Would be nice to be able to not only runn benchmarks and compare results with baseline, but also to be able to define 2 implementations at the same time and compare them. Eg
-module(my_kv).
-export(...).
maps_update(input) ->
maps:from_list([{K, K} || K <- lists:seq(1, 5000)]).
bench_maps_update(Map) ->
Map#{2500 => -1}.
dict_update(input) ->
dict:from_list([{K, K} || K <- lists:seq(1, 5000)]).
bench_dict_update(Dict) ->
dict:store(2500, -1, Dict).
$ rebar3 bench cmp my_kv:maps_update my_kv:dict_update
And this command will run 2 benchmarks and do all the statistical computations and comparision and will tell which implementation is faster.
Would be nice to be able to not only runn benchmarks and compare results with baseline, but also to be able to define 2 implementations at the same time and compare them. Eg
And this command will run 2 benchmarks and do all the statistical computations and comparision and will tell which implementation is faster.