Python providers

Python providers#

Provider ownership and installation#

Public name

Distribution

Runtime or problem definitions

Removal boundary

s2mpj

Bundled in optiprofiler

Bundled S2MPJ subset

Installed and removed with the core.

pycutest

optiprofiler-pycutest

User-managed PyCUTEst, CUTEst, MASTSIF, and compiled problems

Removing the adapter preserves the complete upstream runtime.

solar

optiprofiler-solar

Slim SOLAR source in the adapter; executable in a user cache

Removing the adapter preserves its external cache and user output.

rs13

optiprofiler-rs13 (experimental)

Official RS13 Python problem definitions shipped with the adapter

Removing the adapter removes its packaged problems but preserves benchmark output and unrelated user data.

S2MPJ#

S2MPJ is the default provider and needs no separate installation. Its public selection and loading functions are s2mpj_select() and s2mpj_load().

PyCUTEst#

The PyCUTEst adapter connects OptiProfiler to an existing upstream PyCUTEst installation. Install and validate PyCUTEst first, then install optiprofiler-pycutest. Platform-specific compiler, CUTEst, and MASTSIF setup remains entirely upstream-owned.

SOLAR#

The SOLAR adapter includes the slim runtime source used by its wrapper. It builds the executable lazily and caches the result outside the Python distribution. Set SOLAR_CACHE_DIR only when a non-default cache root is required.

RS13#

RS13 is an experimental provider for the Rios–Sahinidis derivative-free optimization test set. Its distribution includes the official Python problem definitions required by rs13_load together with adapter-maintained selection metadata. Complete solution vectors and independent upstream audit records are maintainer test inputs; they are not required for a benchmark and are not installed for users.

Sources and provenance#

The OptiProfiler repositories below are integration layers or maintained subsets, not replacements for the original projects. Please follow the upstream projects’ citation and license guidance when using their problem collections.