几中垃圾收集器组合测试

测试代码：

public class demonopara {

private static final int _1MB = 1024*1024;

public static void main(String[] args) {

// TODO Auto-generated method stub

byte[] allocation1,allocation2,allocation3,allocation4;

allocation1 = new byte[2 * _1MB];

allocation2 = new byte[2 * _1MB];

allocation3 = new byte[2 * _1MB];

allocation4 = new byte[4 * _1MB];

}

}

JVM参数：-verbose:gc -Xms20M -Xmx20M -Xmn10M -XX:+PrintGCDetails -XX:SurvivorRatio=8

(堆大小20M，不可拓展，10M新生代，10M老生代，Eden和Survivor的比例是8：1，OK)

1.Parallel Scavenge + Serial Old(-XX:+UseParallelGC)

Heap

 PSYoungGen      total 9216K, used 7159K [0x00000007ff600000, 0x0000000800000000, 0x0000000800000000)

  eden space 8192K, 87% used [0x00000007ff600000,0x00000007ffcfde20,0x00000007ffe00000)

  from space 1024K, 0% used [0x00000007fff00000,0x00000007fff00000,0x0000000800000000)

  to   space 1024K, 0% used [0x00000007ffe00000,0x00000007ffe00000,0x00000007fff00000)

 ParOldGen       total 10240K, used 4096K [0x00000007fec00000, 0x00000007ff600000, 0x00000007ff600000)

  object space 10240K, 40% used [0x00000007fec00000,0x00000007ff000010,0x00000007ff600000)

 PSPermGen       total 21504K, used 2611K [0x00000007f9a00000, 0x00000007faf00000, 0x00000007fec00000)

  object space 21504K, 12% used [0x00000007f9a00000,0x00000007f9c8cc78,0x00000007faf00000)

//前三个对象allocation1,allocation2,allocation3都分配到了，Eden区，第四个4M的直接去了老生代。。。

//-XX：+PretenureSizeThreshold  直接晋升老生代的对象的大小，设置这个参数后，大于这个参数的对象直接在老生代分配。但是！！！请注意！！！PretenureSizeThreshold这个参数只对Serial和parNew有用，Parallel Scavenge不认识他。。

2.Serial+Serial Old(-XX:+UseSerialGC)

[GC[DefNew: 6995K->240K(9216K), 0.0044170 secs] 6995K->6384K(19456K), 0.0044380 secs] [Times: user=0.01 sys=0.01, real=0.01 secs] 

Heap

 def new generation   total 9216K, used 4666K [0x00000007f9a00000, 0x00000007fa400000, 0x00000007fa400000)

  eden space 8192K,  54% used [0x00000007f9a00000, 0x00000007f9e527b0, 0x00000007fa200000)

  from space 1024K,  23% used [0x00000007fa300000, 0x00000007fa33c198, 0x00000007fa400000)

  to   space 1024K,   0% used [0x00000007fa200000, 0x00000007fa200000, 0x00000007fa300000)

 tenured generation   total 10240K, used 6144K [0x00000007fa400000, 0x00000007fae00000, 0x00000007fae00000)

   the space 10240K,  60% used [0x00000007fa400000, 0x00000007faa00030, 0x00000007faa00200, 0x00000007fae00000)

 compacting perm gen  total 21248K, used 2614K [0x00000007fae00000, 0x00000007fc2c0000, 0x0000000800000000)

   the space 21248K,  12% used [0x00000007fae00000, 0x00000007fb08d840, 0x00000007fb08da00, 0x00000007fc2c0000)

No shared spaces configured.

