A simple Gleam benchmark runner which wraps Benchee for the heavy lifting.
Named after Gleam, Benchee and their fruity Lychee offspring:
Imaginary Glychees- Requires Gleam 1.0 or later.
- For benchmarking on target JavaScript see https://hex.pm/packages/gleamy_bench, as Glychee only allows benchmarking on target Erlang.
- Glychee is dependency free except for Benchee and Elixir.
- A recent Elixir and Hex must be installed. You might be required to run
mix local.hex
after installing Elixir.
-
Add Glychee to your project:
gleam add glychee --dev
. -
Create a custom benchmarking module for example named
my_benchmark
that contains amain
-function. In that module you will define whichFunction
s to benchmark with one or manyData
. -
Run the benchmark:
gleam clean && \ gleam build && \ gleam run -m my_benchmark
If you do not have a Gleam project yet, create it with:
gleam new foobar
cd foobar
To add and run a demo of Glychee:
-
gleam add glychee --dev
-
In your project create a file named
src/my_benchmark.gleam
with following source code:import gleam/int import gleam/list import glychee/benchmark import glychee/configuration pub fn main() { // Configuration is optional configuration.initialize() configuration.set_pair(configuration.Warmup, 2) configuration.set_pair(configuration.Parallel, 2) // Run the benchmarks benchmark.run( [ benchmark.Function(label: "list.sort()", callable: fn(test_data) { fn() { list.sort(test_data, int.compare) } }), ], [ benchmark.Data(label: "pre-sorted list", data: list.range(1, 100_000)), benchmark.Data( label: "reversed list", data: list.range(1, 100_000) |> list.reverse, ), ], ) }
-
Then run in your terminal via:
gleam clean && \ gleam build && \ gleam run -m my_benchmark
Now you can alter the functions and data specified in above's example to whichever function of your application or library you want to benchmark.
Note that you can benchmark multiple functions with different data sets in one go.
Glychee's documentation can be found at https://hexdocs.pm/glychee.