Day 6: For Loop vs Foreach Loop Performance, and Why My First Benchmark Was Misleading

Update, May 2026: the original benchmark below should not be treated as evidence that a classic for loop is meaningfully faster than foreach. It used System.currentTimeMillis(), no warmup, no repeated forks, and loops whose work could be optimized away. A proper Java microbenchmark should use JMH and should consume real work so the JIT cannot remove the measurement target.

This post is kept as a learning snapshot, but the conclusion has changed: prefer the loop that makes the code clearer unless a production measurement with representative data shows otherwise.

Transcribed from the original LinkedIn image post.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
import java.util.ArrayList;
import java.util.List;

public class Day06 {
    public static void main(String args[]) {
        List<String> fileTypeList = new ArrayList<>();
        for (int i = 0; i < 1000000; i++) {
            fileTypeList.add("fileType");
        }

        long beforeForLoop = System.currentTimeMillis();
        for (int i = 0; i < fileTypeList.size(); i++) {
            fileTypeList.get(i);
        }
        long afterForLoop = System.currentTimeMillis();
        System.out.println("Time took in millis for for " + (afterForLoop - beforeForLoop));

        long beforeForeachLoop = System.currentTimeMillis();
        for (String s : fileTypeList) {
        }

        long afterForeachLoop = System.currentTimeMillis();
        System.out.println("Time took in millis for foreach " + (afterForeachLoop - beforeForeachLoop));
    }
}

What a better benchmark needs

A useful version of this test needs:

JMH warmup and measurement iterations.
Several forks so one JVM run does not dominate the result.
Real work inside each loop.
A consumed result so the JIT cannot remove the loop.
Representative data structures, because arrays, ArrayList, LinkedList, and custom collections do not behave the same way.

For this series, Day 69 covers JMH in more detail: Day 69 - Unlocking Java Performance Secrets.

The original LinkedIn graphic is preserved below.

What a better benchmark needs#

What a better benchmark needs