EdgeDB Benchmark Toolbench

This is a collection of benchmarks intended to compare EdgeDB against different databases and ORMs.

Installation and Use

1. Make a Python 3.8 venv (to be compatible with synth).

2. Install dependencies from requirements.txt.

3. Install EdgeDB, MongoDB, PostgreSQL 13, Golang toolchain.

4. Install NodeJS (at least v14.16.0) and TypeScript toolchains.

5. Install docker and docker-compose (needed for Hasura).

6. Install Prisma via npm install prisma -D.

7. Install synth. (https://www.getsynth.com)

   Note: Synth v0.5.0 replaces python faker with fake-rs, and loses support for some generators used by this project. The previous supported version, Synth v0.4.7, can be installed from https://github.com/getsynth/synth/releases/tag/v0.4.7

8. Create the benchmark data via $ make new-dataset.

9. Load data via $ make load.

10. Compile Go benchmarks: $ make go

11. Compile TypeScript benchmarks: $ make ts

12. Run benchmarks via bench.py, for example:

python bench.py --html out.html --concurrency 10 -D 10 all

Dataset

The default dataset (100000 people, 100000 users, 500000 reviews) is included in the repository. A new dataset can be generated via $ make new-dataset.

License

Apache 2.0.