In this assigment you will study the relative performance of regular strided
data with MPI_Send and MPI_Recv. Send a vector of 1000 elements of type
MPI_DOUBLE, with a stride of 24 between each element.
Test these three forms:
- Use MPI_Type_vector
- Use MPI_Type_struct to form a "striding" structure
- Use MPI_DOUBLE and a loop to pack and unpack the vector your self (that
is, do not use the MPI Datatypes)
Time these in the same way you timed memcpy
; do several
interations to get a good average, and repeat the test 10 times.