OllamaBenchmarks

Run the Ollama benchmark harness in this repository, capture host metadata, archive per-system benchmark snapshots, and update cross-system comparison tables. Use when Codex needs to benchmark the current machine, add another system to the benchmark matrix, or refresh summaries under results/systems and results/cross-system-summary.md.

Use when deploying Ollama models as Docker containers for use with Docker Agent, when selecting the best local models from benchmark results, or when creating docker-agent YAML configurations. Covers the full pipeline from benchmarking through containerised deployment.