这对组合就是中规中矩的，Eden发现4MB放不下了，触发MinorGC,GC期间发现3个2MB全部无法放到Survivor里面，所以只好通过担保机制提前转移到老年代中去。然后Eden放进了4MB的对象。

3.ParNew + Serial Old(-XX:+UseParNewGC)

[GC[ParNew: 6995K->282K(9216K), 0.0039460 secs] 6995K->6426K(19456K), 0.0039770 secs] [Times: user=0.01 sys=0.01, real=0.00 secs] 

Heap

 par new generation   total 9216K, used 4708K [0x00000007f9a00000, 0x00000007fa400000, 0x00000007fa400000)

  eden space 8192K,  54% used [0x00000007f9a00000, 0x00000007f9e527b0, 0x00000007fa200000)

  from space 1024K,  27% used [0x00000007fa300000, 0x00000007fa346a50, 0x00000007fa400000)

  to   space 1024K,   0% used [0x00000007fa200000, 0x00000007fa200000, 0x00000007fa300000)

 tenured generation   total 10240K, used 6144K [0x00000007fa400000, 0x00000007fae00000, 0x00000007fae00000)

   the space 10240K,  60% used [0x00000007fa400000, 0x00000007faa00030, 0x00000007faa00200, 0x00000007fae00000)

 compacting perm gen  total 21248K, used 2614K [0x00000007fae00000, 0x00000007fc2c0000, 0x0000000800000000)

   the space 21248K,  12% used [0x00000007fae00000, 0x00000007fb08d840, 0x00000007fb08da00, 0x00000007fc2c0000)

No shared spaces configured.

跟上面的差不多。。。

4.Parallel Scavenge + Parallel Old(-XX:+UseParallelOldGC)

Heap

 PSYoungGen      total 9216K, used 7159K [0x00000007ff600000, 0x0000000800000000, 0x0000000800000000)

  eden space 8192K, 87% used [0x00000007ff600000,0x00000007ffcfde20,0x00000007ffe00000)

  from space 1024K, 0% used [0x00000007fff00000,0x00000007fff00000,0x0000000800000000)

  to   space 1024K, 0% used [0x00000007ffe00000,0x00000007ffe00000,0x00000007fff00000)

 ParOldGen       total 10240K, used 4096K [0x00000007fec00000, 0x00000007ff600000, 0x00000007ff600000)

  object space 10240K, 40% used [0x00000007fec00000,0x00000007ff000010,0x00000007ff600000)

 PSPermGen       total 21504K, used 2611K [0x00000007f9a00000, 0x00000007faf00000, 0x00000007fec00000)

  object space 21504K, 12% used [0x00000007f9a00000,0x00000007f9c8cc78,0x00000007faf00000)

parallel Scavenge。。。新分配的4MB去了老生代

5.ParNew + CMS + Serial Old(-XX:+UseConcMarkSweepGC)

[GC[ParNew: 7003K->289K(9216K), 0.0046330 secs] 7003K->6435K(19456K), 0.0046730 secs] [Times: user=0.02 sys=0.01, real=0.00 secs] 

[GC [1 CMS-initial-mark: 6146K(10240K)] 10531K(19456K), 0.0002410 secs] [Times: user=0.00 sys=0.00, real=0.00 secs] 

Heap

 par new generation   total 9216K, used 4715K [0x00000007f9a00000, 0x00000007fa400000, 0x00000007fa400000)

  eden space 8192K,  54% used [0x00000007f9a00000, 0x00000007f9e527a8, 0x00000007fa200000)

  from space 1024K,  28% used [0x00000007fa300000, 0x00000007fa348520, 0x00000007fa400000)

  to   space 1024K,   0% used [0x00000007fa200000, 0x00000007fa200000, 0x00000007fa300000)

 concurrent mark-sweep generation total 10240K, used 6146K [0x00000007fa400000, 0x00000007fae00000, 0x00000007fae00000)

 concurrent-mark-sweep perm gen total 21248K, used 2615K [0x00000007fae00000, 0x00000007fc2c0000, 0x0000000800000000)

6MB去了老生代，4MB在新生代，而且，持久代也用了CMS~~~

6.G1（-XX：+UseG1GC）

[GC pause (young) (initial-mark), 0.0014180 secs]

   [Parallel Time: 1.0 ms, GC Workers: 8]

      [GC Worker Start (ms): Min: 120.1, Avg: 120.2, Max: 120.6, Diff: 0.5]

