mlos_bench ========== .. py:module:: mlos_bench .. autoapi-nested-parse:: mlos_bench is a framework to help automate benchmarking and OS/application parameter autotuning and the data management of the results. .. contents:: Table of Contents :depth: 3 Overview ++++++++ ``mlos_bench`` can be installed from `pypi <https://pypi.org/project/mlos-bench>`_ via ``pip install mlos-bench`` and executed using the ``mlos_bench`` `command <../../mlos_bench.run.usage.html>`_ using a collection of `json` `configs <https://github.com/microsoft/MLOS/tree/main/mlos_bench/mlos_bench/config/>`_ which are managed and documented in the :py:mod:`~mlos_bench.config` module and elsewhere in this package documentation. It is intended to be used with :py:mod:`mlos_core` via :py:class:`~mlos_bench.optimizers.mlos_core_optimizer.MlosCoreOptimizer` to help navigate complex parameter spaces more efficiently, though other :py:mod:`~mlos_bench.optimizers` are also available to help customize the search process easily by simply swapping out the :py:class:`~mlos_bench.optimizers.base_optimizer.Optimizer` class in the associated json configs. For instance, :py:class:`~mlos_bench.optimizers.grid_search_optimizer.GridSearchOptimizer` can be used to perform a grid search over the parameter space instead, or no tuning at all and ``mlos_bench`` can be used as only a repeatable benchmarking tool. Goals ^^^^^ The overall goal of the MLOS project is to enable *reproducible* and *trackable* benchmarking and *efficient* autotuning for systems software. In this, automation of the benchmarking process is a key component that ``mlos_bench`` seeks to enable. Interaction ^^^^^^^^^^^ Users are expected to provide JSON :py:mod:`mlos_bench.config` s that instruct the framework how to automate their benchmark, autotuning, or other Experiment. This may involve several steps such as 1. deploying a VM 2. installing some software 3. loading a dataset 4. running a benchmark 5. collecting and storing the results 6. repeating for statistical and noise measures (we call each iteration a ``Trial``) 7. (optionally) repeating with a different configuration (e.g., for autotuning purposes) 8. analyzing the results Since many of these phases are common across different benchmarks, the framework is intended to be modular and composable to allow reuse and faster development of new benchmarking environments or autotuning experiments. Where possible, the framework will provide common configs for reuse (e.g., deploying a VM on Azure, or run `benchbase <https://github.com/cmu-db/benchbase>`_ against a database system) to allow users to focus on the specifics of their experiments. Where none are currently available, one can create them external to MLOS, however users are also encouraged to `submit PRs or Issues <https://github.com/microsoft/MLOS/CONTRIBUTING.md>`_ to add new classes or config and script snippets for others to use as well. For more details on the configuration files, please see the documentation in the :py:mod:`mlos_bench.config` module. Classes Overview ++++++++++++++++ The other core classes in this package are: - :py:mod:`~mlos_bench.environments` which provide abstractions for representing an execution environment. These are generally the target of the optimization process and are used to evaluate the performance of a given configuration, though can also be used to simply run a single benchmark. They can be used, for instance, to provision a :py:mod:`VM <mlos_bench.environments.remote.vm_env>`, run benchmarks or execute any other arbitrary code on a :py:mod:`remote machine <mlos_bench.environments.remote.remote_env>`, and many other things. - Environments are often associated with :py:mod:`~mlos_bench.tunables` which provide a language for specifying the set of configuration parameters that can be optimized or searched over with the :py:mod:`~mlos_bench.optimizers`. - :py:mod:`~mlos_bench.services` provide the necessary abstractions to run interact with the :py:mod:`~mlos_bench.environments` in different settings. For instance, the :py:class:`~mlos_bench.services.remote.azure.azure_vm_services.AzureVMService` can be used to run commands on Azure VMs for a remote :py:mod:`~mlos_bench.environments.remote.vm_env.VMEnv`. Alternatively, one could swap out that service for :py:class:`~mlos_bench.services.remote.ssh.ssh_host_service.SshHostService` in order to target a different VM without having to change the :py:class:`~mlos_bench.environments.base_environment.Environment` configuration at all since they both implement the same :py:class:`~mlos_bench.services.types.remote_exec_type.SupportsRemoteExec` :py:mod:`Services type<mlos_bench.services.types>` interfaces. This is particularly useful when running the same benchmark on different ecosystems and makes the configs more modular and composable. - :py:mod:`~mlos_bench.storage` which provides abstractions for storing and retrieving data from the experiments. For instance, nearly any :py:mod:`SQL <mlos_bench.storage.sql>` backend that `sqlalchemy <https://www.sqlalchemy.org>`_ supports can be used. The data management and automation portions of experiment data is a key component of MLOS as it provides a unified way to manage experiment data across different Environments, enabling more reusable visualization and analysis by mapping benchmark metrics into common semantic types (e.g., via `OpenTelemetry <https://opentelemetry.io>`_). Without this most experiments are effectively siloed and require custom, and more critically, non-reusable scripts to setup and later parse results and are hence harder to scale to many users. With these features as a part of the MLOS ecosystem, benchmarking can become a *service* that any developer, admin, researcher, etc. can use and adapt. See below for more information on the classes in this package. .. rubric:: Notes Note that while the docstrings in this package are generated from the source code and hence sometimes more focused on the implementation details, we do try to provide some (testable) examples here, but most user interactions with the package will be through the `json configs <https://github.com/microsoft/MLOS/tree/main/mlos_bench/mlos_bench/config/>`_. Even so it may be useful to look at the documentation here (perhaps especially starting with :py:mod:`mlos_bench.config`) and, if necessary, the `source code <https://github.com/microsoft/MLOS>`_ to understand how those are interpreted. .. rubric:: Examples Here is an example that shows how to run a simple benchmark using the command line. The entry point for these configs can be found `here <https://github.com/microsoft/MLOS/blob/main/mlos_bench/mlos_bench/tests/config/cli/test-cli-local-env-bench.jsonc>`_. >>> from subprocess import run >>> # Note: we show the command wrapped in python here for testing purposes. >>> # Alternatively replace test-cli-local-env-bench.jsonc with >>> # test-cli-local-env-opt.jsonc for one that does an optimization loop. >>> cmd = "mlos_bench \ ... --config mlos_bench/mlos_bench/tests/config/cli/test-cli-local-env-bench.jsonc \ ... --globals experiment_test_local.jsonc \ ... --tunable_values tunable-values/tunable-values-local.jsonc" >>> print(f"Here's the shell command you'd actually run:\n# {cmd}") Here's the shell command you'd actually run: # mlos_bench --config mlos_bench/mlos_bench/tests/config/cli/test-cli-local-env-bench.jsonc --globals experiment_test_local.jsonc --tunable_values tunable-values/tunable-values-local.jsonc >>> # Now we run the command and check the output. >>> result = run(cmd, shell=True, capture_output=True, text=True, check=True) >>> assert result.returncode == 0 >>> lines = result.stderr.splitlines() >>> first_line = lines[0] >>> last_line = lines[-1] >>> expected = "INFO Launch: mlos_bench" >>> assert first_line.endswith(expected) >>> expected = "INFO Final score: {'score': 123.4, 'total_time': 123.4, 'throughput': 1234567.0}" >>> assert last_line.endswith(expected) .. rubric:: Notes - `mlos_bench/README.md <https://github.com/microsoft/MLOS/tree/main/mlos_bench/>`_ for additional documentation and examples in the source tree. - `mlos_bench/DEVNOTES.md <https://github.com/microsoft/MLOS/tree/main/mlos_bench/mlos_bench/DEVNOTES.md>`_ for additional developer notes in the source tree. - There is also a working example of using ``mlos_bench`` in a *separate config repo* (the more expected case for most users) in the `sqlite-autotuning <https://github.com/Microsoft-CISL/sqlite-autotuning>`_ repo. Submodules ---------- .. toctree:: :maxdepth: 1 /autoapi/mlos_bench/config/index /autoapi/mlos_bench/dict_templater/index /autoapi/mlos_bench/environments/index /autoapi/mlos_bench/event_loop_context/index /autoapi/mlos_bench/launcher/index /autoapi/mlos_bench/optimizers/index /autoapi/mlos_bench/os_environ/index /autoapi/mlos_bench/run/index /autoapi/mlos_bench/schedulers/index /autoapi/mlos_bench/services/index /autoapi/mlos_bench/storage/index /autoapi/mlos_bench/tunables/index /autoapi/mlos_bench/util/index /autoapi/mlos_bench/version/index Attributes ---------- .. autoapisummary:: mlos_bench.__version__ Package Contents ---------------- .. py:data:: __version__ :value: '0.6.2'