UPDATE: ruth-distributed parameters

ruth/zeromq/README.md

@@ -40,7 +40,6 @@ distributed spawning of workers accross nodes.
 In order to run the simulation in distributed fashion we can use ```bench.py```, specifically function ```run```.
 For correct incorporation of nodes spawned and configuration this function has to be used inside ```ruth```, otherwise 
 we may simply use ```bench.py``` and edit it's parameters since it spawns the ```run``` with following parameters:
-
 ```
 WORK_DIR = Path(os.getcwd()).absolute()
 WORKER_DIR = WORK_DIR / str(sys.argv[1])
@@ -61,6 +60,10 @@ try_to_kill = False
 
 run(...)
 ```
+Or run the command:
+```
+ruth-distributed EXPERIMENT_NAME EVKIT_DIR_PATH --config-file="config.json" --workers=32
+```
 
 ## Submitting a job
 In order to submit a job to a cluster, for example SLURM, we may use:

ruth/zeromq/bench.py

@@ -1,4 +1,3 @@
-import itertools
 import logging
 import pandas as pd
 import time
@@ -7,16 +6,12 @@
 import subprocess
 import os
 import sys
-import signal
 
 from typing import List
 from pathlib import Path
-from collections import defaultdict
 from contextlib import closing
 from dataclasses import dataclass
 from cluster.cluster import Cluster, start_process
-from cluster import cluster
-from src.client import Client
 
 
 def get_pbs_nodes() -> List[str]:

ruth/zeromq/distributed.py

@@ -10,14 +10,15 @@
 
 
 @click.command()
-@click.argument("worker-dir", type=str)
-@click.argument("evkit-path", type=click.Path(exists=True))
+@click.argument("experiment-name", type=str)
+@click.argument("evkit-dir-path", type=click.Path(exists=True))
 @click.option("--config-file", type=click.Path(exists=True), help="Path to simulation config.", default="config.json")
 @click.option("--workers", type=int, default=32, help="Number of workers. Default 32.")
+@click.option("--spawn-workers-at-main-node", is_flag=True, help="Spawn workers at main node.")
 @click.option("--try-to-kill", is_flag=True, help="Try to kill workers after simulation is computed.")
-def distributed(worker_dir, evkit_path, config_file, workers, try_to_kill):
+def distributed(experiment_name, evkit_path, config_file, workers, spawn_workers_at_main_node, try_to_kill):
     work_dir = Path(os.getcwd()).absolute()
-    worker_dir = work_dir / worker_dir
+    worker_dir = work_dir / experiment_name
     env_path = os.environ["VIRTUAL_ENV"]
     modules = [
         "Python/3.10.8-GCCcore-12.2.0",
@@ -39,7 +40,8 @@ def distributed(worker_dir, evkit_path, config_file, workers, try_to_kill):
         EVKIT_PATH=evkit_path,
         MODULES=modules,
         ENV_PATH=env_path,
-        try_to_kill=try_to_kill
+        try_to_kill=try_to_kill,
+        spawn_workers_at_main_node=spawn_workers_at_main_node
     )
     # result = bench(nodes, WORKER_DIR)