View Source fprof - The File Trace Profiler

fprof is a profiling tool that can be used to get a picture of how much processing time different functions consumes and in which processes.

fprof uses tracing with timestamps to collect profiling data. Therefore there is no need for special compilation of any module to be profiled.

fprof presents wall clock times from the host machine OS, with the assumption that OS scheduling will randomly load the profiled functions in a fair way. Both own time, that is, the time used by a function for its own execution, and accumulated time, that is, execution time including called functions.

Profiling is essentially done in 3 steps:

  • Tracing to a file.

  • Profiling: the trace file is read and raw profile data is collected into an internal RAM storage on the node. During this step the trace data may be dumped in text format to file or console.

  • Analysing: the raw profile data is sorted and dumped in text format either to file or console.

Since fprof stores trace data to a file, the runtime performance degradation is minimized, but still far from negligible, especially for programs that themselves use the filesystem heavily. Where the trace file is placed is also important; on Unix systems /tmp is usually a good choice, while any network-mounted disk is a bad choice.

fprof can also skip the file step and trace to a tracer process of its own that does the profiling in runtime.

The following sections show some examples of how to profile with fprof.

Profiling from the source code

If you can edit and recompile the source code, it is convenient to insert fprof:trace(start) and fprof:trace(stop) before and after the code to be profiled. All spawned processes are also traced. If you want some other filename than the default, use fprof:trace(start, "my_fprof.trace").

When execution is finished, the raw profile can be processed using fprof:profile(), or fprof:profile(file, "my_fprof.trace") for a non-default filename.

Finally create an informative table dumped on the console with fprof:analyse(), or on file with fprof:analyse(dest, []), or fprof:analyse([{dest, "my_fprof.analysis"}, {cols, 120}]) for a wider listing of a non-default filename.

Profiling a function

If you have one function that does the task that you want to profile, and the function returns when the profiling should stop, it is convenient to use fprof:apply(Module, Function, Args) for the tracing step.

If the tracing should continue after the function has returned, for example if it is a start function that spawns processes to be profiled, use fprof:apply(M, F, Args, [continue | OtherOpts]). The tracing has to be stopped at a suitable later time using fprof:trace(stop).

Immediate profiling

It is also possible to trace immediately into the profiling process that creates the raw profile data, that is to short circuit the tracing and profiling steps so that the filesystem is not used for tracing.

Do something like this:

{ok, Tracer} = fprof:profile(start),
fprof:trace([start, {tracer, Tracer}]),
%% Run code to profile
fprof:trace(stop);

This puts less load on the filesystem, but much more load on the Erlang runtime system.