jvm基础知识1

堆放实例对象，栈放实例对象的引用，方法区存储创建类的信息

上面堆和垃圾回收的关系，垃圾回收回收的是堆内存的数据，s0和s1区域，

例如现在我们要清除s0中的堆对象，将s0中正在运行的对象从s0区域移动到s1区域中，然后将s0中要被清除的堆对象清除，现在对象都存储在s1中，s0中为空，即下来要清除s1中的对象，将不被清除的对象从s1中移动到s0中，将s1中的对象清除，依次类推

接下来学习下虚拟机的参数：

1、第一点给堆设置参数

-Xms5m -Xmx20m

初始化堆大小5m，最大设置为20M,-XX:+PrintCommandLineFlags将配置的参数从控制台打印输出

配置参数：

我们来看下面代码的运行结果：

package com.bjsxt.base001;


public class Test01 {

    public static void main(String[] args) {

        //-Xms5m -Xmx20m -XX:+PrintGCDetails -XX:+UseSerialGC -XX:+PrintCommandLineFlags
        
        //查看GC信息
        System.out.println("max memory:" + Runtime.getRuntime().maxMemory());
        System.out.println("free memory:" + Runtime.getRuntime().freeMemory());
        System.out.println("total memory:" + Runtime.getRuntime().totalMemory());
        
        byte[] b1 = new byte[1*1024*1024];
        System.out.println("分配了1M");
        System.out.println("max memory:" + Runtime.getRuntime().maxMemory());
        System.out.println("free memory:" + Runtime.getRuntime().freeMemory());
        System.out.println("total memory:" + Runtime.getRuntime().totalMemory());
        
        byte[] b2 = new byte[4*1024*1024];
        System.out.println("分配了4M");
        System.out.println("max memory:" + Runtime.getRuntime().maxMemory());
        System.out.println("free memory:" + Runtime.getRuntime().freeMemory());
        System.out.println("total memory:" + Runtime.getRuntime().totalMemory());
        
    }
    
}

-XX:InitialHeapSize=5242880 -XX:MaxHeapSize=20971520 -XX:+PrintCommandLineFlags -XX:+PrintGCDetails -XX:+UseCompressedOops -XX:-UseLargePagesIndividualAllocation -XX:+UseSerialGC
max memory:20316160
free memory:4418752
total memory:5111808
[GC[DefNew: 676K->127K(1536K), 0.0014042 secs] 676K->481K(4992K), 0.0014266 secs] [Times: user=0.00 sys=0.00, real=0.00 secs]
分配了1M
max memory:20316160
free memory:3541280
total memory:5111808
[GC[DefNew: 1180K->0K(1536K), 0.0010578 secs][Tenured: 1505K->1505K(3456K), 0.0022471 secs] 1533K->1505K(4992K), [Perm : 2511K->2511K(21248K)], 0.0033312 secs] [Times: user=0.02 sys=0.00, real=0.00 secs]
分配了4M
max memory:20316160
free memory:3625944
total memory:9441280
Heap
def new generation total 1664K, used 108K [0x00000000f9a00000, 0x00000000f9bc0000, 0x00000000fa0a0000)
eden space 1536K, 7% used [0x00000000f9a00000, 0x00000000f9a1b100, 0x00000000f9b80000)
from space 128K, 0% used [0x00000000f9b80000, 0x00000000f9b80000, 0x00000000f9ba0000)
to space 128K, 0% used [0x00000000f9ba0000, 0x00000000f9ba0000, 0x00000000f9bc0000)
tenured generation total 7556K, used 5601K [0x00000000fa0a0000, 0x00000000fa801000, 0x00000000fae00000)
the space 7556K, 74% used [0x00000000fa0a0000, 0x00000000fa618670, 0x00000000fa618800, 0x00000000fa801000)
compacting perm gen total 21248K, used 2520K [0x00000000fae00000, 0x00000000fc2c0000, 0x0000000100000000)
the space 21248K, 11% used [0x00000000fae00000, 0x00000000fb076370, 0x00000000fb076400, 0x00000000fc2c0000)
No shared spaces configured.

