Entry point for agent context in this repo.
dpnp is the Data Parallel Extension for NumPy: a NumPy-compatible Python API with accelerated execution backends and cross-platform packaging/CI.
For detailed documentation, build instructions, and contribution guidelines, see:
README.md(project overview and quickstart)CONTRIBUTING.md(if exists, or refer to project docs)
- Keep changes small and single-purpose (bug fix, perf, docs, or infra).
- Preserve NumPy-compatible API behavior by default; call out intentional behavior changes in PR text.
- Pair user-visible changes with tests and docs/examples updates.
- Avoid duplicating mutable details (versions/flags/matrix) in instruction docs.
For Copilot behavior policy (flow, source-of-truth, edit hygiene), see:
.github/copilot-instructions.md(canonical for Copilot behavior)
Below directories have local AGENTS.md for deeper context:
.github/AGENTS.md— CI/workflow/automation contextdpnp/AGENTS.md— core NumPy-compatible implementation guardrailsdoc/AGENTS.md— documentation updates and consistencyexamples/AGENTS.md— runnable user-facing examplesbenchmarks/AGENTS.md— reproducible performance evaluationscripts/AGENTS.md— maintenance/developer scriptstests_external/AGENTS.md— external/integration validation scope