Skip to content

andygrove/datafusion-comet

 
 

Repository files navigation

Apache DataFusion Comet

Apache licensed Discord chat

logo

Apache DataFusion Comet is a high-performance accelerator for Apache Spark, built on top of the powerful Apache DataFusion query engine. Comet is designed to significantly enhance the performance of Apache Spark workloads while leveraging commodity hardware and seamlessly integrating with the Spark ecosystem without requiring any code changes.

Benefits of Using Comet

Run Spark Queries at DataFusion Speeds

Comet delivers a performance speedup for many queries, enabling faster data processing and shorter time-to-insights.

The following chart shows the time it takes to run the 22 TPC-H queries against 100 GB of data in Parquet format using a single executor with 8 cores. See the Comet Benchmarking Guide for details of the environment used for these benchmarks.

When using Comet, the overall run time is reduced from 615 seconds to 364 seconds, a 1.7x speedup, with query 1 running 9x faster than Spark.

Running the same queries with DataFusion standalone (without Spark) using the same number of cores results in a 3.6x speedup compared to Spark.

Comet is not yet achieving full DataFusion speeds in all cases, but with future work we aim to provide a 2x-4x speedup for a broader set of queries.

Here is a breakdown showing relative performance of Spark, Comet, and DataFusion for each TPC-H query.

The following charts shows how much Comet currently accelerates each query from the benchmark.

Relative speedup

Absolute speedup

These benchmarks can be reproduced in any environment using the documentation in the Comet Benchmarking Guide. We encourage you to run your own benchmarks.

Results for our benchmark derived from TPC-DS are available in the benchmarking guide.

Use Commodity Hardware

Comet leverages commodity hardware, eliminating the need for costly hardware upgrades or specialized hardware accelerators, such as GPUs or FPGA. By maximizing the utilization of commodity hardware, Comet ensures cost-effectiveness and scalability for your Spark deployments.

Spark Compatibility

Comet aims for 100% compatibility with all supported versions of Apache Spark, allowing you to integrate Comet into your existing Spark deployments and workflows seamlessly. With no code changes required, you can immediately harness the benefits of Comet's acceleration capabilities without disrupting your Spark applications.

Tight Integration with Apache DataFusion

Comet tightly integrates with the core Apache DataFusion project, leveraging its powerful execution engine. With seamless interoperability between Comet and DataFusion, you can achieve optimal performance and efficiency in your Spark workloads.

Active Community

Comet boasts a vibrant and active community of developers, contributors, and users dedicated to advancing the capabilities of Apache DataFusion and accelerating the performance of Apache Spark.

Getting Started

To get started with Apache DataFusion Comet, follow the installation instructions. Join the DataFusion Slack and Discord channels to connect with other users, ask questions, and share your experiences with Comet.

Follow Apache DataFusion Comet Overview to get more detailed information

Contributing

We welcome contributions from the community to help improve and enhance Apache DataFusion Comet. Whether it's fixing bugs, adding new features, writing documentation, or optimizing performance, your contributions are invaluable in shaping the future of Comet. Check out our contributor guide to get started.

License

Apache DataFusion Comet is licensed under the Apache License 2.0. See the LICENSE.txt file for details.

Acknowledgments

We would like to express our gratitude to the Apache DataFusion community for their support and contributions to Comet. Together, we're building a faster, more efficient future for big data processing with Apache Spark.

About

Apache Arrow DataFusion Comet Spark Accelerator

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Rust 45.8%
  • Scala 39.9%
  • Java 12.8%
  • Shell 0.9%
  • Python 0.2%
  • Dockerfile 0.2%
  • Makefile 0.2%