micro-benchmark

This egg provides means to benchmark your code in a simple manner. It features traditional benchmarking with microsecond resolution as well as a different approach that has been inspired by: https://github.com/evanphx/benchmark-ips

Documentation

Types

real : (or integer float ratnum)statistics-alist : (list-of (pair symbol *))

Usage

(import micro-benchmark)

current-benchmark-iterations

[parameter] current-benchmark-iterations

The default number of iterations. Initially 100.

current-benchmark-statistics-set

[parameter] current-benchmark-statistics-set

The set of generated statistics, 'verbose or 'normal.

benchmark-measure

[syntax] (benchmark-measure EXPR ...) -> real

Runs the EXPR ... once & returns the runtime in nanoseconds.

benchmark-run

(benchmark-run [ITERATIONS] EXPR) -> statistics-alist(benchmark-run ([ITERATIONS [STAT-ID [VAR]]]) EXPR ...) -> statistics-alist

Runs the EXPR ... ITERATIONS times and returns a micro-stats (generate-statistics) report.

ITERATIONS: fixnum | true ; defaults to (current-benchmark-iterations)
STAT-ID: symbol | true ; defaults to (current-benchmark-statistics-set)
VAR: identifier ; variable to bind the current interation index

benchmark-ips

(benchmark-ips [SECONDS] EXPR) -> statistics-alist(benchmark-ips ([SECONDS [WARMUPS [STAT-ID]]]) EXPR ...) -> statistics-alist

Determines how many times one can run the given EXPR ... per second and returns a (generate-statistics) report.

SECONDS: fixnum ; seconds of run while counting, defaults to 5
WARMUPS: fixnum ; seconds of run before counting, defaults to 2
STAT-ID: symbol ; statistics set generated

benchmark-measure-run

[syntax] (benchmark-measure-run ([ITERATIONS [VAR]]) EXPR ...) -> (list-of float)

Returns the ITERATIONS X (EXPR ...) runtimes, in nanoseconds.

ITERATIONS: fixnum | true ; defaults to (current-benchmark-iterations)
VAR: identifier ; variable to bind the current interation index

run-benchmark-measure

[procedure] (run-benchmark-measure THUNK) -> real

Returns the (THUNK) runtime in nanoseconds.

benchmark-measure-ips

(benchmark-measure-ips (begin EXPR ...)) -> (list-of float)(benchmark-measure-ips ([SECONDS [WARMUPS]]) EXPR ...) -> (list-of float)

Determines how many times one can run the given EXPR ... per second and returns a list of the iterations/second.

SECONDS: real ; seconds of run while counting, defaults to 5
WARMUPS: real ; seconds of run before counting, defaults to 2

run-benchmark-ips

[procedure] (run-benchmark-ips THUNK [SECONDS [WARMUPS]]) -> (list-of float)

Determines how many times one can run the given (THUNK) per second and returns a list of the iterations/second.

SECONDS: real ; seconds of run while counting, defaults to 5
WARMUPS: real ; seconds of run before counting, defaults to 2

current-nanoseconds

Note How unreliable time is handled is a build choice, from reliable-time, eventual-time, or warning-time. See the micro-benchmark.egg file in the source distribution.

Usage

(import current-nanoseconds)

current-flnanoseconds

[syntax] (current-flnanoseconds) -> float

current-nanoseconds

[procedure] (current-nanoseconds) -> integer

current-microseconds

Usage

(import current-microseconds)

current-flmicroseconds

[syntax] (current-flmicroseconds) -> float

current-microseconds

[procedure] (current-microseconds) -> integer

Examples

(import micro-benchmark)

;YMMV

;; simply measure the runtime of the given fragment
(benchmark-measure (sleep 2))
;=> 2003822.22601318

;; run code 3 times and return results
(parameterize ((current-benchmark-iterations 3))
  (benchmark-run () (sleep 1)) )
;=> ((min . 1001227.66900635) (max . 1003503.40200806) ...)

;; find out how many iterations we can make per second
(benchmark-ips () (sleep 2))
;=> ((mean . 0.499209372899189) ...)

(import micro-benchmark slib-charplot)
(parameterize ((plot-dimensions '(20 60)))
  ;suggest (vector-map ns->ms (benchmark-me...)) 1st
  (histograph (benchmark-measure-run (40) (sleep 1))))

Requirements

micro-stats

Authors

David Krentzlin Kon Lovett

Repository

This egg is hosted on the CHICKEN Subversion repository:

https://anonymous@code.call-cc.org/svn/chicken-eggs/release/5/micro-benchmark

If you want to check out the source code repository of this egg and you are not familiar with Subversion, see this page.

Version history

0.3.6: Fix benchmark-measure-run.
0.3.5: C6 release.

License

This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or (at
your option) any later version.

This program is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
General Public License for more details.

A full copy of the GPL license can be found at
<http://www.gnu.org/licenses/>.