initial_commit

launch_bench_multinode.py

@@ -0,0 +1,79 @@
+import os
+import subprocess
+from datetime import datetime
+
+
+def load_template(template_path: str):
+    output_template = ""
+    with open(template_path, "r") as handle:
+        output_template = handle.read()
+    return output_template
+
+
+def write_batch(batch_fpath: str, batch_content: str):
+    with open(batch_fpath, "w") as handle:
+        _ = handle.write(batch_content)
+
+
+collectives = ["Reduce",
+               # "Reduce_scatter",
+               # "Reduce_scatter_block",
+               # "Allreduce",
+               # "Allgather",
+               # "Allgatherv",
+               # "Scatter",
+               # "Scatterv",
+               # "Gather",
+               # "Gatherv",
+               # "Alltoall",
+               # "Bcast",
+               # "Barrier"
+               ]
+
+procnt = [
+    18,
+    # 36,
+    # 54,
+    # 72
+]
+mpi1_bin = "/home/hpc/ihpc/ihpc136h/workspace/prototyping/bin"
+slurm_template = load_template("templates/bench.template")
+
+template_parameter = {"time_stamp": datetime.now().strftime("%y_%m_%d_%H-%M-%S"),
+                      "job_name": "",
+                      "output_dir": os.getcwd()+"/output/",
+                      "err_dir": os.getcwd()+"/error/",
+                      "data_dir": os.getcwd()+"/data/",
+                      "n_procs": 18,
+                      "off_mem_flag": "",
+                      "bin": mpi1_bin
+                      }
+
+output_dir = os.getcwd()+"/output/"
+err_dir = os.getcwd()+"/error/"
+batch_files_dir = os.getcwd()+"/batchs/"
+data_dir = os.getcwd()+"/data/"
+
+if os.path.isdir(output_dir) == False:
+    os.mkdir(output_dir)
+if os.path.isdir(err_dir) == False:
+    os.mkdir(err_dir)
+if os.path.isdir(data_dir) == False:
+    os.mkdir(data_dir)
+if os.path.isdir(batch_files_dir) == False:
+    os.mkdir(batch_files_dir)
+
+log = ""
+
+for n_procs in procnt:
+    template_parameter["n_procs"] = n_procs
+    for collective in collectives:
+        template_parameter["job_name"] = collective
+        write_batch(batch_files_dir+collective+".sh",
+                    slurm_template.format(**template_parameter))
+        result = subprocess.run(["sbatch", batch_files_dir+collective+".sh"],
+                                capture_output=True, text=True)
+        log += f"#{collective} {n_procs}" + "\n"
+        log += "\tSTDOUT:" + result.stdout + "\n"
+        log += "\tSTDERR:" + result.stderr + "\n"
+print(log)