Disruptor-伪共享

Disruptor之所以快的原因之一是因为考虑内存的伪共享问题
伪共享测试地址：http://mechanical-sympathy.blogspot.com/2011/07/false-sharing.html
根据文章内容述说的内容是内存中有三级缓存，越接近cpu的缓存越快也越小，当两个线程同时更新数据，刚好两个数据都在同一个缓存行，更新一个数据，会引起读取另一个数据失效，会使数据来回获取保持更新，来回会耗时较大，所以使用填充沾满一个缓存行，使更新没那么频繁。

Disruptor怎么使用的

class LhsPadding
{
    protected long p1, p2, p3, p4, p5, p6, p7;
}

class Value extends LhsPadding
{
    protected volatile long value;
}

class RhsPadding extends Value
{
    protected long p9, p10, p11, p12, p13, p14, p15;
}

/**
 * 
Concurrent sequence class used for tracking the progress of
 * the ring buffer and event processors.  Support a number
 * of concurrent operations including CAS and order writes.
 *
 * 
Also attempts to be more efficient with regards to false
 * sharing by adding padding around the volatile field.
 */
public class Sequence extends RhsPadding
{
    static final long INITIAL_VALUE = -1L;
    private static final Unsafe UNSAFE;
    private static final long VALUE_OFFSET;

伪共享填充.png

左右填充的目的是为了使真是的数据正好处在一个缓存行内，而不受别的数据影响

应用场景
Log4J2