Performance Comparisons

Detailed performance analysis comparing Kapper with other popular JVM database libraries using real benchmark data.

Benchmark Overview

Our benchmarks test realistic database operations using JMH (Java Microbenchmark Harness):

• find100 - Retrieve 100 records
• findById - Single record lookup by primary key
• insertSingleRow - Single record insertion
• updateSingleRow - Single record update
• simpleJoin - Basic JOIN query with mapping

Libraries Tested

Library	Version	Type	Notes
Kapper	1.6.1	Micro ORM	SQL-first mapping library
Hibernate	Latest	Full ORM	Traditional Java ORM
Ktorm	Latest	Kotlin DSL	Type-safe query DSL
JDBC	-	Direct API	Raw database access

Test Environment

• Platform: Oracle Cloud ARM instances (4 OCPUs, 24GB RAM)
• JVM: GraalVM JDK 21 with optimised JIT compilation
• Framework: JMH (Java Microbenchmark Harness)
• Database: SQLite 3.50 (embedded), PostgreSQL 14 (Testcontainers)
• Dataset: 1000-row tables with realistic data structures

SQLite Performance (Embedded Database)

Find Operations Performance

SQLite find operations comparison showing Kapper's competitive positioning

Key Findings:

• JDBC maintains raw performance leadership as expected for direct SQL
• Kapper with auto-mapping shows excellent ORM competitiveness especially vs Ktorm
• Ktorm significantly slower in single-record operations (10x slower findById)

Modification Operations Performance

SQLite modification and JOIN operations showing Kapper's near-JDBC performance

Key Findings:

• Kapper nearly matches JDBC performance for modifications (10-13% overhead)
• Dramatic advantages over ORMs - 2-20x faster than Hibernate/Ktorm
• Consistent low-overhead pattern across all modification operations

Performance Categories Summary

Performance Trade-offs

JDBC vs Kapper:

• JDBC Advantages: Raw performance, no mapping overhead
• Kapper Advantages: Type safety, automatic mapping, cleaner code
• Trade-off: 10-70% performance cost for significant development productivity

Kapper vs Full ORMs (Hibernate):

• Kapper Advantages: 69-282% faster, predictable performance, no lazy loading
• ORM Advantages: Advanced features (caching, lazy loading, change tracking)
• Sweet Spot: High performance with essential mapping features

Kapper vs Kotlin DSLs (Ktorm):

• Kapper Advantages: 36-1970% faster, direct SQL control
• DSL Advantages: Type-safe query building, fluent API
• Trade-off: Kapper's SQL-first approach wins on performance

Kapper Mapping Strategy Comparison

Within Kapper itself, different mapping approaches offer performance trade-offs:

Mapping Strategy Performance (SQLite)

Detailed comparison of Kapper's different mapping strategies showing performance trade-offs

Performance Impact Analysis

Most Significant Gains:

• find100 (Bulk Operations): Manual mapping 35.5% faster than auto-mapping (161.6μs vs 250.4μs for Kotlin data classes)

Minimal Impact Operations:

• Single-row modifications: All strategies perform similarly (~0-2% difference)
• Simple lookups: Small but consistent manual mapping advantage (~4%)

Strategic Mapping Choices

// High-frequency, performance-critical paths
connection.query("SELECT id, name FROM users") { rs, _ ->
    User(rs.getLong("id"), rs.getString("name"))
}

// Balanced performance with convenience
data class User(val id: Long, val name: String)
connection.query<User>("SELECT id, name FROM users")

Next Steps

• Performance History - Track performance evolution across versions