堆空间被分割成大约2048个区域。region大小是2的次幂，最小1M，最大不超过32M，它的大小是在jvm初始化的时候就计算出来了，一旦计算出来以后就固定不变了，这个值jdk1.7与jdk1.8中的算法是不一样的;
jdk7中region计算的源码：
http://hg.openjdk.java.net/jdk7/jdk7/hotspot/file/9b0ca45cd756/src/share/vm/gc_implementation/g1/heapRegion.cpp
计算公式如下：
// Minimum region size; we won't go lower than that.
// We might want to decrease this in the future, to deal with small
// heaps a bit more efficiently.
//最小的region大小
#define MIN_REGION_SIZE ( 1024 * 1024 )
// Maximum region size; we don't go higher than that. There's a good
// reason for having an upper bound. We don't want regions to get too
// large, otherwise cleanup's effectiveness would decrease as there
// will be fewer opportunities to find totally empty regions after
// marking.
//最大的region大小
#define MAX_REGION_SIZE ( 32 * 1024 * 1024 )
// The automatic region size calculation will try to have around this
// many regions in the heap (based on the min heap size).
//期望regions区域个数，只是期望，一般不会是2048
#define TARGET_REGION_NUMBER 2048
void HeapRegion::setup_heap_region_size(uintx min_heap_size) {
// region_size in bytes
uintx region_size = G1HeapRegionSize;
if (FLAG_IS_DEFAULT(G1HeapRegionSize)) {
// We base the automatic calculation on the min heap size. This
// can be problematic if the spread between min and max is quite
// wide, imagine -Xms128m -Xmx32g. But, if we decided it based on
// the max size, the region size might be way too large for the
// min size. Either way, some users might have to set the region
// size manually for some -Xms / -Xmx combos.
// region的计算公式 min_heap_size就是 -Xms参数指定的值，如果没有指定默认为0
region_size = MAX2(min_heap_size / TARGET_REGION_NUMBER,
(uintx) MIN_REGION_SIZE);
}
int region_size_log = log2_long((jlong) region_size);
// Recalculate the region size to make sure it's a power of
// 2. This means that region_size is the largest power of 2 that's
// <= what we've calculated so far.
region_size = ((uintx)1 << region_size_log);
// Now make sure that we don't go over or under our limits.
if (region_size < MIN_REGION_SIZE) {
region_size = MIN_REGION_SIZE;
} else if (region_size > MAX_REGION_SIZE) {
region_size = MAX_REGION_SIZE;
}
// And recalculate the log.
region_size_log = log2_long((jlong) region_size);说白了就是-Xms /2048 ,如果这个值小于1，则取值为1，大于32则取值32；其它值则取与2,4,8,16相近的；
jdk1.8，以及jdk9中的实现：
http://hg.openjdk.java.net/jdk8/jdk8/hotspot/file/87ee5ee27509/src/share/vm/gc_implementation/g1/heapRegion.cpp：
// Minimum region size; we won't go lower than that.
// We might want to decrease this in the future, to deal with small
// heaps a bit more efficiently.
#define MIN_REGION_SIZE ( 1024 * 1024 )
// Maximum region size; we don't go higher than that. There's a good
// reason for having an upper bound. We don't want regions to get too
// large, otherwise cleanup's effectiveness would decrease as there
// will be fewer opportunities to find totally empty regions after
// marking.
#define MAX_REGION_SIZE ( 32 * 1024 * 1024 )
// The automatic region size calculation will try to have around this
// many regions in the heap (based on the min heap size).
#define TARGET_REGION_NUMBER 2048
size_t HeapRegion::max_region_size() {
return (size_t)MAX_REGION_SIZE;
}
void HeapRegion::setup_heap_region_size(size_t initial_heap_size, size_t max_heap_size) {
uintx region_size = G1HeapRegionSize;
if (FLAG_IS_DEFAULT(G1HeapRegionSize)) {
//这里是代码的关键，(-Xms+-Xmx)/2/2048
size_t average_heap_size = (initial_heap_size + max_heap_size) / 2;
region_size = MAX2(average_heap_size / TARGET_REGION_NUMBER,
(uintx) MIN_REGION_SIZE);
}
int region_size_log = log2_long((jlong) region_size);
// Recalculate the region size to make sure it's a power of
// 2. This means that region_size is the largest power of 2 that's
// <= what we've calculated so far.
