Lam/mpi-specific function to enable run-time tracing
#include <mpi.h> int MPIL_Trace_on(void)
These functions give the application the flexibility to generate traces only during certain interesting phases of the application's execution. This technique can considerably reduce the size of the trace files and burden of displaying them.
Both functions are collective over the MPI_COMM_WORLD communicator. In typical usage, the -toff option of mpirun (1) would be used to enable tracing, but start with the runtime switch in the off position. At the beginning of an interesting phase of program execution, MPIL_Trace_on would be called. MPIL_Trace_off would be called after the interesting phase. Tracing can be turned on and off many times. Each period of tracing eventually forms a trace segment in the trace file extracted by lamtrace (1). If the on/off functions are never used and tracing is enabled with the -ton option of mpirun (1), a single trace segment is produced.
The on/off functions have no effect if tracing is not enabled by mpirun (1) with either the -ton or -toff switches. Thus, an application can be littered with these functions but run without trace collection and very little additional overhead due to the no-operation function calls.
This is a LAM/MPI-specific function and is intended mainly for debugging. If this function is used, it should be used in conjunction with the LAM_MPI C preprocessor macro
#if LAM_MPI MPIL_Trace_on(); #endif
After the volume of generated traces exceeds a preset limit, the oldest traces are discarded in favour of new traces. Avoiding discarded traces is further incentive to use MPIL_Trace_on and MPIL_Trace_off .
All MPI routines in Fortran (except for MPI_WTIME and MPI_WTICK ) have an additional argument ierr at the end of the argument list. ierr is an integer and has the same meaning as the return value of the routine in C. In Fortran, MPI routines are subroutines, and are invoked with the call statement.
All MPI objects (e.g., MPI_Datatype , MPI_Comm ) are of type INTEGER in Fortran.
If an error occurs in an MPI function, the current MPI error handler is called to handle it. By default, this error handler aborts the MPI job. The error handler may be changed with MPI_Errhandler_set ; the predefined error handler MPI_ERRORS_RETURN may be used to cause error values to be returned (in C and Fortran; this error handler is less useful in with the C++ MPI bindings. The predefined error handler MPI::ERRORS_THROW_EXCEPTIONS should be used in C++ if the error value needs to be recovered). Note that MPI does not guarantee that an MPI program can continue past an error.
All MPI routines (except MPI_Wtime and MPI_Wtick ) return an error value; C routines as the value of the function and Fortran routines in the last argument. The C++ bindings for MPI do not return error values; instead, error values are communicated by throwing exceptions of type MPI::Exception (but not by default). Exceptions are only thrown if the error value is not MPI::SUCCESS .
Note that if the MPI::ERRORS_RETURN handler is set in C++, while MPI functions will return upon an error, there will be no way to recover what the actual error value was.
- No error; MPI routine completed successfully.
- Other error; use MPI_Error_string to get more information about this error code.