3.8 Function netcellfun

Function File: netcellfun (connections, fun, …)

Evaluates function fun in a parallel cluster and collects results.

This function handles arguments and options equivalently to parcellfun and returnes equivalent output. Differently, the first argument specifies server machines for parallel remote execution, see pconnect for a description of the connections variable. A further difference is that the option "ChunksPerProc" is ignored and instead the chunk size can be specified directly with an option "ChunkSize".

This function can only be successfully called at the client machine of the parallel cluster. connections can contain all connections of the network, a subset of them, or a single connection. The local machine (client), if contained in connections, is ignored. However, one of the servers can run at the local machine under certain conditions (see pconnect) and will not be ignored in this case, so that the local machine can take part in parallel execution with netcellfun.

As a second level of parallelism, fun is executed at each server machine (using parcellfun or pararrayfun) by default in as many local processes in parallel as the server has processor cores available. The number of local parallel processes can be configured for each server with the "nlocaljobs" option (see network_set), a value of 0 means that the default value will be used, a value of 1 means that execution is not parallel within the server (but still parallel over the cluster).

See parallel_doc ("limitations") for possible limitations e.g. on how fun can be defined.

Cluster execution incurs a considerable overhead. A speedup is likely if the computation time of fun is long. To speed up execution of a large set of arguments with short computation times of fun, increase "ChunkSize", possibly use "Vectorize" (see pararrayfun), and possibly experiment with increasing "nlocaljobs" from the default.

See also: netarrayfun, pconnect, pserver, sclose, rfeval, install_vars.