region_size = ((uintx)1 << region_size_log);
// Now make sure that we don't go over or under our limits.
if (region_size < MIN_REGION_SIZE) {
region_size = MIN_REGION_SIZE;
} else if (region_size > MAX_REGION_SIZE) {
region_size = MAX_REGION_SIZE;
}
// And recalculate the log.
region_size_log = log2_long((jlong) region_size);
// Now, set up the globals.
guarantee(LogOfHRGrainBytes == 0, "we should only set it once");
LogOfHRGrainBytes = region_size_log;
guarantee(LogOfHRGrainWords == 0, "we should only set it once");
LogOfHRGrainWords = LogOfHRGrainBytes - LogHeapWordSize;
guarantee(GrainBytes == 0, "we should only set it once");
// The cast to int is safe, given that we've bounded region_size by
// MIN_REGION_SIZE and MAX_REGION_SIZE.
GrainBytes = (size_t)region_size;
guarantee(GrainWords == 0, "we should only set it once");
GrainWords = GrainBytes >> LogHeapWordSize;
guarantee((size_t) 1 << LogOfHRGrainWords == GrainWords, "sanity");
guarantee(CardsPerRegion == 0, "we should only set it once");
CardsPerRegion = GrainBytes >> CardTableModRefBS::card_shift;
}
关键代码在这两句：
size_t average_heap_size = (initial_heap_size + max_heap_size) / 2;
region_size = MAX2(average_heap_size / TARGET_REGION_NUMBER,
(uintx) MIN_REGION_SIZE);
其实就是(-Xms+-Xmx)/2/2048 ，如果-Xms没有配置，那么就是0；
由此得出以下结论：
- region1.7计算方式与1.8，1.9是不同；
- region是在jvm启动的时候根据-Xms与-Xmx来确定的，它的范围在1M、2M、4M、8M、16M、32M；
- region的大小都是一样的，一经确定就不会改变，即使是humongous，只不过是多个连续的region区域；
- region数量并不是2048个，2048只是一个期望值，从计算方式可以看出来，除非我们将-Xms与-Xmx设置成一样大，并且大小必须大于2GB，小于64GB的情况下，这个值才能约等于2048，为什么是约等于？因为region大小必须是2的次幂，取值只能是1、2、4、8、16、32，如果不是正好等于这些值，会取一个近似的大小，如果内存很小2GB以下，region数量一定小于2048，如果内存很大64GB以上，则region数量会超过2048；
另外这里有个疑问，为什么要取2048？而不是其它值？期望大神能给我解答一下？
极大区域(Humongous regions)。这些区域被设计成保持标准区域大小的50%或者更大的对象。它们被保存在一个连续的区域集合里。最后，最后一个类型的区域就是堆空间里没有使用的区域
垃圾回收的区域有四种：
eden，survivor、old、Humongous 区域的回收；
使用G1垃圾回收方式有三种young、mixed、Full，young和mixed是G1回收器，Full是G1失败的情况下采用serial回收；
young会回收全部的eden和survivor区域；
mixed会回收全部的eden和survivor区域，同时会回收部分old 或Humongous区域；
Full gc则是G1回收失败的情况下，使用serial回收全部的区域，这时候我们会发现FullGc的次数会+1；
old和Humongous 回收都会跟随者GC pause (G1 Evacuation Pause) (mixed)
年轻带回收
2522.958: [GC pause (G1 Evacuation Pause) (young), 0.1063552 secs]
[Parallel Time: 98.4 ms, GC Workers: 8]
[GC Worker Start (ms): Min: 2522958.7, Avg: 2522958.8, Max: 2522958.9, Diff: 0.2]
[Ext Root Scanning (ms): Min: 0.6, Avg: 0.9, Max: 2.0, Diff: 1.4, Sum: 7.1]
[Update RS (ms): Min: 14.9, Avg: 16.4, Max: 16.9, Diff: 2.0, Sum: 131.3]
[Processed Buffers: Min: 16, Avg: 28.1, Max: 71, Diff: 55, Sum: 225]
