[QUESTION]某些场景下json字符串转为json对象性能比fastjson1差

Int100是官方测试文档中字段数为100、类型为Int的类，在本地测试发现fastjson1的运行时长反而小于fastjson2的运行时长，请问这是什么原因
// java对象序列化为字符串
@Test
public void testFastjson1JsonObjectToString() {
ThreadMXBean threadMXBean = ManagementFactory.getThreadMXBean();
int num = 100000;
// 对象
Int100 int100 = new Int100();
long start = System.currentTimeMillis();
long startCpuTime = threadMXBean.getCurrentThreadCpuTime();

for (int i = 0; i < num; i++) {
// 循环计算处理时间
String jsonString = JSON.toJSONString(int100);
}

long end = System.currentTimeMillis();
long endCpuTime = threadMXBean.getCurrentThreadCpuTime();

long elapsedTime = end - start;
long elapsedCpuTime = endCpuTime - startCpuTime;

// cpu时间单位为纳秒，1纳秒 = 1000000毫秒
double cpuUsage = (elapsedCpuTime / (elapsedTime * 1000000.0)) * 100.0;
// 计算时间
System.out.println("运行时间：" + elapsedTime);
System.out.println("CPU 占用率: " + cpuUsage + "%");
}

// java对象序列化为字符串
@Test
public void testFastjson2JsonObjectToString() {
ThreadMXBean threadMXBean = ManagementFactory.getThreadMXBean();
int num = 100000;
// 对象
Int100 int100 = new Int100();
long start = System.currentTimeMillis();
long startCpuTime = threadMXBean.getCurrentThreadCpuTime();

for (int i = 0; i < num; i++) {
// 循环计算处理时间
String jsonString = com.alibaba.fastjson2.JSON.toJSONString(int100);
}

long end = System.currentTimeMillis();
long endCpuTime = threadMXBean.getCurrentThreadCpuTime();

long elapsedTime = end - start;
long elapsedCpuTime = endCpuTime - startCpuTime;

double cpuUsage = (elapsedCpuTime / (elapsedTime * 1000000.0)) * 100.0;

System.out.println("运行时间：" + elapsedTime);
System.out.println("CPU 占用率: " + cpuUsage + "%");
}

Start with the two inline tests, testFastjson1JsonObjectToString and testFastjson2JsonObjectToString, and reproduce them with the referenced Int100 class under identical Java and library versions. Record comparable runtime and CPU measurements; done means the performance difference is explained or reduced to a reproducible benchmark result.

Written by the indexing model from the issue text.