Dask on Ray

Dask is a Python parallel computing library geared towards scaling analytics and scientific computing workloads. It provides big data collections that mimic the APIs of the familiar NumPy and Pandas libraries, allowing those abstractions to represent larger-than-memory data and/or allowing operations on that data to be run on a multi-machine cluster, while also providing automatic data parallelism, smart scheduling, and optimized operations. Operations on these collections create a task graph, which is executed by a scheduler.

Ray provides a scheduler for Dask (dask_on_ray) which allows you to build data analyses using Dask’s collections and execute the underlying tasks on a Ray cluster.

dask_on_ray uses Dask’s scheduler API, which allows you to specify any callable as the scheduler that you would like Dask to use to execute your workload. Using the Dask-on-Ray scheduler, the entire Dask ecosystem can be executed on top of Ray.

Scheduler

The Dask-on-Ray scheduler can execute any valid Dask graph, and can be used with any Dask .compute() call. Here’s an example:

import ray
from ray.util.dask import ray_dask_get
import dask
import dask.array as da
import dask.dataframe as dd
import numpy as np
import pandas as pd

# Start Ray.
# Tip: If you're connecting to an existing cluster, use ray.init(address="auto").
ray.init()

d_arr = da.from_array(np.random.randint(0, 1000, size=(256, 256)))

# The Dask scheduler submits the underlying task graph to Ray.
d_arr.mean().compute(scheduler=ray_dask_get)

# Set the scheduler to ray_dask_get in your config so you don't have to specify it on
# each compute call.
dask.config.set(scheduler=ray_dask_get)

df = dd.from_pandas(pd.DataFrame(
    np.random.randint(0, 100, size=(1024, 2)),
    columns=["age", "grade"]), npartitions=2)
df.groupby(["age"]).mean().compute()

Note

For execution on a Ray cluster, you should not use the Dask.distributed client; simply use plain Dask and its collections, and pass ray_dask_get to .compute() calls or set the scheduler in one of the other ways detailed here. Follow the instructions for using Ray on a cluster to modify the ray.init() call.

Why use Dask on Ray?

  1. To take advantage of Ray-specific features such as the

    launching cloud clusters and shared-memory store.

  2. If you’d like to use Dask and Ray libraries in the same application without having two different clusters.

  3. If you’d like to create data analyses using the familiar NumPy and Pandas APIs provided by Dask and execute them on a fast, fault-tolerant distributed task execution system geared towards production, like Ray.

Dask-on-Ray is an ongoing project and is not expected to achieve the same performance as using Ray directly. All Dask abstractions should run seamlessly on top of Ray using this scheduler, so if you find that one of these abstractions doesn’t run on Ray, please open an issue.

Best Practice for Large Scale workloads

For Ray 1.3, the default scheduling policy is to pack tasks to the same node as much as possible. It is more desirable to load balance tasks if you run a large scale / memory intensive Dask on Ray workloads.

In this case, there are two recommended setup. - Setting an internal config flag scheduler_loadbalance_spillback to change the scheduler to load balance tasks. - Setting the head node’s num-cpus to 0 so that tasks are not scheduled on a head node.

# Head node. Set `num_cpus=0` to avoid tasks are being scheduled on a head node.
RAY_SCHEDULER_LOADBALANCE_SPILLBACK=1 ray start --head --num-cpus=0

# Worker node.
RAY_SCHEDULER_LOADBALANCE_SPILLBACK=1 ray start --address=[head-node-address]

Out-of-Core Data Processing

Processing datasets larger than cluster memory is supported via Ray’s object spilling: if the in-memory object store is full, objects will be spilled to external storage (local disk by default). This feature is available but off by default in Ray 1.2, and is on by default in Ray 1.3+. Please see your Ray version’s object spilling documentation for steps to enable and/or configure object spilling.

Persist

Dask-on-Ray patches dask.persist() in order to match Dask Distributed’s persist semantics <https://distributed.dask.org/en/latest/manage-computation.html#client-persist>; namely, calling dask.persist() with a Dask-on-Ray scheduler will submit the tasks to the Ray cluster and return Ray futures inlined in the Dask collection. This is nice if you wish to compute some base collection (such as a Dask array), followed by multiple different downstream computations (such as aggregations): those downstream computations will be faster since that base collection computation was kicked off early and referenced by all downstream computations, often via shared memory.

import ray
from ray.util.dask import ray_dask_get
import dask
import dask.array as da
import numpy as np
# Start Ray.
# Tip: If you're connecting to an existing cluster, use ray.init(address="auto").
ray.init()

# Set the scheduler to ray_dask_get in your config so you don't have to specify it on
# each compute call.
dask.config.set(scheduler=ray_dask_get)

d_arr = da.ones(100)
print(dask.base.collections_to_dsk([d_arr]))
# {('ones-c345e6f8436ff9bcd68ddf25287d27f3',
#   0): (functools.partial(<function _broadcast_trick_inner at 0x7f27f1a71f80>,
#   dtype=dtype('float64')), (5,))}

# This submits all underlying Ray tasks to the cluster and returns a Dask array with
# the Ray futures inlined.
d_arr_p = d_arr.persist()

# Notice that the Ray ObjectRef is inlined. The dask.ones() task has been submitted
# to and is running on the Ray cluster.
dask.base.collections_to_dsk([d_arr_p])
# {('ones-c345e6f8436ff9bcd68ddf25287d27f3',
#   0): ObjectRef(8b4e50dc1ddac855ffffffffffffffffffffffff0100000001000000)}

# Future computations on this persisted Dask Array will be fast since we already
# started computing d_arr_p in the background.
d_arr_p.sum().compute()
d_arr_p.min().compute()
d_arr_p.max().compute()

