Test

benchmarks/benchmark_tests.yaml

@@ -36,6 +36,6 @@
     compute_template: distributed.yaml
 
   run:
-    timeout: 3600 # 1hr
+    timeout: 7200 # 2hr
     prepare: sleep 0
     script: python distributed/test_distributed.py

benchmarks/distributed/config.yaml

@@ -38,6 +38,15 @@ available_node_types:
         resources:
           node: 1
         max_workers: 999999
+    tiny_worker_node:
+        node_config:
+            InstanceType: t3.large
+            ImageId: ami-0a2363a9cff180a64
+        resources:
+          node: 1
+          tiny: 1
+          CPU: 5
+        max_workers: 999999
 
 head_node_type: head_node
 

benchmarks/distributed/test_actors.py

@@ -0,0 +1,73 @@
+import ray
+import ray.autoscaler.sdk
+
+import json
+import os
+from time import sleep, perf_counter
+from tqdm import tqdm, trange
+
+TEST_NUM_NODES = 1000
+MAX_ACTORS_IN_CLUSTER = 20000
+
+
+def num_alive_nodes():
+    n = 0
+    for node in ray.nodes():
+        if node["Alive"]:
+            n += 1
+    return n
+
+
+def scale_to(target):
+    while num_alive_nodes() != target:
+        ray.autoscaler.sdk.request_resources(bundles=[{"node": 1}] * target)
+        print(f"Current # nodes: {num_alive_nodes()}, target: {target}")
+        print("Waiting ...")
+        sleep(5)
+
+
+def test_max_actors():
+    # TODO (Alex): Dynamically set this based on number of cores
+    cpus_per_actor = 0.24
+
+    @ray.remote(num_cpus=cpus_per_actor)
+    class Actor:
+        def foo(self):
+            pass
+
+    actors = [
+        Actor.options(num_cpus=0.24).remote()
+        for _ in trange(MAX_ACTORS_IN_CLUSTER, desc="Launching actors")
+    ]
+
+    for actor in tqdm(actors, desc="Ensuring actors have started"):
+        assert ray.get(actor.foo.remote()) is None
+
+
+ray.init(address="auto")
+
+scale_to(TEST_NUM_NODES)
+assert num_alive_nodes(
+) == TEST_NUM_NODES, "Wrong number of nodes in cluster " + len(ray.nodes())
+
+cluster_resources = ray.cluster_resources()
+
+available_resources = ray.available_resources()
+assert available_resources == cluster_resources, (
+    str(available_resources) + " != " + str(cluster_resources))
+
+actor_start = perf_counter()
+test_max_actors()
+actor_end = perf_counter()
+actor_time = actor_end - actor_start
+
+print(f"Actor time: {actor_time} ({MAX_ACTORS_IN_CLUSTER} actors)")
+
+if "TEST_OUTPUT_JSON" in os.environ:
+    out_file = open(os.environ["TEST_OUTPUT_JSON"], "w")
+    results = {
+        "actor_time": actor_time,
+        "num_actors": MAX_ACTORS_IN_CLUSTER,
+        "success": "1"
+    }
+    json.dump(results, out_file)

benchmarks/distributed/test_distributed.py

@@ -9,10 +9,10 @@
 from tqdm import tqdm, trange
 
 TEST_NUM_NODES = 65
-MAX_ACTORS_IN_CLUSTER = 10000
 MAX_RUNNING_TASKS_IN_CLUSTER = 10000
 MAX_PLACEMENT_GROUPS = 1000
-MAX_NUM_NODES = 250
+MAX_ACTORS_IN_CLUSTER = 50000
+MAX_NUM_NODES = 1000
 
 
 def num_alive_nodes():
@@ -144,7 +144,7 @@ def f3():
 
 scale_to(TEST_NUM_NODES)
 assert num_alive_nodes(
-) == TEST_NUM_NODES, "Wrong number of nodes in cluster " + len(ray.nodes())
+) == TEST_NUM_NODES, f"Wrong number of nodes in cluster {len(ray.nodes())}"
 
 cluster_resources = ray.cluster_resources()
 
@@ -153,24 +153,13 @@ def f3():
     str(available_resources) + " != " + str(cluster_resources))
 print("Done launching nodes")
 
-actor_start = perf_counter()
-test_max_actors()
-actor_end = perf_counter()
-
-sleep(1)
-assert num_alive_nodes(
-) == TEST_NUM_NODES, "Wrong number of nodes in cluster " + len(ray.nodes())
-assert available_resources == cluster_resources, (
-    str(available_resources) + " != " + str(cluster_resources))
-print("Done testing actors")
-
 task_start = perf_counter()
 test_max_running_tasks()
 task_end = perf_counter()
 
 sleep(1)
 assert num_alive_nodes(
-) == TEST_NUM_NODES, "Wrong number of nodes in cluster " + len(ray.nodes())
+) == TEST_NUM_NODES, f"Wrong number of nodes in cluster {len(ray.nodes())}"
 assert available_resources == cluster_resources, (
     str(available_resources) + " != " + str(cluster_resources))
 print("Done testing tasks")
@@ -181,7 +170,7 @@ def f3():
 
 sleep(1)
 assert num_alive_nodes(
-) == TEST_NUM_NODES, "Wrong number of nodes in cluster " + len(ray.nodes())
+) == TEST_NUM_NODES, f"Wrong number of nodes in cluster {len(ray.nodes())}"
 assert available_resources == cluster_resources, (
     str(available_resources) + " != " + str(cluster_resources))
 print("Done testing placement groups")
@@ -192,7 +181,16 @@ def f3():
 
 sleep(1)
 assert num_alive_nodes(
-) == MAX_NUM_NODES, "Wrong number of nodes in cluster " + len(ray.nodes())
+) == MAX_NUM_NODES, f"Wrong number of nodes in cluster {len(ray.nodes())}"
+
+actor_start = perf_counter()
+test_max_actors()
+actor_end = perf_counter()
+
+sleep(1)
+assert num_alive_nodes(
+) == MAX_NUM_NODES, f"Wrong number of nodes in cluster {len(ray.nodes())}"
+print("Done testing actors")
 print("Done.")
 
 actor_time = actor_end - actor_start