Does a try‑catch Inside a Java Loop Really Slow Down Your Code?

Two developers discuss a Java snippet where a try-catch block is placed inside a for loop, questioning its performance impact.

for (int i = 0; i < 5000; i++) {
    try {
        dosth;
    } catch (Exception e) {
        e.printStackTrace();
    }
}

One suggests moving the try-catch outside the loop, while the other argues that the loop’s purpose is to keep execution going even if a single iteration fails.

To settle the debate, they write two benchmark methods using JMH:

public class TryCatchTest {
    @Benchmark
    public void tryfor(Blackhole blackhole) {
        try {
            for (int i = 0; i < 5000; i++) {
                blackhole.consume(i);
            }
        } catch (Exception e) {
            e.printStackTrace();
        }
    }

    @Benchmark
    public void fortry(Blackhole blackhole) {
        for (int i = 0; i < 5000; i++) {
            try {
                blackhole.consume(i);
            } catch (Exception e) {
                e.printStackTrace();
            }
        }
    }
}

The JMH results show that the performance difference between the two approaches is negligible:

fortry: 86,261 (~ 100,359 ± 14,098) to 114,457 (~ 100,359 ± 14,098)

tryfor: 95,961 (~ 103,216 ± 7,255) to 110,471 (~ 103,216 ± 7,255)

Even when the loop count is reduced to 1,000, the scores remain similar.

Bytecode inspection confirms that both methods generate almost identical bytecode, differing only in the range of the exception table entries.

The discussion then references Effective Java , which warns against using try-catch for normal control flow because it can hinder JVM optimizations such as instruction reordering.

Key take‑aways about try-catch performance:

Using try-catch instead of normal code can introduce some overhead, so it should not replace regular logic when unnecessary.

Placing try-catch inside a loop versus wrapping the whole loop yields similar performance; the choice mainly affects program logic, not speed.

While try-catch does have a cost, it is usually negligible in typical business code; avoid broad catch blocks and only catch what you need.