API for profile
- ()
by Stuart Sierra
Full namespace name:
clojure.contrib.profile
Overview
Simple code profiling & timing measurement.
Wrap any section of code in the prof macro, giving it a name, like this:
(defn my-function [x y]
(let [sum (prof :addition (+ x y))
product (prof :multiplication (* x y))]
[sum product]))
The run your code in the profile macro, like this:
(profile (dotimes [i 10000] (my-function 3 4)))
Which prints a report for each named section of code:
Name mean min max count sum
addition 265 0 37000 10000 2655000
multiplication 274 0 53000 10000 2747000
Times are measured in nanoseconds, to the maximum precision available
under the JVM. See the function documentation for more details.
Public Variables and Functions
*enable-profiling*
var
Set this to false before loading/compiling to omit
profiling code.
Source
print-summary
function
Usage: (print-summary profile-summary)
Prints a table of the results returned by summarize.
Source
prof
macro
Usage: (prof name & body)
If *enable-profiling* is true, wraps body in profiling code.
Returns the result of body. Profile timings will be stored in
*profile-data* using name, which must be a keyword, as the key.
Timings are measured with System/nanoTime.
Source
profile
macro
Usage: (profile & body)
Runs body with profiling enabled, then prints a summary of
results. Returns nil.
Source
summarize
function
Usage: (summarize profile-data)
Takes the raw data returned by with-profile-data and returns a map
from names to summary statistics. Each value in the map will look
like:
{:mean ..., :min ..., :max ..., :count ..., :sum ...}
:mean, :min, and :max are how long the profiled section took to run,
in nanoseconds. :count is the total number of times the profiled
section was executed. :sum is the total amount of time spent in the
profiled section, in nanoseconds.
Source
with-profile-data
macro
Usage: (with-profile-data & body)
Executes body with *profile-data* bound to an atom of a new map.
Returns the raw profile data as a map. Keys in the map are profile
names (keywords), and values are lists of elapsed time, in
nanoseconds.
Source