This project compares the execution speed of a simple computational task across different programming languages.
Inspired by Ben Dicken's loop performance comparison and his excellent visual analysis.
Each benchmark performs the same computation:
- Creates an array of 10,000 elements
- Runs nested loops to perform arithmetic operations
- Measures and reports execution time
- Python 🐍 (using PyPy)
- JavaScript 🟨
- Go 🦫
- C 🎯
- Rust 🦀
Make sure you have the following installed:
- PyPy
- Node.js
- Go
- GCC (GNU Compiler Collection)
- Rust
Simply execute the benchmark script:
./run_benchmarks.shMy machine:
Chip: Apple M3 Pro
Total Number of Cores: 11 (5 performance and 6 efficiency)
Memory: 18 GB
Running benchmarks with hyperfine...
-----------------------------------
Benchmark 1: node src/loop_benchmark.js
Time (mean ± σ): 764.0 ms ± 24.4 ms [User: 751.3 ms, System: 7.6 ms]
Range (min … max): 744.8 ms … 815.3 ms 10 runs
Benchmark 2: pypy src/loop_benchmark.py
Time (mean ± σ): 1.035 s ± 0.014 s [User: 1.022 s, System: 0.010 s]
Range (min … max): 1.025 s … 1.061 s 10 runs
Benchmark 3: ./build/loop_benchmark_go
Time (mean ± σ): 292.8 ms ± 36.6 ms [User: 289.7 ms, System: 1.8 ms]
Range (min … max): 268.4 ms … 387.1 ms 10 runs
Benchmark 4: ./build/loop_benchmark_c
Time (mean ± σ): 1.8 ms ± 0.5 ms [User: 0.4 ms, System: 0.7 ms]
Range (min … max): 1.3 ms … 2.5 ms 10 runs
Benchmark 5: ./build/loop_benchmark_rust
Time (mean ± σ): 1.9 ms ± 0.5 ms [User: 0.7 ms, System: 0.7 ms]
Range (min … max): 1.3 ms … 2.8 ms 10 runs
Summary
./build/loop_benchmark_c ran
1.08 ± 0.42 times faster than ./build/loop_benchmark_rust
162.60 ± 47.66 times faster than ./build/loop_benchmark_go
424.30 ± 113.33 times faster than node src/loop_benchmark.js
574.87 ± 152.64 times faster than pypy src/loop_benchmark.py