Skip to main content

A Validation Orchestrator designed by and for software at scale.

Project description

Parallel Computing Validation System

PyPI version License Python versions Docs

Parallel Computing Validation System (PCVS) is a Validation Orchestrator designed by and for software at scale. Its primary target is HPC applications & runtimes but can flawlessly address smaller use cases. PCVS can help users to create their test scenarios and reuse them among multiples implementations, a high value when it comes to validating programming standards (like APIs & ABIs). No matter the number of programs, benchmarks, languages, or tech non-regression bases use, PCVS gathers in a single execution, and, with a focus on interacting with HPC batch managers efficiently, run jobs concurrently to reduce the time-to-result overall.

Quick installation guide

PCVS is a Python program distributed on PyPI. It can be easily installed using pip.

$ pip3 install pcvs

Alternatively, you can install and test it from source.

How to install from source
# Considering python3.10+
$ pip3 install .
# For dev/testing purposes, use the following to install development dependencies
$ pip3 install '.[dev]'
# Basic tests
$ tox -e pcvs-coverage
# OR
$ coverage run

Once installed, the program should be available as pcvs.

[!TIP] Some PCVS specific test-suite exists, PCVS-Benchmarks is a curated set of HPC benchmarks and test suite with PCVS test descriptions

Basic usage

[!NOTE] You can add autocompletion for PCVS commands to your favorite shell

# ZSH support
$ eval "$(_PCVS_COMPLETE=zsh_source pcvs)"
# BASH support
$ eval "$(_PCVS_COMPLETE=bash_source pcvs)"

Profiles and test description

PCVS relies on two YAML-based sets of configurations to understand the environment and the suite.

  • The profiles describe the testing environment (compilers, machine, criterions...). They are split in multiple files to help compose profile from basic blocks. They can be manipulated through the pcvs config subcommand. For more details, you can refer to the documentation.

  • The test descriptions describe the test suite (how to build, run and validate). They are independent from the profiles to help share suites adapted to PCVS across different platforms. They can be static (a plain YAML file) or dynamic (generated on the fly by a script). A complete example of a test description can be found in the documentation.

The validity of the YAML files can be asserted using the pcvs check subcommand.

Run a test suite

Once the configurations setup, they can be reused indefinitely to run tests.

$ pcvs run -p <profile> <path/to/suite>

Using the above command will launch PCVS on a given test suite, building and validating it. A quick summary is available at the end of the run.

Commonly used options
  • Print per test validation
    $ pcvs -v run <path/to/suite>
    
  • Print the output of the tests as well
    $ pcvs -v run --print [errors|all] <path/to/suite>
    
  • Exit with a non-zero error code on test failure
    $ pcvs run -S <path/to/suite>
    
  • Filter tests to run (based on tags)
    $ pcvs run --run-filter <list/of/tags> <path/to/suite>
    

Visualize the results

Once the PCVS run completed, there are two ways to visualize the results:

  • Using the webview with pcvs report subcommand and opening localhost:5000 in a browser
  • Using the TUI with pcvs --tui report

The output of the different tests can be retrieved from either as well as the complete command ran.

Complete documentation

The user guide is hosted on GitHub pages: https://cea-hpc.github.io/pcvs. Every CLI command has a builtin help that can be displayed using the -h option. Alternatively, the documentation can be generated locally from the source code.

How to generate the documentation
  • The CLI is managed and documented through click. The manpages can be automatically built with the third-party tool click-man (not a dependency, should be installed manually). Note that these manpages may not contain more information than the content of each --help command.
  • The general documentation use sphinx and can be built by running make html in the docs directory. The entry of the documentation will be build/html/index.html.

Authors

This work is currently supported by the French Alternative Energies and Atomic Energy Commission (CEA). For any question and/or remarks, please contact:

Licensing

PCVS is distributed under the terms of the CeCILL-C Free Software Licence Agreement.

All new contributions must be made under the CeCILL-C terms.

For more details, see COPYING or CeCILL-C website.

SPDX-Licence-Identifier: CECILL-C

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

pcvs-1.1.0.tar.gz (544.3 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

pcvs-1.1.0-py3-none-any.whl (588.9 kB view details)

Uploaded Python 3

File details

Details for the file pcvs-1.1.0.tar.gz.

File metadata

  • Download URL: pcvs-1.1.0.tar.gz
  • Upload date:
  • Size: 544.3 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for pcvs-1.1.0.tar.gz
Algorithm Hash digest
SHA256 1e46f99ebd391e49aff056362216c555ce8fe3fea02145919b16cf153eb17987
MD5 85eeced7bde97486256d197919daa88c
BLAKE2b-256 e0c534a989738befcf82f3bcf3cc73a5f73b4f40b6eeab31cd2506063a7f73cf

See more details on using hashes here.

Provenance

The following attestation bundles were made for pcvs-1.1.0.tar.gz:

Publisher: release.yml on cea-hpc/pcvs

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file pcvs-1.1.0-py3-none-any.whl.

File metadata

  • Download URL: pcvs-1.1.0-py3-none-any.whl
  • Upload date:
  • Size: 588.9 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for pcvs-1.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 b9b4973798317b955553534391493fa342895463dda93b244378898f691106a8
MD5 07d7caf08e36bc25fb94039734fcb348
BLAKE2b-256 6029dbcefd654208ecb1f4d4d1144ff77fc33447f5e724a1b155b755cb6dc9e2

See more details on using hashes here.

Provenance

The following attestation bundles were made for pcvs-1.1.0-py3-none-any.whl:

Publisher: release.yml on cea-hpc/pcvs

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page