[Bug 289292] [NEW PORT] benchmarks/py-reframe: HPC system validation & performance regression framework

From: <bugzilla-noreply_at_freebsd.org>
Date: Thu, 04 Sep 2025 09:43:38 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=289292

            Bug ID: 289292
           Summary: [NEW PORT] benchmarks/py-reframe: HPC system
                    validation & performance regression framework
           Product: Ports & Packages
           Version: Latest
          Hardware: Any
                OS: Any
            Status: New
          Severity: Affects Only Me
          Priority: ---
         Component: Individual Port(s)
          Assignee: ports-bugs@FreeBSD.org
          Reporter: rikka.goering@outlook.de

This adds a new port: benchmarks/py-reframe

ReFrame is a framework for writing and running portable regression and
performance tests for HPC systems. Tests are plain Python and can target
different systems, programming environments, and job schedulers (e.g., Slurm).
Typical use-cases include CI-style cluster validation, smoke tests after
maintenance, and comparative benchmarking across nodes/partitions.

Rationale:
- Complements the existing HPC stack in ports; I maintain sysutils/slurm-wlm
  and also submitted sysutils/py-clustershell. ReFrame fits naturally for
  scheduler-aware testing and benchmarking on FreeBSD clusters.

Port notes:
- Category: benchmarks
- Language: Python (installed as a console script, architecture-independent)
- NO_ARCH=yes
- Follows PEP 517/518 build if upstream uses it (USES=python …)
- Installed entry point: `reframe`
- Depends on py-clustershell (see bug #289176)

QA:
- Builds and packages cleanly in poudriere (14.3-RELEASE/amd64) with default
  options.
- `portfmt`/`portclippy` clean.
- `stage-qa`/`check-plist` clean.
- Runtime test: `reframe --version` and a minimal tutorial test executed
  via Slurm on a local node.

Upstream:
- Homepage/Docs: https://github.com/reframe-hpc/reframe
- License: BSD 3-Clause License

Reviewer focus:
- Placement in `benchmarks/` vs. `devel/`—I chose benchmarks due to the
  performance/validation focus and scheduler integration.
- Python version range and RUN_DEPENDS as packaged here.

-- 
You are receiving this mail because:
You are the assignee for the bug.