A DESCRIPTION OF THE PROBLEM :
When I executed the java file, I noticed a significant performance difference between jdk11, jdk21 and jdk8 when executing a large number of loops. jdk11 ran over 400 times longer than jdk8. I think the difference is in the helperMethod method, so I should use "inline optimization" in this part to reduce the runtime.

EXPECTED VERSUS ACTUAL BEHAVIOR :
EXPECTED -
time ./hotspot-jdk8u352/build/linux-x86_64-normal-server-fastdebug/jdk/bin/java Test
output:
2

real 0m2.967s
user 0m0.842s
sys 0m1.549s
ACTUAL -
time ./hotspot-jdk-21.0.2/build/linux-x86_64-server-fastdebug/images/jdk/bin/java Test

output:
2

real 3m44.658s
user 3m46.811s
sys 0m1.244s

time ./hotspot-jdk-11.0.15/build/linux-x86_64-normal-server-fastdebug/jdk/bin/java Test

output:
2

real 14m12.330s
user 54m51.359s
sys 0m1.800s

---------- BEGIN SOURCE ----------
import java.util.concurrent.CountDownLatch;

public class Test {

    static int n = 847734685;

    String countAndSay(int n) throws Exception {
        byte[] sessionId = { -96, 68, -45, 10, 31, -126, -47, 68, -76, -67 };
        if (sessionId == null) {
            throw new NullPointerException("session id cannot be null");
        }
        
        for (int i = 0; i < n; i++) {
            int j = helperMethod(i);
        }
        
        String[] strs = new String[n + 1];
        strs[0] = "2";
        return strs[0];
    }
    
    int helperMethod(int index) {
        int num1 = index * 2;
        int num2 = num1 + 5;
        int result = calculateResult(num1, num2);
        return result;
    }
    
    int calculateResult(int a, int b) {
        return a * b;
    }

    public static void main(String[] args) throws Exception {
        System.out.println(new Test().countAndSay(n));
    }
}
---------- END SOURCE ----------