[Scan RS (ms): Min: 32.6, Avg: 32.9, Max: 33.2, Diff: 0.6, Sum: 263.4]
[Code Root Scanning (ms): Min: 0.0, Avg: 0.0, Max: 0.0, Diff: 0.0, Sum: 0.1]
[Object Copy (ms): Min: 47.7, Avg: 47.8, Max: 47.8, Diff: 0.1, Sum: 382.0]
[Termination (ms): Min: 0.0, Avg: 0.1, Max: 0.1, Diff: 0.1, Sum: 0.8]
[GC Worker Other (ms): Min: 0.0, Avg: 0.1, Max: 0.1, Diff: 0.1, Sum: 0.7]
[GC Worker Total (ms): Min: 98.0, Avg: 98.2, Max: 98.3, Diff: 0.3, Sum: 785.4]
[GC Worker End (ms): Min: 2523056.9, Avg: 2523057.0, Max: 2523057.1, Diff: 0.1]
[Code Root Fixup: 0.1 ms]
[Code Root Migration: 0.1 ms]
[Code Root Purge: 0.0 ms]
[Clear CT: 0.9 ms]
[Other: 6.7 ms]
[Choose CSet: 0.0 ms]
[Ref Proc: 0.5 ms]
[Ref Enq: 0.0 ms]
[Redirty Cards: 1.7 ms]
[Free CSet: 3.5 ms]
[Eden: 1524.0M(1524.0M)->0.0B(1508.0M) Survivors: 72.0M->80.0M Heap: 6957.7M(8192.0M)->5441.7M(8192.0M)]会计算eden和survivor的大小被保存在清单统计信息中，用来辅助计算size，停顿时间目标之类的都会纳入考虑
老年代回收日志
//初始标记(stop the world事件)
//先执行一个常规的年轻代收集，然后标记那些可能有引用到年老代的对象的survivor区(g1中survivor区有很多，有引用到老年代对象的survivor区被称作根区，这是为了找老年代的gc root，jvm垃圾回收采用可达性分析算法，需要从gc root开始查找)
//还可能是: [GC pause (G1 Humongous Allocation) (young) (initial-mark), 0.0404763 secs]
155.623: [GC pause (G1 Evacuation Pause) (young) (initial-mark), 0.0720050 secs]
[Parallel Time: 67.9 ms, GC Workers: 8]
[GC Worker Start (ms): Min: 155623.0, Avg: 155623.1, Max: 155623.2, Diff: 0.2]
[Ext Root Scanning (ms): Min: 0.4, Avg: 0.8, Max: 3.7, Diff: 3.3, Sum: 6.6]
[Code Root Marking (ms): Min: 0.0, Avg: 1.3, Max: 3.6, Diff: 3.6, Sum: 10.1]
[Update RS (ms): Min: 16.6, Avg: 18.7, Max: 19.8, Diff: 3.2, Sum: 149.7]
[Processed Buffers: Min: 17, Avg: 35.6, Max: 81, Diff: 64, Sum: 285]
[Scan RS (ms): Min: 22.5, Avg: 23.0, Max: 23.3, Diff: 0.8, Sum: 184.0]
[Code Root Scanning (ms): Min: 0.0, Avg: 0.0, Max: 0.2, Diff: 0.2, Sum: 0.3]
[Object Copy (ms): Min: 23.6, Avg: 23.7, Max: 23.8, Diff: 0.2, Sum: 189.7]
[Termination (ms): Min: 0.0, Avg: 0.0, Max: 0.0, Diff: 0.0, Sum: 0.0]
[GC Worker Other (ms): Min: 0.0, Avg: 0.0, Max: 0.1, Diff: 0.1, Sum: 0.3]
[GC Worker Total (ms): Min: 67.5, Avg: 67.6, Max: 67.7, Diff: 0.2, Sum: 540.8]
[GC Worker End (ms): Min: 155690.7, Avg: 155690.7, Max: 155690.8, Diff: 0.1]
[Code Root Fixup: 0.2 ms]
[Code Root Migration: 0.5 ms]
[Code Root Purge: 0.0 ms]
[Clear CT: 0.3 ms]
[Other: 3.1 ms]
[Choose CSet: 0.0 ms]
[Ref Proc: 0.3 ms]
[Ref Enq: 0.0 ms]
[Redirty Cards: 1.2 ms]
[Free CSet: 0.9 ms]
[Eden: 580.0M(580.0M)->0.0B(576.0M) Survivors: 48.0M->48.0M Heap: 1291.0M(1512.0M)->715.8M(1512.0M)]
