分类： LINUX
2011-08-02 16:00:57
备注： 原文地址：
笔者英语不好，又没人翻译，只好自己动手，希望大家多提意见，我好及时修改，以免误导他人。本文仅供参考，笔者对使用者产生的任何后果，概不负责。 转载请注明出处！
正文：
The MySQL distribution comes with a rather comprehensive set of generic tests that have been bundled together so you can run them as a group and examine the results. The tests will do little to help you figure out whether a configuration change will speed up your application. But they're very helpful when used as a high-level benchmark, meaning they provide a good overall indication of how well one server performs relative to another.
MySQL分发包捆绑了一套相当全面的系统基准测试工具，这样你就可以单独运行他们并检查测试结果。每次测试的结果对于确定一个配置更改是否会加快您的应用程序帮助不大。但作为一个高层次的基准，他们非常有帮助，这意味着他们提供一个良好的整体性能指示，一台服务器性能相对于另一台的整体表现。
You can also run the tests individually if you'd like compare a subset of the results from several servers. If you're mainly interested in UPDATE speed, run one of the UPDATE-intensive tests a few times on each server.
如果你想比较几台服务器的性能，可以单独测试每台服务器，然后比较测试结果。如果你主要想测试更新速度，每台服务器上多做几次更新更密集的测试。
The benchmark suite can be used to test non-MySQL servers as well. According to the README, PostgreSQL, Solid, and mSQL have been tested. This may be helpful if you're trying to choose between MySQL and PostgreSQL. All the benchmark code is relatively generic Perl using the DBI and Benchmark modules. If needed, you can add support for nearly any database server that has a DBI driver (Oracle, Sybase, Informix, DB2, etc.). If you do so, be sure to look at the bench-init.pl for any global options you may need to add or change.
基准套件，也可以用来测试非MySQL服务器。根据自述文件，PostgreSQL，Solid，mSQL的已经过测试。这个测试会对你选择MySQL或PostgreSQL哪种做为数据库有所帮助。所有基准的代码是用相对通用的Perl语言编写,使用了DBI和基准模块。如果需要，您可以添加具有DBI驱动程序（ORACLE，SYBASE，Informix和DB2等）几乎所有的数据库服务器的支持。如果您这样做，一定要检查bench-init.pl文件，寻找任何您可能需要添加或更改的全局选项。
By running the benchmarks against several different servers, you'll get an idea of how much faster one server is than another. The tests are largely CPU-bound, but there are portions of the test that demand a lot of disk I/O (for short times). You'll likely find that the 2.4-GHz CPU doesn't necessarily make MySQL run twice as fast as the 1.2-GHz CPU.
通过对几个不同的服务器上运行基准测试，你会想要知道一台服务器比另一种是快多少。测试数据主要是针对CPU的，但也有部分的测试，需求大量的磁盘I / O交换（很短的时间）。您可能会发现，MySQL在2.4 - GHz CPU上的运行速度不一定能达到在1.2 GHz的CPU的运行速度的两倍。
The benchmark suite will not help you test the benefits of multi-CPU machines because the benchmark process is completely serialized. It executes one query after another, so MySQL will not benefit from the addition of a second CPU. To test that, you'll need to use MySQL super-smack or a home-grown solution. Both are covered in the following sections.
基准套件不能帮助您测试多核CPU的机器，因为基准测试的过程完全序列化。它执行了一个又一个的查询，所以MySQL不会从第二个CPU中受益。为了测试多核CPU的性能，你需要使用MySQL super-smack或一个手动设计的解决方案。下面的章节将会讲到这些。
To run the tests, use the run-all-tests script located in the sql-bench directory. Be sure to read the README in that directory. It provides a complete list of the command-line options you can use.
运行测试，使用位于sql-bench目录下的run-all-tests测试脚本。请务必阅读该目录中的自述。它提供了一个可以使用的命令行选项的完整列表。
