Profiling for Ray Developers

This document details, for Ray developers, how to use pprof to profile Ray binaries.


These instructions are for Ubuntu only. Attempts to get pprof to correctly symbolize on Mac OS have failed.

sudo apt-get install google-perftools libgoogle-perftools-dev

Launching the to-profile binary

If you want to launch Ray in profiling mode, define the following variables:

export RAYLET_PERFTOOLS_PATH=/usr/lib/x86_64-linux-gnu/
export RAYLET_PERFTOOLS_LOGFILE=/tmp/pprof.out

The file /tmp/pprof.out will be empty until you let the binary run the target workload for a while and then kill it via ray stop or by letting the driver exit.

Visualizing the CPU profile

The output of pprof can be visualized in many ways. Here we output it as a zoomable .svg image displaying the call graph annotated with hot paths.

# Use the appropriate path.

google-pprof -svg $RAYLET /tmp/pprof.out > /tmp/pprof.svg
# Then open the .svg file with Chrome.

# If you realize the call graph is too large, use -focus=<some function> to zoom
# into subtrees.
google-pprof -focus=epoll_wait -svg $RAYLET /tmp/pprof.out > /tmp/pprof.svg

Here’s a snapshot of an example svg output, taken from the official documentation:

Running Microbenchmarks

To run a set of single-node Ray microbenchmarks, use:

ray microbenchmark

The following are the results for the 0.7.5 release on a Python 3 / a m4.16xl instance:

single core get calls per second 12169.8 +- 386.41
single core put calls per second 3117.45 +- 94.17
single core put gigabytes per second 11.32 +- 3.4
multi core put calls per second 16221.06 +- 895.13
multi core put gigabytes per second 24.14 +- 0.29
single core tasks sync per second 887.77 +- 3.69
single core tasks async per second 4524.45 +- 196.39
multi core tasks async per second 6963.49 +- 161.31
single core actor calls sync per second 762.4 +- 56.47
single core actor calls async per second 1030.44 +- 45.42
multi core actor calls async per second 6065.92 +- 175.05