Skip to main content

A memory profiler for data batch processing applications.

Project description

The Fil memory profiler for Python

Your code reads some data, processes it, and uses too much memory. In order to reduce memory usage, you need to learn what code is responsible, and specifically what code is responsible for peak memory usage.

And that's exactly what Fil will help you find. Fil an open source memory profiler designed for data processing applications written in Python, and includes native support for Jupyter.

At the moment it only runs on Linux and macOS.

For more information, including an example of the output, see https://pythonspeed.com/products/filmemoryprofiler/

Installation

Assuming you're on macOS or Linux, and are using Python 3.6 or later, you can use either Conda or pip (or any tool that is pip-compatible and can install manylinux2010 wheels).

Conda

To install on Conda:

$ conda install -c conda-forge filprofiler

Pip

To install the latest version of Fil you'll need Pip 19 or newer. You can check like this:

$ pip --version
pip 19.3.0

If you're using something older than v19, you can upgrade by doing:

$ pip install --upgrade pip

If that doesn't work, try running that in a virtualenv.

Assuming you have a new enough version of pip:

$ pip install filprofiler

Using Fil

Measuring peak (high-water mark) memory usage in Jupyter

To measure memory usage of some code in Jupyter you need to do three things:

  1. Use an alternative kernel, "Python 3 with Fil". You can choose this kernel when you create a new notebook, or you can switch an existing notebook in the Kernel menu; there should be a "Change Kernel" option in there in both Jupyter Notebook and JupyterLab.
  2. Load the extension by doing %load_ext filprofiler.
  3. Add the %%filprofile magic to the top of the cell with the code you wish to profile.

Screenshot of JupyterLab

Measuring peak (high-water mark) memory usage for Python scripts

Instead of doing:

$ python yourscript.py --input-file=yourfile

Just do:

$ fil-profile run yourscript.py --input-file=yourfile

And it will generate a report.

Debugging out-of-memory crashes

First, run free to figure out how much memory is available—in this case about 6.3GB—and then set a corresponding limit on virtual memory with ulimit:

$ free -h
       total   used   free  shared  buff/cache  available
Mem:   7.7Gi  1.1Gi  6.3Gi    50Mi       334Mi      6.3Gi
Swap:  3.9Gi  3.0Gi  871Mi
$ ulimit -Sv 6300000

Then, run your program under Fil, and it will generate a SVG at the point in time when memory runs out:

$ fil-profile run oom.py 
...
=fil-profile= Wrote memory usage flamegraph to fil-result/2020-06-15T12:37:13.033/out-of-memory.svg

Reducing memory usage in your code

You've found where memory usage is coming from—now what?

If you're using data processing or scientific computing libraries, I have written a relevant guide to reducing memory usage.

What Fil tracks

Fil will track memory allocated by:

  • Normal Python code.
  • C code using malloc()/calloc()/realloc()/posix_memalign().
  • C++ code using new (including via aligned_alloc()).
  • Anonymous mmap()s.
  • Fortran 90 explicitly allocated memory (tested with gcc's gfortran).

Still not supported, but planned:

  • mremap() (resizing of mmap()).
  • File-backed mmap(). The usage here is inconsistent since the OS can swap it in or out, so probably supporting this will involve a different kind of resource usage.
  • Other forms of shared memory, need to investigate if any of them allow sufficient allocation.
  • Anonymous mmap()s created via /dev/zero (not common, since it's not cross-platform, e.g. macOS doesn't support this).
  • memfd_create().
  • Possibly memalign, valloc(), pvalloc(), reallocarray().

License

Copyright 2020 Hyphenated Enterprises LLC

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

 http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for filprofiler, version 0.10.0
Filename, size File type Python version Upload date Hashes
Filename, size filprofiler-0.10.0-cp36-cp36m-macosx_10_14_x86_64.whl (271.7 kB) File type Wheel Python version cp36 Upload date Hashes View
Filename, size filprofiler-0.10.0-cp36-cp36m-manylinux2010_x86_64.whl (1.8 MB) File type Wheel Python version cp36 Upload date Hashes View
Filename, size filprofiler-0.10.0-cp37-cp37m-macosx_10_14_x86_64.whl (271.7 kB) File type Wheel Python version cp37 Upload date Hashes View
Filename, size filprofiler-0.10.0-cp37-cp37m-manylinux2010_x86_64.whl (1.8 MB) File type Wheel Python version cp37 Upload date Hashes View
Filename, size filprofiler-0.10.0-cp38-cp38-macosx_10_14_x86_64.whl (271.7 kB) File type Wheel Python version cp38 Upload date Hashes View
Filename, size filprofiler-0.10.0-cp38-cp38-manylinux2010_x86_64.whl (1.8 MB) File type Wheel Python version cp38 Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page