Add MPI-IO timings

Author: gjbex

source-code/mpi4py/README.md

@@ -27,7 +27,8 @@ C/C++ or Fortran implementations.
 1. `run_pi_mpipool.sh`: Bash script to run `pi_mpipool.py`.
 1. `file_trafficker.py`: file write/read test application that can run
    serially, multi-threaded, multi-process and MPI.
-
+1. `mpi_io.py`: timing of MPI-IO operations.
+1. `translate_bin.py`: translate binary to ASCII data.
 
 ## How to run?
 

source-code/mpi4py/mpi_io.py

@@ -0,0 +1,60 @@
+#!/usr/bin/env python
+
+import argparse
+from mpi4py import MPI
+import numpy as np
+import re
+import time
+
+
+def compute_size(size_str):
+    if (match := re.match(r'\s*(\d+)\s*([a-z]+)?', size_str, re.IGNORECASE)) is not None:
+        size = int(match.group(1))
+        if match.group(2):
+            units = match.group(2).lower()
+            if units == 'b':
+                pass
+            elif units == 'kb':
+                size *= 1024
+            elif units == 'mb':
+                size *= 1024**2
+            elif units == 'gb':
+                size *= 1024**3
+            elif units == 'tb':
+                size *= 1024**4
+            else:
+                raise ValueError(f"invalid unit in size expression: '{size_str}'")
+        return size
+    else:
+        raise ValueError(f"invalid size expression: '{size_str}'")
+
+
+if __name__ == '__main__':
+    arg_parser = argparse.ArgumentParser(description='write a file using MPI-IO')
+    arg_parser.add_argument('--file', required=True,
+                            help='file name to write')
+    arg_parser.add_argument('--size', default='100mb',
+                            help='file size to write')
+    options = arg_parser.parse_args()
+
+    comm = MPI.COMM_WORLD
+    comm_rank = comm.Get_rank()
+    comm_size = comm.Get_size()
+    total_file_size = compute_size(options.size)//8
+    data_size = total_file_size//comm_size
+    offset = comm_rank*data_size*8
+
+    start_val = comm_rank*data_size
+    end_val = (1 + comm_rank)*data_size
+    if comm_rank == comm_size - 1:
+        end_val += total_file_size % comm_size
+    data = np.arange(start_val, end_val, 1.0)
+
+    mode = MPI.MODE_WRONLY | MPI.MODE_CREATE
+    start_time = time.time()
+    file = MPI.File.Open(comm, options.file, mode)
+    file.Write_at_all(offset, data)
+    file.Close()
+    end_time = time.time()
+    total_time = end_time - start_time
+    print(f'{comm_rank}, {total_time} s, {compute_size(options.size)/(total_time*2**20):.3f} MB/s') 

source-code/mpi4py/translate_bin.py

@@ -0,0 +1,13 @@
+#!/usr/bin/env python
+
+import argparse
+import numpy as np
+
+
+if __name__ == '__main__':
+    arg_parser = argparse.ArgumentParser(description="read and print MPI-IO data")
+    arg_parser.add_argument('file', help='file to read')
+    options = arg_parser.parse_args()
+    data = np.fromfile(options.file)
+    for value in data:
+        print(value)