GPUPhot v1.0.0 — Canonical Benchmarks & Scientific Pipeline

GPUPhot v1.0.0: Initial Release & Benchmark Reproduction Suite

This is the initial official release of GPUPhot, corresponding to the manuscript submission:
"GPUPHOT: A Python Framework for High-Performance GPU-Accelerated Photometry and Distributed Astronomical Data Reduction" (Lemes-Perera et al., 2026).

Highlights

• Deterministic GPU Detection & Photometry: Sub-pixel recentering, explicit cuML centroid seeding, and full SVD PCA solving.
• Microservices Worker Architecture: Containerized Celery worker queue with PostgreSQL/Q3C persistence and decoupled networking.
• Canonical Benchmark Provenance: Fully reproducible benchmark suite generating all 11 manuscript tables and 8 figures.
• Documentation & Tutorials: 7 interactive Jupyter notebooks and full Sphinx documentation suite.

Reproducibility

All telemetry datasets, calibration tables, and figure generation scripts for the paper are permanently archived in this release.

To clone this exact release:

git clone --branch v1.0.0 https://github.com/Light-Bridges/GPUPhot.git