[Times: user=0.56 sys=0.00, real=0.08 secs]
//根区扫描
//扫描上一阶段标记的根区（也就是有引用指向到老年代的survivor区）,找到指向老年代的引用；这个发生在应用继续运行时。此阶段必须在下次年轻代收集执行前完成；
155.695: [GC concurrent-root-region-scan-start]
155.737: [GC concurrent-root-region-scan-end, 0.0418092 secs]
//并发标记阶段
//遍历整个堆寻找活跃对象，这个发生在应用运行时，这个阶段可以被年轻代垃圾回收打断。
155.737: [GC concurrent-mark-start]
156.295: [GC concurrent-mark-end, 0.5581926 secs]
//重新标记(stop the world事件)
//完全标记堆中的活跃对象，使用一个叫作snapshot-at-the-beginning(SATB)的比CMS收集器的更快的算法
156.296: [GC remark 156.296: [GC ref-proc, 0.0031235 secs], 0.0065449 secs]
[Times: user=0.02 sys=0.01, real=0.00 secs]
//清理(stop the world事件)
//在存活对象和完全空闲的区域上执行统计（accounting）操作，擦写Remembered Sets.；
156.303: [GC cleanup 997M->995M(1512M), 0.0064480 secs]
[Times: user=0.05 sys=0.00, real=0.01 secs]
//并发清理
//重置空heap区并将它们返还给空闲列表（free list）；
156.310: [GC concurrent-cleanup-start]
156.310: [GC concurrent-cleanup-end, 0.0000178 secs]
//(*)复制(stop the world事件)
//一般都是紧跟着一个young 复制，外加n个mixed复制操作
//这些是stop the world暂停为了疏散或者复制活跃对象到新的未使用的区域。这个可以由被记录为[GC Pause (young)]的年轻代区域或者被记录为[GC Pause (mixed)]年轻代和年老代区域完成
157.032: [GC pause (G1 Evacuation Pause) (young), 0.0661416 secs]
[Parallel Time: 62.2 ms, GC Workers: 8]
[GC Worker Start (ms): Min: 157032.1, Avg: 157032.1, Max: 157032.2, Diff: 0.1]
[Ext Root Scanning (ms): Min: 0.5, Avg: 0.7, Max: 1.8, Diff: 1.3, Sum: 5.9]
[Update RS (ms): Min: 17.1, Avg: 18.4, Max: 18.9, Diff: 1.8, Sum: 146.8]
[Processed Buffers: Min: 22, Avg: 36.8, Max: 51, Diff: 29, Sum: 294]
[Scan RS (ms): Min: 21.3, Avg: 21.7, Max: 21.8, Diff: 0.6, Sum: 173.7]
[Code Root Scanning (ms): Min: 0.0, Avg: 0.0, Max: 0.1, Diff: 0.1, Sum: 0.2]
[Object Copy (ms): Min: 21.0, Avg: 21.0, Max: 21.1, Diff: 0.1, Sum: 168.2]
[Termination (ms): Min: 0.0, Avg: 0.0, Max: 0.0, Diff: 0.0, Sum: 0.0]
[GC Worker Other (ms): Min: 0.0, Avg: 0.0, Max: 0.1, Diff: 0.1, Sum: 0.3]
[GC Worker Total (ms): Min: 61.8, Avg: 61.9, Max: 62.0, Diff: 0.1, Sum: 495.2]
[GC Worker End (ms): Min: 157094.0, Avg: 157094.0, Max: 157094.1, Diff: 0.1]
[Code Root Fixup: 0.2 ms]
[Code Root Migration: 0.6 ms]
[Code Root Purge: 0.0 ms]
[Clear CT: 0.3 ms]
[Other: 2.8 ms]
[Choose CSet: 0.0 ms]
[Ref Proc: 0.3 ms]
[Ref Enq: 0.0 ms]
[Redirty Cards: 1.1 ms]
[Free CSet: 0.9 ms]
[Eden: 576.0M(576.0M)->0.0B(28.0M) Survivors: 48.0M->46.0M Heap: 1289.8M(1512.0M)->716.1M(1512.0M)]
[Times: user=0.51 sys=0.00, real=0.07 secs]
157.126: [GC pause (G1 Evacuation Pause) (mixed), 0.1268152 secs]
[Parallel Time: 117.4 ms, GC Workers: 8]
