asv.commands.compare

Classes

Compare

Helper class that provides a standard way to create an ABC using

Functions

mean(values)

unroll_result(benchmark_name, params, *values)

Iterate through parameterized result values

_isna(value)

_is_result_better(a, b, a_ss, b_ss, factor[, use_stats])

Check if result 'a' is better than 'b' by the given factor,

Module Contents

asv.commands.compare.mean(values)[source]
asv.commands.compare.unroll_result(benchmark_name, params, *values)[source]

Iterate through parameterized result values

Yields

name

Strings of the form “benchmark_name(value1, value2)” with parameter values substituted in. For non-parameterized results, simply the benchmark name.

value

Benchmark timing or other scalar value.

asv.commands.compare._isna(value)[source]
asv.commands.compare._is_result_better(a, b, a_ss, b_ss, factor, use_stats=True)[source]

Check if result ‘a’ is better than ‘b’ by the given factor, possibly taking confidence intervals into account.

class asv.commands.compare.Compare[source]

Helper class that provides a standard way to create an ABC using inheritance.

classmethod setup_arguments(subparsers)[source]

Method to add the parser and arguments of the command.

In most cases, subparser.add_parser(cmd_name) should be called to add the command (where cmd_name is for example run or continuous). And then call add_argument to the returned parser, to add the command arguments.

classmethod run_from_conf_args(conf, args)[source]

Call the run method (i.e. cls.run) with the right arguments from conf and args. In most cases this is just something like clf.run(conf=conf, my_arg_1=args.my_arg_1, my_arg_2=args.my_arg_2, …).

classmethod run(conf, hash_1, hash_2, factor=None, split=False, only_changed=False, sort='default', machine=None, env_spec=None, use_stats=True)[source]
classmethod print_table(conf, hash_1, hash_2, factor, split, resultset_1=None, resultset_2=None, machine=None, only_changed=False, sort='default', use_stats=True, env_names=None, commit_names=None)[source]