Benchmark the performance of Zarr implementations.

This repo defines Zarr-specific workloads and datasets. These are run by our general-purpose, IO-centric benchmarking framework, perfcapture.

Code maturity

This project started in early October 2023, and is currently in the planning stage! This code does not run yet.

Usage

1. Install perfcapture.
2. pip install zarr tensorstore
3. Clone the zarr-benchmark GitHub repository.
4. Run perfcapture/scripts/cli.py with --recipe-path set to your zarr-benchmark/recipes path.