      [Ext Root Scanning (ms): Min: 0.0, Avg: 0.2, Max: 0.4, Diff: 0.4, Sum: 1.9]

      [Code Root Marking (ms): Min: 0.0, Avg: 0.0, Max: 0.0, Diff: 0.0, Sum: 0.0]

      [Update RS (ms): Min: 0.0, Avg: 0.0, Max: 0.0, Diff: 0.0, Sum: 0.0]

         [Processed Buffers: Min: 0, Avg: 1.1, Max: 5, Diff: 5, Sum: 9]

      [Scan RS (ms): Min: 0.0, Avg: 0.0, Max: 0.0, Diff: 0.0, Sum: 0.0]

      [Code Root Scanning (ms): Min: 0.0, Avg: 0.0, Max: 0.0, Diff: 0.0, Sum: 0.0]

      [Object Copy (ms): Min: 0.4, Avg: 0.5, Max: 0.6, Diff: 0.2, Sum: 3.7]

      [Termination (ms): Min: 0.0, Avg: 0.1, Max: 0.1, Diff: 0.1, Sum: 0.6]

      [GC Worker Other (ms): Min: 0.0, Avg: 0.0, Max: 0.1, Diff: 0.1, Sum: 0.3]

      [GC Worker Total (ms): Min: 0.5, Avg: 0.8, Max: 1.0, Diff: 0.5, Sum: 6.6]

      [GC Worker End (ms): Min: 121.0, Avg: 121.1, Max: 121.1, Diff: 0.1]

   [Code Root Fixup: 0.0 ms]

   [Code Root Migration: 0.0 ms]

   [Clear CT: 0.1 ms]

   [Other: 0.3 ms]

      [Choose CSet: 0.0 ms]

      [Ref Proc: 0.2 ms]

      [Ref Enq: 0.0 ms]

      [Free CSet: 0.0 ms]

   [Eden: 1024.0K(10.0M)->0.0B(9216.0K) Survivors: 0.0B->1024.0K Heap: 6588.3K(20.0M)->6560.0K(20.0M)]

 [Times: user=0.01 sys=0.01, real=0.00 secs] 

[GC concurrent-root-region-scan-start] 第一阶段开始，初始标记

[GC concurrent-root-region-scan-end, 0.0002720 secs]初始标记完毕

[GC concurrent-mark-start]    并发标记

[GC concurrent-mark-end, 0.0000430 secs]

[GC remark [GC ref-proc, 0.0000610 secs], 0.0005240 secs]  重新标记

 [Times: user=0.00 sys=0.00, real=0.01 secs] 

[GC cleanup 10M->10M(20M), 0.0004140 secs]   筛选回收，就是清理

 [Times: user=0.00 sys=0.00, real=0.00 secs] 

Heap

 garbage-first heap   total 20480K, used 10656K [0x00000007f9a00000, 0x00000007fae00000, 0x00000007fae00000)

  region size 1024K, 2 young (2048K), 1 survivors (1024K)

 compacting perm gen  total 20480K, used 2614K [0x00000007fae00000, 0x00000007fc200000, 0x0000000800000000)

   the space 20480K,  12% used [0x00000007fae00000, 0x00000007fb08d840, 0x00000007fb08da00, 0x00000007fc200000)

No shared spaces configured.

其实我也知道这样测G1什么也测不出什么特性来，我就是好奇。。。玩玩~

首先，发现G1的时间很棒哎~其他的都是0.004.。。G1:0.001

region size 1024K

 Heap: 6588.3K(20.0M)->6560.0K(20.0M)。。。我很想知道我的4MB去哪了。。。

garbage-first heap   total 20480K, used 10656K 只能在这看到10M是全的。。。

这里我会更新的。

另外，由于-XX：+UseParallelGC和-XX:+UseParallelOldGC看上去效果差不多，大R解释：

只用UseParallelGC与用了UseParallelOldGC背后实际的collector不一样，前者在HotSpot VM内是PSMarkSweep，后者是PSParallelCompact。 

http://hllvm.group.iteye.com/group/topic/37945