IntelĀ® MPI Benchmarks 4.0
This pattern tests the case when all participating processes perform concurrent I/O to different private files. This benchmark is particularly useful for the systems that allow completely independent I/O operations from different processes. The benchmark pattern is expected to show parallel scaling and obtain optimum results. See the basic definitions and a schematic view of the pattern below.
Property |
Description |
---|---|
Measured pattern |
As symbolized in figure I/O benchmarks, aggregation for output |
Elementary I/O action |
As symbolized in the figure below. In this figure, Nproc is the number of processes. |
MPI routines for the blocking mode |
MPI_File_write/MPI_File_read |
MPI routines for the nonblocking mode |
MPI_File_iwrite/MPI_File_iread |
etype |
MPI_BYTE |
File type |
MPI_BYTE |
MPI data type |
MPI_BYTE |
Reported timings |
Δt (in μsec), aggregate and non-aggregate for the Write flavor. |
Reported throughput |
x/Δt, aggregate and non-aggregate for the Write flavor |