JCTools, 场景特化的并发类库，某些程度上能够提高性能

同上一篇同样，在jmap -histo中发现MpscChunkedArrayQueue类的实例比较多，javadoc看了下，其原来是出自JC Tools，https://github.com/JCTools/JCTools。

经过官网，咱们能够看到，它针对实际的场景对JDK自带的并发队列作了一些特定的优化，以下：

Java Concurrency Tools for the JVM. This project aims to offer some concurrent data structures currently missing from the JDK:

• SPSC/MPSC/SPMC/MPMC variations for concurrent queues:

  • SPSC - Single Producer Single Consumer (Wait Free, bounded and unbounded)
  • MPSC - Multi Producer Single Consumer (Lock less, bounded and unbounded)
  • SPMC - Single Producer Multi Consumer (Lock less, bounded)
  • MPMC - Multi Producer Multi Consumer (Lock less, bounded)

• SPSC/MPSC linked array queues offer a balance between performance, allocation and footprint

• An expanded queue interface (MessagePassingQueue):

  • relaxedOffer/Peek/Poll: trade off conflated guarantee on full/empty queue state with improved performance.
  • drain/fill: batch read and write methods for increased throughput and reduced contention

不过总体浏览了下，文档较少，benchmark提供了用例但没有看到结果，真用到的时候，能够看下到底性能能够提高多少。有一文章对源码进行了分析（http://blog.csdn.net/youaremoon/article/details/50351929），不过说实话，不知道如今为何不少人喜欢share源码分析，仅仅只是源码分析，对场景化的适用条件以及原理和实现机制却是考虑的不多。举个例子，https://www.qcloud.com/community/article/904925001482373849?utm_source=Community&utm_medium=article904925001482373849&utm_campaign=kyzg就是一篇典型的文章，是为了说明看过源码了、仍是理解了。一般，我的认为理解如何实现、适用场景及优缺点是同等重要的，不然就只顾赶路、却不关心本身在哪里了。