How can I visualize the benchmarking results?

Criterium supports various viewers, including human-readable text output, structured data (pprint), interactive charts in Portal, and Vega-Lite charts for Clay/Clerk notebooks via the Kindly viewer.

Does Criterium support benchmarking with randomized or generated data?

Yes, through the criterium/arg-gen extension, you can use test.check generators to provide fresh, diverse inputs for every benchmark iteration to avoid biased results.

What is 'Domain Analysis' in the context of Criterium?

Domain Analysis benchmarks code across a parameter space (like increasing input sizes), allowing you to analyze scaling behavior and fit results to complexity models like O(n) or O(n log n).

Why should I use Criterium instead of Clojure's built-in 'time' macro?

Criterium accounts for JVM-specific factors like JIT compilation and garbage collection, providing statistically significant results with confidence intervals rather than a single, potentially skewed timing measurement.

Criterium Clojure Benchmarking

Name: Criterium Clojure Benchmarking
Author: hugoduncan

byhugoduncan

•

1,220

Security & Testing

Conducts statistically rigorous performance benchmarks for Clojure code by accounting for JVM warmup, garbage collection, and measurement overhead.

About

Criterium is the industry-standard benchmarking tool for Clojure developers, moving beyond naive timing methods to provide high-fidelity performance metrics. This skill enables Claude to guide users through complex performance analysis, including JIT-aware measurements, statistical outlier detection, and algorithmic complexity modeling (O-notation). It supports a wide range of output formats like Portal and interactive charts, and integrates with test.check for sophisticated argument generation, making it indispensable for optimizing performance-critical Clojure applications.

Key Features

Statistical rigor with bootstrap confidence intervals and outlier detection
Integration with test.check for automated, generative benchmark inputs
1,220 GitHub stars
JVM-aware measurements handling JIT warmup and GC interference
Domain analysis for comparing implementations and modeling algorithmic complexity
Support for multiple viewers including Portal, Clay/Clerk, and text output

Use Cases

Determining the algorithmic complexity (O-notation) of a function across varied input sizes
Comparing the performance of multiple Clojure function implementations with statistical confidence
Identifying performance bottlenecks and regressions in production-grade Clojure systems

About

Key Features

Statistical rigor with bootstrap confidence intervals and outlier detection
Integration with test.check for automated, generative benchmark inputs
1,220 GitHub stars
JVM-aware measurements handling JIT warmup and GC interference
Domain analysis for comparing implementations and modeling algorithmic complexity
Support for multiple viewers including Portal, Clay/Clerk, and text output

Use Cases

Determining the algorithmic complexity (O-notation) of a function across varied input sizes
Comparing the performance of multiple Clojure function implementations with statistical confidence
Identifying performance bottlenecks and regressions in production-grade Clojure systems