标准的能力 [英] capabilities of criterion

问题描述

我发现了一些使用标准的旧例子这里。它看起来好像回到了2009年，它支持用于绘制图形的命令行选项。

现在，当我导入标准0.6时，将其包装在基本应用程序中（以测试它）并说./application - 帮助它只给出垂直可用的选项，并没有任何关系到绘图。在我看来，就像在输出格式和功能方面，它在2009年的功能比2012年更多？

似乎有一个 criterion-to-html ，它创建了基本的html报告，但没有任何东西接近eg png输出显示在上面的网站或标准源数据包的示例目录中。

在哪里可以找到准确和最新的信息？

  ./ application  - 帮助
 
用法：application [OPTIONS] [BENCHMARKS] 
 -h， - ？ --help打印帮助，然后退出
 -G --no-gc不要在迭代之间收集垃圾
 -g --gc在迭代之间收集垃圾
 -I CI --ci = CI bootstrap置信区间
 -l --list仅列出基准测试名称列表
 -o FILENAME --output = FILENAME要写入
的报告文件-q --quiet print less output 
 --resamples = N个自举重采样数量，用于执行
 -s N --samples = N个要采集的采样数量
 -t FILENAME --template = FILENAME模板文件，使用
 -u FILENAME --summary = FILENAME产生所有结果的汇总CSV文件
 -V --version显示版本，然后退出
 -v --verbose打印更多输出
如果没有基准测试名称给出，全部运行
否则，基准测试通过前缀匹配运行
 
 $ b $ / ./application -k win 
错误：无法识别的选项`-k'
使用信息运行application --help
  




解决方案

由于monadic链接的评论 - 图表（通过图表库）不再在Criterion中受到支持。它提供的最好是CSV输出和 HTML输出，两者都非常有用。 CSV输出可以使用任何你想要的图（pgfplots，Chart，gnuplot等）来绘制，但是需要一些解析。 HTML非常好，如果你不想自动查找任何东西，你可以拍摄一个截图。

如果你有时间，那么也许你可以建立/上传一个读取Criterion CSV并通过Chart生成旧条形图的软件包（甚至更好的！）。

I found some old examples of the use of criterion here. It looks as if back in 2009 it was supporting command line options for drawing graphs.

Now, when I import criterion 0.6, wrap it in a basic application (to test it) and say ./application --help it gives only vert little available options and nothing related to drawing. It seems to me as if concerning output formats and capabilities it had more features in 2009 than it has in 2012?

There seems to be a criterion-to-html that creates basic html reports, but nothing that comes close to e.g. the png output show on the above website or in the examples directory of the criterion source packet.

Where can I find accurate and up to date information?

./application --help

Usage: application [OPTIONS] [BENCHMARKS]
  -h, -?       --help               print help, then exit
  -G           --no-gc              do not collect garbage between iterations
  -g           --gc                 collect garbage between iterations
  -I CI        --ci=CI              bootstrap confidence interval
  -l           --list               print only a list of benchmark names
  -o FILENAME  --output=FILENAME    report file to write to
  -q           --quiet              print less output
               --resamples=N        number of bootstrap resamples to perform
  -s N         --samples=N          number of samples to collect
  -t FILENAME  --template=FILENAME  template file to use
  -u FILENAME  --summary=FILENAME   produce a summary CSV file of all results
  -V           --version            display version, then exit
  -v           --verbose            print more output
If no benchmark names are given, all are run
Otherwise, benchmarks are run by prefix match


 ./application -k win
Error: unrecognized option `-k'
Run "application --help" for usage information

解决方案

As monadic linked in the comments - charting (via the Chart library) is no longer supported in Criterion. The best it supplies is CSV output and the HTML output, both are quite useful. The CSV output can be graphed using whatever you want (pgfplots, Chart, gnuplot, etc), but some parsing will be needed. The HTML is very nice and you could take a screen shot of that if you're not looking for anything automated.

If you have time then perhaps you can build/upload a package that reads Criterion CSVs and produces the old bar graphs via Chart (or even better ones!).

这篇关于标准的能力的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！