Module

BenchLib

Package: purescript-benchlib
Repository: m-bock/purescript-benchlib

A benchmark, slightly simplified, can be seen as this three step transformation:

prepare :: size -> input
benchmark :: input -> return
finalize :: return -> output

Common type variables:

inp: type of input for benchmark, can be diifferent for each benchmark in a group. It is created by the prepare function.
ret: type of result of benchmark, can be diifferent for each benchmark in a group.
out: type of output of the benchmark, must be equal for all benchmarks in a group. Created by the finalize function. Having this extra type variable e.g. allows to check if benchmark results of a group do return values considered equal.
m: the monad in which the benchmark is run. It must be an instance of MonadBench. Instances are provided for Effect and Aff.

#BenchResult Source

type BenchResult = { benchName :: String, duration :: Milliseconds, iterations :: Int, size :: Size }

The result of a benchmark.

#GroupResults Source

type GroupResults = { benchs :: Array BenchResult, groupName :: String }

The result of a benchmark group.

#SuiteResults Source

type SuiteResults = { groups :: Array GroupResults, suiteName :: String }

The result of a benchmark suite.

#CheckResults Source

type CheckResults = { results :: Array { benchName :: String, output :: String }, success :: Boolean }

#Reporter Source

type Reporter = { onBenchFinish :: BenchResult -> Effect Unit, onBenchStart :: { benchName :: String, size :: Size } -> Effect Unit, onCheckResults :: CheckResults -> Effect Unit, onGroupFinish :: GroupResults -> Effect Unit, onGroupStart :: String -> Effect Unit, onSizeFinish :: Size -> Effect Unit, onSizeStart :: Size -> Effect Unit, onSuiteFinish :: SuiteResults -> Effect Unit, onSuiteStart :: String -> Effect Unit }

A reporter is a set of functions that are called at different stages of the benchmark. It allows to customize the output of the benchmark suite or perform other actions like writing to a file.

#Size Source

type Size = Int

#benchSuite Source

benchSuite :: String -> (SuiteOpts -> SuiteOpts) -> Array Group -> Suite

Create a benchmark suite of a given name. The suite will be run with the provided options. The suite is a collection of groups, each containing multiple benchmarks.

#benchGroup Source

benchGroup :: forall @out. Eq out => Show out => String -> (GroupOpts -> GroupOpts) -> Array (Bench out) -> Group

Create a benchmark group of a given name. The group will be run with the provided options. The group is a collection of benchmarks

#bench Source

bench :: forall inp ret out. Eq out => String -> (BenchOptsPure Unit out out -> BenchOptsPure inp ret out) -> (inp -> ret) -> Bench out

Create a benchmark of a given name. The benchmark will be run with the provided options. The benchmark is a function that takes an input and returns an output. The input is generated by the prepare function, and the output is processed by the finalize function.

#benchM Source

benchM :: forall m inp ret out. MonadBench m => Eq out => String -> (BenchOptsM m Unit out out -> BenchOptsM m inp ret out) -> (inp -> m ret) -> Bench out

Like `bench``, but with a monadic function.

#benchSuite_ Source

benchSuite_ :: String -> Array Group -> Suite

Like benchSuite, but with default options.

#benchGroup_ Source

benchGroup_ :: forall @out. Eq out => Show out => String -> Array (Bench out) -> Group

Like benchGroup, but with default options.

#bench_ Source

bench_ :: forall out. Eq out => String -> (Unit -> out) -> Bench out

Like bench, but with default options.

#benchM_ Source

benchM_ :: forall @m out. Eq out => MonadBench m => String -> (Unit -> m out) -> Bench out

Like benchM, but with default options.

#checkEq Source

checkEq :: forall a. Eq a => Array a -> Boolean

Check if all elements in the array are equal. Useful for checking the results of benchmarks.

#MonadBench Source

class MonadBench :: (Type -> Type) -> Constraintclass (Monad m) <= MonadBench m  where

A class for monadic benchmarks. It allows to run benchmarks in different monads as long as they are capable of getting turned into an Aff monad.

Members

toAff :: forall a. m a -> Aff a

Instances

MonadBench Effect
MonadBench Aff

#defaultReporter Source

defaultReporter :: Reporter

Default reporter useful for selective overriding. It will do nothing.

#reportConsole Source

reportConsole :: Reporter

Console reporter. It will print the results to the console in human readable format.

#run Source

run :: Suite -> Effect Unit

Run the benchmark suite.

#eval Source

eval :: Suite -> Aff SuiteResults

Evaluate the benchmark suite.

Modules: BenchLib; BenchLib.Reporters.Html; BenchLib.Reporters.Json