Skip to main content
Version: 0.16

Benchmark

Note: Different serialization frameworks excel in different scenarios. Benchmark results are for reference only. For your specific use case, conduct benchmarks with appropriate configurations and workloads.

Java Benchmark​

The Java benchmark section compares Fory against popular Java serialization frameworks using the current benchmark suite from docs/benchmarks/java.

Serialization Throughput:

Java Serialization Throughput

Deserialization Throughput:

Java Deserialization Throughput

Zero-Copy Serialize Throughput:

Zero-Copy Deserialize Throughput:

Important: Fory's runtime code generation requires proper warm-up for performance measurement:

For additional benchmark notes, raw data, and the complete Java benchmark README, see Java Benchmarks.

Python Benchmark​

Fory Python demonstrates strong performance compared to pickle and Protobuf across object and list workloads.

Python Throughput

For benchmark setup, raw results, and reproduction steps, see Python Benchmarks.

Rust Benchmark​

Fory Rust demonstrates competitive performance compared to other Rust serialization frameworks.

Note: Results depend on hardware, dataset, and implementation versions. See the Rust guide for how to run benchmarks yourself: https://github.com/apache/fory/blob/main/benchmarks/rust_benchmark/README.md

C++ Benchmark​

Fory C++ demonstrates competitive performance compared to Protobuf C++ serialization framework.

Go Benchmark​

Fory Go demonstrates strong performance compared to Protobuf and Msgpack across single-object and list workloads.

Note: Results depend on hardware, dataset, and implementation versions. See the Go benchmark report for details: https://fory.apache.org/docs/benchmarks/go/

C# Benchmark​

Fory C# demonstrates strong performance compared to Protobuf and Msgpack across typed object serialization and deserialization workloads.

Note: Results depend on hardware and runtime versions. See the C# benchmark report for details: https://fory.apache.org/docs/benchmarks/csharp/

Swift Benchmark​

Fory Swift demonstrates strong performance compared to Protobuf and Msgpack across both scalar-object and list workloads.

Note: Results depend on hardware and runtime versions. See the Swift benchmark report for details: https://fory.apache.org/docs/benchmarks/swift/

JavaScript Benchmark​

The data used for this bar graph includes a complex object that has many kinds of field types, and the size of the JSON data is 3KB.

See benchmarks for the benchmark code.