• Documentation: https://spla.readthedocs.io/en/latest/?badge=latest
• 36141 total downloads
• Last upload: 3 months and 29 days ago

Description

SPLA provides specialized functions for distributed matrix multiplication, as required in some computational material science codes. It aims for maximum computation and communication overlap when possible and allows any combination of host and device pointers if compiled with GPU support. C++, C and Fortran interfaces are available.