[GC[DefNew: 1180K->0K(1536K), 0.0010578 secs][Tenured: 1505K->1505K(3456K), 0.0022471 secs] 1533K->1505K(4992K), [Perm : 2511K->2511K(21248K)], 0.0033312 secs] [Times: user=0.02 sys=0.00, real=0.00 secs]

表示进行了垃圾回收，DefNew表示回收了新生代，Tenured表示回收了老年代，Perm 表示回收了永久区

新生代会频繁被GC不稳定，新生代包含eden空间 s1 和s0空间

我们来看下面的列子：

package com.bjsxt.base001;

public class Test02 {

    public static void main(String[] args) {
        
        //第一次配置
        //-Xms20m -Xmx20m -Xmn1m -XX:SurvivorRatio=2 -XX:+PrintGCDetails -XX:+UseSerialGC
        
        //第二次配置
        //-Xms20m -Xmx20m -Xmn7m -XX:SurvivorRatio=2 -XX:+PrintGCDetails -XX:+UseSerialGC
        
        //第三次配置
        //-XX:NewRatio=老年代/新生代
        //-Xms20m -Xmx20m -XX:SurvivorRatio=2 -XX:+PrintGCDetails -XX:+UseSerialGC
        
        byte[] b = null;
        //连续向系统申请10MB空间
        for(int i = 0 ; i <10; i ++){
            b = new byte[1*1024*1024];
        }
    }
}

程序运行的效果是：

[GC[DefNew: 512K->255K(768K), 0.0013618 secs] 512K->439K(20224K), 0.0013894 secs] [Times: user=0.00 sys=0.00, real=0.00 secs]
Heap
def new generation total 768K, used 428K [0x00000000f9a00000, 0x00000000f9b00000, 0x00000000f9b00000)
eden space 512K, 33% used [0x00000000f9a00000, 0x00000000f9a2b300, 0x00000000f9a80000)
from space 256K, 99% used [0x00000000f9ac0000, 0x00000000f9affff8, 0x00000000f9b00000)
to space 256K, 0% used [0x00000000f9a80000, 0x00000000f9a80000, 0x00000000f9ac0000)
tenured generation total 19456K, used 10424K [0x00000000f9b00000, 0x00000000fae00000, 0x00000000fae00000)
the space 19456K, 53% used [0x00000000f9b00000, 0x00000000fa52e040, 0x00000000fa52e200, 0x00000000fae00000)
compacting perm gen total 21248K, used 2517K [0x00000000fae00000, 0x00000000fc2c0000, 0x0000000100000000)
the space 21248K, 11% used [0x00000000fae00000, 0x00000000fb075440, 0x00000000fb075600, 0x00000000fc2c0000)
No shared spaces configured.

eden space 512K，from space 256K是比例是2，SurvivorRatio=2 就是这个含义设置的意思 xmn表示新生代的值设置成1M，from区域就是s0区域，to区域就是s1区域，二者的大小是一样的

eden space 512K, 33% used [0x00000000f9a00000, 0x00000000f9a2b300, 0x00000000f9a80000)
from space 256K, 99% used [0x00000000f9ac0000, 0x00000000f9affff8, 0x00000000f9b00000)
to space 256K, 0% used [0x00000000f9a80000, 0x00000000f9a80000, 0x00000000f9ac0000)

上面三者之和就是1M

真正在应用的场景中，已经尽量减少GC，将初始化的内存大小和最大内存大小配置成一样，这样可以减少GC

package com.bjsxt.base001;

import java.util.Vector;

public class Test03 {

    public static void main(String[] args) {
        
        //-Xms1m -Xmx1m -XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=d:/Test03.dump
        //堆内存溢出
        Vector v = new Vector();
        for(int i=0; i < 5; i ++){
            v.add(new Byte[1*1024*1024]);
        }
        
    }
}

上面这个代码我申请5M的内存，但是在设置的时候却是

/-Xms1m -Xmx1m 设置成1M，就会出现堆内存异常，使用-XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath就会将信息导出来查看

package com.bjsxt.base001;

public class Test04 {

    //-Xss1m  
    //-Xss5m
    
    //栈调用深度
    private static int count;
    
    public static void recursion(){
        count++;
        recursion();
    }
    public static void main(String[] args){
        try {
            recursion();
        } catch (Throwable t) {
            System.out.println("调用最大深入：" + count);
            t.printStackTrace();
        }
    }
}