[GC Worker Start (ms): Min: 157127.7, Avg: 157127.8, Max: 157127.9, Diff: 0.1]
[Ext Root Scanning (ms): Min: 0.6, Avg: 0.8, Max: 2.0, Diff: 1.5, Sum: 6.5]
[Update RS (ms): Min: 31.5, Avg: 32.9, Max: 33.1, Diff: 1.6, Sum: 263.0]
[Processed Buffers: Min: 43, Avg: 64.5, Max: 83, Diff: 40, Sum: 516]
[Scan RS (ms): Min: 34.5, Avg: 34.8, Max: 35.0, Diff: 0.4, Sum: 278.5]
[Code Root Scanning (ms): Min: 0.0, Avg: 0.1, Max: 0.4, Diff: 0.4, Sum: 0.8]
[Object Copy (ms): Min: 48.3, Avg: 48.4, Max: 48.5, Diff: 0.3, Sum: 387.0]
[Termination (ms): Min: 0.0, Avg: 0.1, Max: 0.2, Diff: 0.2, Sum: 1.0]
[GC Worker Other (ms): Min: 0.0, Avg: 0.0, Max: 0.0, Diff: 0.0, Sum: 0.1]
[GC Worker Total (ms): Min: 117.1, Avg: 117.1, Max: 117.2, Diff: 0.1, Sum: 936.9]
[GC Worker End (ms): Min: 157244.9, Avg: 157244.9, Max: 157244.9, Diff: 0.0]
[Code Root Fixup: 0.7 ms]
[Code Root Migration: 2.0 ms]
[Code Root Purge: 0.0 ms]
[Clear CT: 0.3 ms]
[Other: 6.4 ms]
[Choose CSet: 1.2 ms]
[Ref Proc: 0.2 ms]
[Ref Enq: 0.0 ms]
[Redirty Cards: 2.5 ms]
[Free CSet: 2.0 ms]
[Eden: 28.0M(28.0M)->0.0B(66.0M) Survivors: 46.0M->8192.0K Heap: 744.1M(1512.0M)->607.7M(1512.0M)]
[Times: user=0.96 sys=0.00, real=0.12 secs]
157.315: [GC pause (G1 Evacuation Pause) (mixed), 0.1369814 secs]
[Parallel Time: 131.4 ms, GC Workers: 8]
[GC Worker Start (ms): Min: 157315.8, Avg: 157315.8, Max: 157315.9, Diff: 0.1]
[Ext Root Scanning (ms): Min: 0.5, Avg: 0.8, Max: 2.2, Diff: 1.7, Sum: 6.2]
[Update RS (ms): Min: 83.5, Avg: 85.2, Max: 85.7, Diff: 2.1, Sum: 681.9]
[Processed Buffers: Min: 112, Avg: 135.2, Max: 155, Diff: 43, Sum: 1082]
[Scan RS (ms): Min: 26.1, Avg: 26.4, Max: 26.6, Diff: 0.4, Sum: 211.3]
[Code Root Scanning (ms): Min: 0.0, Avg: 0.0, Max: 0.2, Diff: 0.2, Sum: 0.4]
[Object Copy (ms): Min: 18.3, Avg: 18.5, Max: 18.6, Diff: 0.4, Sum: 147.9]
[Termination (ms): Min: 0.0, Avg: 0.1, Max: 0.1, Diff: 0.1, Sum: 0.8]
[GC Worker Other (ms): Min: 0.0, Avg: 0.0, Max: 0.0, Diff: 0.0, Sum: 0.1]
[GC Worker Total (ms): Min: 131.0, Avg: 131.1, Max: 131.1, Diff: 0.1, Sum: 1048.6]
[GC Worker End (ms): Min: 157446.9, Avg: 157446.9, Max: 157446.9, Diff: 0.0]
[Code Root Fixup: 0.4 ms]
[Code Root Migration: 1.2 ms]
[Code Root Purge: 0.0 ms]
[Clear CT: 0.3 ms]
[Other: 3.7 ms]
[Choose CSet: 0.7 ms]
[Ref Proc: 0.3 ms]
[Ref Enq: 0.0 ms]
[Redirty Cards: 1.2 ms]
[Free CSet: 1.1 ms]
[Eden: 66.0M(66.0M)->0.0B(752.0M) Survivors: 8192.0K->10.0M Heap: 673.7M(1512.0M)->579.6M(1512.0M)]
[Times: user=1.05 sys=0.00, real=0.13 secs]
扫一扫
227
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
