nightly benchmark run configureation

scripts/ray_distributed_train.py

@@ -15,6 +15,15 @@ def _has_flag(tokens: list[str], name: str) -> bool:
     return any(tok == name or tok.startswith(f"{name}=") for tok in tokens)
 
 
+def _entry_tokens(run_kind: str, entry_args: str) -> list[str]:
+    tokens = shlex.split(entry_args)
+    if run_kind == "benchmark" and not (
+        _has_flag(tokens, "--run-kind") or _has_flag(tokens, "--run-mode")
+    ):
+        return ["--run-kind", "benchmark", *tokens]
+    return tokens
+
+
 def _alive_node_ips() -> list[str]:
     seen: set[str] = set()
     ips: list[str] = []
@@ -33,13 +42,18 @@ def _alive_node_ips() -> list[str]:
 def _train_on_node(
     *,
     root: str,
-    train_args: str,
+    run_kind: str,
+    entry_args: str,
     rank: int,
     world_size: int,
     coordinator_ip: str,
     coordinator_port: int,
     base_seed: int,
     run_group: str,
+    compare_robust: bool,
+    output_root: str,
+    wandb_entity: str,
+    wandb_project: str,
     sync_jax: bool,
 ) -> int:
     env = dict(os.environ)
@@ -53,6 +67,12 @@ def _train_on_node(
     env["JAX_PLATFORMS"] = requested_platform
     # Keep each train process in single-host mode to avoid accidental global stalls.
     env["CLOUD_TPU_TASK_ID"] = "0"
+    if run_kind == "benchmark":
+        env["PHANTOM_BENCHMARK_COMPARE_ROBUST"] = "1" if compare_robust else "0"
+    if wandb_entity:
+        env["WANDB_ENTITY"] = wandb_entity
+    if wandb_project:
+        env["WANDB_PROJECT"] = wandb_project
 
     cwd = str(Path(root))
 
@@ -75,30 +95,62 @@ def _train_on_node(
             [sys.executable, "-c", probe], cwd=cwd, env=env_probe, check=True
         )
 
-    tokens = shlex.split(train_args)
+    tokens = _entry_tokens(run_kind, entry_args)
+    seed = int(base_seed + rank)
     if not _has_flag(tokens, "--seed"):
-        tokens.extend(["--seed", str(base_seed + rank)])
-    if not _has_flag(tokens, "--group"):
+        tokens.extend(["--seed", str(seed)])
+
+    if run_kind == "train" and not _has_flag(tokens, "--group"):
         tokens.extend(["--group", run_group])
 
+    if (
+        run_kind == "benchmark"
+        and output_root
+        and not _has_flag(tokens, "--output-dir")
+    ):
+        out_dir = Path(output_root) / f"rank_{rank}" / f"seed_{seed}"
+        out_dir.parent.mkdir(parents=True, exist_ok=True)
+        tokens.extend(["--output-dir", str(out_dir)])
+
     cmd = [sys.executable, "-m", "engine.train", *tokens]
+    print(
+        {
+            "rank": int(rank),
+            "run_kind": run_kind,
+            "seed": int(seed),
+            "compare_robust": bool(compare_robust),
+            "wandb_entity": str(env.get("WANDB_ENTITY", "")),
+            "wandb_project": str(env.get("WANDB_PROJECT", "")),
+            "command": " ".join(cmd),
+        }
+    )
     proc = subprocess.run(cmd, cwd=cwd, env=env)
     return int(proc.returncode)
 
 
 def main() -> None:
     parser = argparse.ArgumentParser(
-        description="Launch one train run per Ray TPU node"
+        description="Launch one train/benchmark run per Ray TPU node"
     )
-    parser.add_argument("--train-args", type=str, required=True)
+    parser.add_argument("--run-kind", choices=["train", "benchmark"], default="train")
+    parser.add_argument("--entry-args", type=str, default="")
+    parser.add_argument("--train-args", type=str, default="")
     parser.add_argument("--num-nodes", type=int, default=0)
     parser.add_argument("--tpu-per-task", type=float, default=8.0)
     parser.add_argument("--base-seed", type=int, default=42)
     parser.add_argument("--sync-jax", action="store_true")
     parser.add_argument("--coordinator-port", type=int, default=12355)
     parser.add_argument("--run-group", type=str, default="")
+    parser.add_argument("--compare-robust", action="store_true")
+    parser.add_argument("--output-root", type=str, default="")
+    parser.add_argument("--wandb-entity", type=str, default="")
+    parser.add_argument("--wandb-project", type=str, default="")
     args = parser.parse_args()
 
+    entry_args = str(args.entry_args or args.train_args).strip()
+    if not entry_args:
+        raise ValueError("--entry-args (or legacy --train-args) is required")
+
     ray.init(address="auto")
 
     node_ips = _alive_node_ips()
@@ -118,8 +170,11 @@ def main() -> None:
             "nodes": node_ips,
             "world_size": world_size,
             "coordinator": f"{coordinator_ip}:{int(args.coordinator_port)}",
-            "train_args": args.train_args,
+            "run_kind": str(args.run_kind),
+            "entry_args": entry_args,
             "run_group": run_group,
+            "compare_robust": bool(args.compare_robust),
+            "output_root": str(args.output_root),
         }
     )
 
@@ -130,14 +185,19 @@ def main() -> None:
         futures.append(
             _train_on_node.options(resources=resources).remote(
                 root=root,
-                train_args=args.train_args,
+                run_kind=str(args.run_kind),
+                entry_args=entry_args,
                 rank=rank,
                 world_size=world_size,
                 coordinator_ip=coordinator_ip,
                 coordinator_port=int(args.coordinator_port),
                 base_seed=int(args.base_seed),
                 run_group=run_group,
-                sync_jax=bool(args.sync_jax),
+                compare_robust=bool(args.compare_robust),
+                output_root=str(args.output_root),
+                wandb_entity=str(args.wandb_entity),
+                wandb_project=str(args.wandb_project),
+                sync_jax=bool(args.sync_jax and str(args.run_kind) == "train"),
             )
         )