count - number of requests in array (integer)
request_array - array of requests to test (array of handles)
out_count - number of completed requests (integer)
index_array - array of indexes of completed requests (array of integer)
status_array - statuses of completed requests (array of structures)
The MPI_Testsome() function is used to test if some of the requests in an array are completed. This check is done in a non-blocking manner. If a request is done, then its index in the request_array will be stored inside the first available slot in index_array . The total number of completed requests is stored in out_count , and the first out_count elements of the index_array and status_array will be valid.
If a request is a receive, then the original tag and source of a recieved message are stored in the MPI_TAG and MPI_SOURCE fields of the required element of status_array respectively. If the received message has been truncated then MPI_ERROR will be set to be MPI_ERR_TRUNCATE. One additional field holds the total size of the message in bytes. If either a send or receive is successfully cancelled, then MPI_ERROR will contain MPI_ERR_CANCEL. Otherwise, it will be set to MPI_ERR_SUCCESS. If any MPI_ERROR field is not MPI_SUCCESS then this function will return MPI_ERR_IN_STATUS. This will not invoke an error handler. If status_array is MPI_STATUSES_IGNORE then this information will not be stored.
If the size ( count ) of the request array is zero, then out_count will be set to MPI_UNDEFINED and MPI_SUCCESS will be returned. PMPI_Testsome() is the profiling version of this function.
The C++ version of this function passess the out_count as the return value.
All MPI routines except for MPI_Wtime and MPI_Wtick return an error code. The the current MPI error handler is invoked if the return value is not MPI_SUCCESS. The default error handler aborts, but this may be changed with by using the MPI_Errhandler_set() function. The predefined error handler MPI_ERRORS_RETURN may be used to cause error values to be returned instead. Note that MPI does not guarentee that an MPI program can continue past an error. In this implementation, all errors except MPI_ERR_INTERN or MPI_ERR_OTHER should always be recoverable.
In C, the error code is passed as the return value. In FORTRAN, all functions have an parameter ierr which returns the error code. MPI C++ functions do not directly return an error code. However, C++ users may want to use the MPI::ERRORS_THROW_EXCEPTIONS handler. This will throw an MPI::Exception with the corresponding error code. To prevent exceptions from being raised from within C and Fortran code, they will see all error return values as MPI_ERR_PENDING when this handler is chosen. In this implementation, call MPI::throw_exception() to throw the correct exception if this occurs.
MPI_SUCCESS - No error;
MPI_ERR_PENDING - Pending exception;
MPI_COUNT - Invalid array size;
MPI_ERR_ARG - Invalid pointer;
MPI_ERR_REQUEST - Invalid request;
MPI_ERR_IN_STATUS - MPI_ERROR is set in a status structure. An error handler is not called.
Company Info |
Product Index |
Category Index |
Copyright © Lockless Inc All Rights Reserved.