Losseau, Tanguy
[UCL]
Van Roy, Peter
[UCL]
Though Erlang is a robust language offering an efficient concurrent environment, it comes short in not having a high performance for native computations being executed in a virtual machine (BEAM). Even so, Erlang offers the possibility to write Native Integrated Functions (NIFs), which provide an interface to functions written in C and in result extends the language with lower level operations. This master thesis goal is to write matrix operations for Erlang in NIFs. Considering a small set of matrix operations, its implementations in pure Erlang and NIFs are compared. The NIF versions are observed to run from five to seventeen times faster then their Erlang equivalents.
Bibliographic reference |
Losseau, Tanguy. Numerl : efficient vector and matrix computation for Erlang. Ecole polytechnique de Louvain, Université catholique de Louvain, 2022. Prom. : Van Roy, Peter. |
Permanent URL |
http://hdl.handle.net/2078.1/thesis:33858 |