def driver_0(redis_address, driver_index):
"""The script for driver 0.
This driver should create five actors that each use one GPU and some actors
that use no GPUs. After a while, it should exit.
"""
ray.init(redis_address=redis_address)
# Wait for all the nodes to join the cluster.
_wait_for_nodes_to_join(total_num_nodes)
# Start some long running task. Driver 2 will make sure the worker running
# this task has been killed.
for i in range(num_long_running_tasks_per_driver):
long_running_task.remote(driver_index, i, redis_address)
# Create some actors that require one GPU.
actors_one_gpu = [Actor1.remote(driver_index, i, redis_address)
for i in range(5)]
# Create some actors that don't require any GPUs.
actors_no_gpus = [Actor0.remote(driver_index, 5 + i, redis_address)
for i in range(5)]
for _ in range(1000):
ray.get([actor.check_ids.remote() for actor in actors_one_gpu])
ray.get([actor.check_ids.remote() for actor in actors_no_gpus])
# Start a long-running method on one actor and make sure this doesn't
# affect anything.
actors_no_gpus[0].long_running_method.remote()
_broadcast_event("DRIVER_0_DONE", redis_address)
def driver_0(redis_address, driver_index):
"""The script for driver 0.
This driver should create five actors that each use one GPU and some actors
that use no GPUs. After a while, it should exit.
"""
ray.init(redis_address=redis_address)
# Wait for all the nodes to join the cluster.
_wait_for_nodes_to_join(total_num_nodes)
# Start some long running task. Driver 2 will make sure the worker running
# this task has been killed.
for i in range(num_long_running_tasks_per_driver):
long_running_task.remote(driver_index, i, redis_address)
# Create some actors that require one GPU.
actors_one_gpu = [
Actor1.remote(driver_index, i, redis_address) for i in range(5)
]
# Create some actors that don't require any GPUs.
actors_no_gpus = [
Actor0.remote(driver_index, 5 + i, redis_address) for i in range(5)
]
for _ in range(1000):
ray.get([actor.check_ids.remote() for actor in actors_one_gpu])
ray.get([actor.check_ids.remote() for actor in actors_no_gpus])
# Start a long-running method on one actor and make sure this doesn't
# affect anything.
actors_no_gpus[0].long_running_method.remote()
_broadcast_event("DRIVER_0_DONE", redis_address)
def driver(redis_address, driver_index):
"""The script for driver 0.
This driver should create five actors that each use one GPU and some actors
that use no GPUs. After a while, it should exit.
"""
ray.init(redis_address=redis_address)
# Wait for all the nodes to join the cluster.
_wait_for_nodes_to_join(total_num_nodes)
# Limit the number of drivers running concurrently.
for i in range(driver_index - max_concurrent_drivers + 1):
_wait_for_event("DRIVER_{}_DONE".format(i), redis_address)
def try_to_create_actor(actor_class, timeout=100):
# Try to create an actor, but allow failures while we wait for the
# monitor to release the resources for the removed drivers.
start_time = time.time()
while time.time() - start_time < timeout:
try:
actor = actor_class.remote()
except Exception as e:
time.sleep(0.1)
else:
return actor
# If we are here, then we timed out while looping.
raise Exception("Timed out while trying to create actor.")
# Create some actors that require one GPU.
actors_one_gpu = []
for _ in range(num_gpus_per_driver):
actors_one_gpu.append(try_to_create_actor(Actor1))
for _ in range(100):
ray.get([actor.check_ids.remote() for actor in actors_one_gpu])
_broadcast_event("DRIVER_{}_DONE".format(driver_index), redis_address)
def driver(redis_address, driver_index):
"""The script for all drivers.
This driver should create five actors that each use one GPU. After a while,
it should exit.
"""
ray.init(redis_address=redis_address)
# Wait for all the nodes to join the cluster.
_wait_for_nodes_to_join(total_num_nodes)
# Limit the number of drivers running concurrently.
for i in range(driver_index - max_concurrent_drivers + 1):
_wait_for_event("DRIVER_{}_DONE".format(i), redis_address)
def try_to_create_actor(actor_class, timeout=500):
# Try to create an actor, but allow failures while we wait for the
# monitor to release the resources for the removed drivers.
start_time = time.time()
while time.time() - start_time < timeout:
try:
actor = actor_class.remote()
except Exception:
time.sleep(0.1)
else:
return actor
# If we are here, then we timed out while looping.
raise Exception("Timed out while trying to create actor.")
# Create some actors that require one GPU.
actors_one_gpu = []
for _ in range(num_gpus_per_driver):
actors_one_gpu.append(try_to_create_actor(Actor1))
for _ in range(100):
ray.get([actor.check_ids.remote() for actor in actors_one_gpu])
_broadcast_event("DRIVER_{}_DONE".format(driver_index), redis_address)
