README.md

serde_qs Benchmarks

This directory contains performance benchmarks for the serde_qs crate using Criterion.rs.

Overview

The benchmarks are organized into several categories to measure performance across different use cases:

1. Simple Data Structure Benchmarks

• Flat structs: Basic structures with primitive types
• Optional fields: Structures with Option<T> fields
• Collections: Vec<T> and HashMap<K,V> performance
• Primitive types: String, integer, boolean serialization/deserialization

2. Complex/Nested Data Structure Benchmarks

• Nested structs: Objects containing other objects (Address within QueryParams)
• Deep nesting: 3-4 levels of nested structures
• Large collections: Arrays with 100+ items
• Mixed structures: Combinations of objects, arrays, and maps

3. Performance Comparisons

• vs serde_urlencoded: Direct comparison for flat structures where both libraries work

Benchmark Results

Performance Baseline (v1.0)

Simple Data Structures

Operation	Structure	Time
Serialize	Simple struct	~183ns
Deserialize	Simple struct	~296ns
Serialize	HashMap (3 items)	~199ns
Deserialize	HashMap (3 items)	~455ns
Serialize	Vec (5 items)	~309ns
Deserialize	Vec (5 items)	~491ns

Complex/Nested Data Structures

Operation	Structure	Time
Serialize	Nested struct (2 levels)	~739ns
Deserialize	Nested struct (2 levels)	~1.28μs
Serialize	Deep nested (4 levels)	~1.10μs
Deserialize	Deep nested (4 levels)	~2.47μs
Serialize	Large vec (100 items)	~4.97μs
Deserialize	Large vec (100 items)	~8.16μs

Comparison with serde_urlencoded

Structure	Library	Serialize	Deserialize
Simple struct	serde_qs	~181ns	~295ns
Simple struct	serde_urlencoded	~136ns (33% faster)	~183ns (61% faster)
HashMap	serde_qs	~198ns	~464ns
HashMap	serde_urlencoded	~174ns (14% faster)	~352ns (32% faster)