Custom optimization for Dask DataFrame shuffling

Dask on Ray provides a Dask DataFrame optimizer that leverages Ray’s ability to execute multiple-return tasks in order to speed up shuffling by as much as 4x on Ray. Simply set the dataframe_optimize configuration option to our optimizer function, similar to how you specify the Dask-on-Ray scheduler:

import ray
from ray.util.dask import ray_dask_get, dataframe_optimize
import dask
import dask.dataframe as dd
import numpy as np
import pandas as pd

# Start Ray.
# Tip: If you're connecting to an existing cluster, use ray.init(address="auto").
ray.init()

# Set the scheduler to ray_dask_get, and set the Dask DataFrame optimizer to our
# custom optimization function, this time using the config setter as a context manager.
with dask.config.set(scheduler=ray_dask_get, dataframe_optimize=dataframe_optimize):
    npartitions = 100
    df = dd.from_pandas(pd.DataFrame(
        np.random.randint(0, 100, size=(10000, 2)),
        columns=["age", "grade"]), npartitions=npartitions)
    # We set max_branch to infinity in order to ensure that the task-based shuffle
    # happens in a single stage, which is required in order for our optimization to
    # work.
    df.set_index(
        ["age"], shuffle="tasks", max_branch=float("inf")).head(10, npartitions=-1)

Callbacks

Dask’s custom callback abstraction is extended with Ray-specific callbacks, allowing the user to hook into the Ray task submission and execution lifecycles. With these hooks, implementing Dask-level scheduler and task introspection, such as progress reporting, diagnostics, caching, etc., is simple.

Here’s an example that measures and logs the execution time of each task using the ray_pretask and ray_posttask hooks:

from ray.util.dask import RayDaskCallback
from timeit import default_timer as timer


class MyTimerCallback(RayDaskCallback):
   def _ray_pretask(self, key, object_refs):
      # Executed at the start of the Ray task.
      start_time = timer()
      return start_time

   def _ray_posttask(self, key, result, pre_state):
      # Executed at the end of the Ray task.
      execution_time = timer() - pre_state
      print(f"Execution time for task {key}: {execution_time}s")


with MyTimerCallback():
   # Any .compute() calls within this context will get MyTimerCallback()
   # as a Dask-Ray callback.
   z.compute(scheduler=ray_dask_get)

The following Ray-specific callbacks are provided:

  1. ray_presubmit(task, key, deps): Run before submitting a Ray task. If this callback returns a non-None value, a Ray task will _not_ be created and this value will be used as the would-be task’s result value.

  2. ray_postsubmit(task, key, deps, object_ref): Run after submitting a Ray task.

  3. ray_pretask(key, object_refs): Run before executing a Dask task within a Ray task. This executes after the task has been submitted, within a Ray worker. The return value of this task will be passed to the ray_posttask callback, if provided.

  4. ray_posttask(key, result, pre_state): Run after executing a Dask task within a Ray task. This executes within a Ray worker. This callback receives the return value of the ray_pretask callback, if provided.

  5. ray_postsubmit_all(object_refs, dsk): Run after all Ray tasks have been submitted.

  6. ray_finish(result): Run after all Ray tasks have finished executing and the final result has been returned.

See the docstring for RayDaskCallback.__init__() <ray.util.dask.callbacks.RayDaskCallback>.__init__() for further details about these callbacks, their arguments, and their return values.

When creating your own callbacks, you can use RayDaskCallback directly, passing the callback functions as constructor arguments:

def my_presubmit_cb(task, key, deps):
   print(f"About to submit task {key}!")

with RayDaskCallback(ray_presubmit=my_presubmit_cb):
   z.compute(scheduler=ray_dask_get)

or you can subclass it, implementing the callback methods that you need:

class MyPresubmitCallback(RayDaskCallback):
   def _ray_presubmit(self, task, key, deps):
      print(f"About to submit task {key}!")

with MyPresubmitCallback():
   z.compute(scheduler=ray_dask_get)

You can also specify multiple callbacks:

# The hooks for both MyTimerCallback and MyPresubmitCallback will be
# called.
with MyTimerCallback(), MyPresubmitCallback():
   z.compute(scheduler=ray_dask_get)

Combining Dask callbacks with an actor yields simple patterns for stateful data aggregation, such as capturing task execution statistics and caching results. Here is an example that does both, caching the result of a task if its execution time exceeds some user-defined threshold:

@ray.remote
class SimpleCacheActor:
   def __init__(self):
      self.cache = {}

   def get(self, key):
      # Raises KeyError if key isn't in cache.
      return self.cache[key]

   def put(self, key, value):
      self.cache[key] = value


class SimpleCacheCallback(RayDaskCallback):
   def __init__(self, cache_actor_handle, put_threshold=10):
      self.cache_actor = cache_actor_handle
      self.put_threshold = put_threshold

   def _ray_presubmit(self, task, key, deps):
      try:
         return ray.get(self.cache_actor.get.remote(str(key)))
      except KeyError:
         return None

   def _ray_pretask(self, key, object_refs):
      start_time = timer()
      return start_time

   def _ray_posttask(self, key, result, pre_state):
      execution_time = timer() - pre_state
      if execution_time > self.put_threshold:
         self.cache_actor.put.remote(str(key), result)


cache_actor = SimpleCacheActor.remote()
cache_callback = SimpleCacheCallback(cache_actor, put_threshold=2)
with cache_callback:
   z.compute(scheduler=ray_dask_get)

Note

The existing Dask scheduler callbacks (start, start_state, pretask, posttask, finish) are also available, which can be used to introspect the Dask task to Ray task conversion process, but note that the pretask and posttask hooks are executed before and after the Ray task is submitted, not executed, and that finish is executed after all Ray tasks have been submitted, not executed.

This callback API is currently unstable and subject to change.