Presto C++ Workshop¶
Deploying Presto C++ and Benchmarking with pbench¶
Welcome to our Presto C++ workshop! This workshop will provide an overview of Presto C++, the next-gen Presto worker. You will learn how to deploy Presto C++ with Docker Compose on your machine and query TPC-H, TPC-DS data from your local file-system with the Hive and Iceberg connectors (using file-based hive metastore). We will then demonstrate how to run queries manually with Presto CLI and look at how the open source benchmarking tool pbench can be used to run TPC-DS benchmarks against your Presto cluster. We will use pbench to compare the performance of Presto and Presto C++ on the TPC-DS benchmark.
By the end of the workshop, you will know
- What Presto C++ is and its benefits
- How to set up and deploy Presto C++ with Docker Compose locally
- How to use the Presto CLI to run queries
- How to use pbench to run benchmarks and view results
- How to get a quantified speed-up factor of Presto C++ vs Java
This workshop is organized into the following sections:
Agenda¶
| Part 1 | Set-up and deployment of Presto C++ and Presto Java locally |
| Part 2 | Downloading and running pbench |
Compatibility¶
This workshop has been tested on the following platforms:
- aarch64: Apple Silicon (M-series chips)
Technology Used¶
- Docker: A container engine to run several applications in self-contained containers
- Docker Compose: A YAML-based tool for defining and running multi-container applications
- Presto: Fast and Reliable SQL Engine for Data Analytics and the Open Lakehouse
- pbench: A new, open-source benchmarking tool for Presto