The MATLAB Interface

Within MATLAB, NetSolve may be used in two ways. It is possible to call NetSolve in a blocking or nonblocking fashion. Here is an example of the MATLAB interface to solve an linear system computation using the blocking call:

       >> a = rand(100); b = rand(100,1);
       >> x = netsolve('ax=b',a,b)

This MATLAB script first creates a random matrix, a, and a vector b of length 100. The call to the netsolve function returns with the solution This call manages all the NetSolve protocol, and the computation may be executed on a remote host.

Here is the same computation performed in a nonblocking fashion:

       >> a = rand(100); b = rand(100,1);
       >> request = netsolve_nb('send','ax=b',a,b)
       >> x = netsolve_nb('probe',request)
            Not Ready Yet
       >> x = netsolve_nb('wait',request)

Here, the first call to netsolve_nb() sends a request to the NetSolve agent and returns immediately with a request identifier. One can then either probe for the request or wait for it. This approach allows user-level parallelism and communication/computation overlapping (see Section 7).

Other functions are provided, for example, to obtain informations on the problems available or on the status of the pending requests.