This work is supported by Anaconda, Inc. and the Data Driven Discovery Initiative from the Moore Foundation.

This is part three of my series on scalable machine learning.

You can download a notebook of this post [here][notebook].

In part one, I talked about the type of constraints that push us to parallelize or distribute a machine learning workload. Today, we’ll be talking about the second constraint, “I’m constrained by time, and would like to fit more models at once, by using all the cores of my laptop, or all the machines in my cluster”.

An Aside on Parallelism

In the case of Python, we have two main avenues of parallelization (which we’ll roughly define as using multiple “workers” to do some “work” in less time). Those two avenues are

  1. multi-threading
  2. multi-processing

For python, the most important differences are that

  1. multi-threaded code can potentially be limited by the GIL
  2. multi-processing code requires that data be serialized between processes

The GIL is the “Global Interpreter Lock”, an implementation detail of CPython that means only one thread in your python process can be executing python code at once.

This talk by Python core-developer Raymond Hettinger does a good job summarizing things for Python, with an important caveat: much of what he says about the GIL doesn’t apply to the scientific python stack. NumPy, scikit-learn, and much of pandas release the GIL and can run multi-threaded, using shared memory and so avoiding serialization costs. I’ll highlight his quote, which summarizes the situation:

Your weakness is your strength, and your strength is your weakness

The strength of threads is shared state. The weakness of threads is shared state.

Another wrinkle here is that when you move to a distributed cluster, you have to have multiple processes. And communication between processes becomes even more expensive since you’ll have network overhead to worry about, in addition to the serialization costs.

Fortunately, modules like concurrent.futures and libraries like dask make it easy to swap one mode in for another. Let’s make a little dask array:

import dask.array as da
import dask
import dask.threaded
import dask.multiprocessing

X = da.random.uniform(size=(10000, 10), chunks=(1000, 10))
result = X / (X.T @ X).sum(1)

We can swap out the scheduler with a context-manager:

with dask.set_options(get=dask.threaded.get):
    # threaded is the default for dask.array anyway
with dask.set_options(get=dask.multiprocessing.get):

Every dask collection (dask.array, dask.dataframe, dask.bag) has a default scheduler that typically works well for the kinds of operations it does. For dask.array and dask.dataframe, the shared-memory threaded scheduler is used.

Cost Models

In this talk, Simon Peyton Jones talks about parallel and distributed computing for Haskell. He stressed repeatedly that there’s no silver bullet when it comes to parallelism. The type of parallelism appropriate for a web server, say, may be different than the type of parallelism appropriate for a machine learning algorithm.

I mention all this, since we’re about to talk about parallel machine learning. In general, for small data and many models you’ll want to use the threaded scheduler. For bigger data (larger than memory), you’ll want want to use the distributed scheduler. Assuming the underlying NumPy, SciPy, scikit-learn, or pandas operation releases the GIL, you’ll be able to get nice speedups without the cost of serialization. But again, there isn’t a silver bullet here, and the best type of parallelism will depend on your particular problem.

Where to Parallelize

In a typical machine-learning workflow, there are typically ample opportunities for parallelism.

  1. Over Hyper-parameters (one fit per combination of parameters)
  2. Over Cross-validation folds (one fit per fold)
  3. Within an algorithm (for some algorithms)

Scikit-learn already uses parallelism in many places, anywhere you see an n_jobs keyword.