tests/checkasm/ext/src/html_data/body.html - third_party/ffmpeg - Git at Google

 <body>
   <div class="content">
     <h1 class='title'>checkasm performance measurements</h1>
     <p class="no-print"><a href="#grokularation">want to understand this report?</a></p>
     <h1 id="overview"><a href="#overview">overview</a></h1>
     <div class="no-print">
       <button id="toggleOverview">Collapse All</button>
       <span class="overview-info">
         <select id="sort-overview" class="select">
           <option value="report-index">index</option>
           <option value="lex">name</option>
           <option value="colex">suffix</option>
           <option value="duration">time ascending</option>
           <option value="rev-duration">time descending</option>
         </select>
         <a href="#controls-explanation" class="info" title="click bar/label to zoom; x-axis to toggle logarithmic scale; background to reset">&#9432;</a>
         <a id="legend-toggle" class="chevron button"></a>
       </span>
     </div>
     <aside id="overview-charts"></aside>
     <main id="reports"></main>
   </div>

   <aside id="controls-explanation" class="explanation no-print">
     <h1><a href="#controls-explanation">controls</a></h1>

     <p>
       The overview chart can be controlled by clicking the following elements:
       <ul>
         <li><em>a bar or its label</em> zooms the x-axis to that bar</li>
         <li><em>the background</em> resets zoom to the entire chart</li>
         <li><em>the x-axis</em> toggles between linear and logarithmic scale</li>
         <li><em>the chevron</em> in the top-right toggles the the legend</li>
         <li><em>a group name in the legend</em> shows/hides that group</li>
       </ul>
     </p>

   </aside>

   <aside id="grokularation" class="explanation">

     <h1><a href="#grokularation">understanding this report</a></h1>

     <p>
       In this report, each function benchmarked by checkasm is assigned a section of its own.
       <span class="no-print">The charts in each section are active; if you hover your mouse over data points and annotations, you will see more details.</span>
     </p>

     <ul>
       <li>
         The chart on the left is a <a href="http://en.wikipedia.org/wiki/Kernel_density_estimation">kernel density estimate</a> (also known as a KDE) of time measurements.
         This graphs the probability of any given time measurement occurring. This function is estimated from a <a href="https://en.wikipedia.org/wiki/Log-normal_distribution">Log-normal
         distribution</a> fitted to the measurement data. The peak corresponds to the most likely execution time (point estimate).
       </li>

       <li>
         The chart on the right is the raw data from which the kernel density estimate is built.
         The <em>x</em>-axis indicates the number of loop iterations, while the <em>y</em>-axis shows measured execution time for the given number of loop iterations.
         The line behind the values is the linear regression estimate of execution time for a given number of iterations.
         Ideally, all measurements will be on (or very near) this line.
         The transparent area behind it shows the 95% confidence interval for the execution time estimate.
         Note that, for technical reasons, when <code>bench_run()</code> is called multiple times in the same
         function, only the last call's measurements are shown in this chart.
       </li>
     </ul>

     <p>
       Under the charts is a small table. They show the results of a log-space linear regression
       run on the measurements displayed in the right-hand chart.
     </p>

     <ul>
       <li>
         <em>Adjusted cycles</em> is the estimated number of CPU cycles taken by a single function
         call, after subtracting the overhead of the measurement loop and function call itself.
       </li>
       <li>
         <em>Adjusted time</em> is the estimated time taken by a single function call, after
         normalization by the estimated hardware timer tick frequency, and includes extra variance
         from the uncertainty of the tick rate estimation.
       </li>
       <li>
         <em>Raw cycles</em> and <em>Raw time</em> reflect the raw data collected from the underlying
         hardware counter, before subtracting the measurement overhead.
       </li>
     </ul>
   </aside>

   <footer>
     <div class="content">
       <h1 class="colophon-header">colophon</h1>
       <p>
         This report was created using the <a id="checkasm-version" href="https://code.videolan.org/videolan/checkasm">checkasm</a>
         assembly testing and performance analysis tool. The data analysis and report template was copied from
         the <a href=" http://hackage.haskell.org/package/criterion">Haskell criterion</a> library.
       </p>
     </div>
   </footer>
 </body>
	<body>
	<div class="content">
	<h1 class='title'>checkasm performance measurements</h1>
	<p class="no-print"><a href="#grokularation">want to understand this report?</a></p>
	<h1 id="overview"><a href="#overview">overview</a></h1>
	<div class="no-print">
	<button id="toggleOverview">Collapse All</button>
	<span class="overview-info">
	<select id="sort-overview" class="select">
	<option value="report-index">index</option>
	<option value="lex">name</option>
	<option value="colex">suffix</option>
	<option value="duration">time ascending</option>
	<option value="rev-duration">time descending</option>
	</select>
	<a href="#controls-explanation" class="info" title="click bar/label to zoom; x-axis to toggle logarithmic scale; background to reset">ⓘ</a>
	<a id="legend-toggle" class="chevron button"></a>
	</span>
	</div>
	<aside id="overview-charts"></aside>
	<main id="reports"></main>
	</div>

	<aside id="controls-explanation" class="explanation no-print">
	<h1><a href="#controls-explanation">controls</a></h1>

	<p>
	The overview chart can be controlled by clicking the following elements:
	<ul>
	<li><em>a bar or its label</em> zooms the x-axis to that bar</li>
	<li><em>the background</em> resets zoom to the entire chart</li>
	<li><em>the x-axis</em> toggles between linear and logarithmic scale</li>
	<li><em>the chevron</em> in the top-right toggles the the legend</li>
	<li><em>a group name in the legend</em> shows/hides that group</li>
	</ul>
	</p>

	</aside>

	<aside id="grokularation" class="explanation">

	<h1><a href="#grokularation">understanding this report</a></h1>

	<p>
	In this report, each function benchmarked by checkasm is assigned a section of its own.
	<span class="no-print">The charts in each section are active; if you hover your mouse over data points and annotations, you will see more details.</span>
	</p>

	<ul>
	<li>
	The chart on the left is a <a href="http://en.wikipedia.org/wiki/Kernel_density_estimation">kernel density estimate</a> (also known as a KDE) of time measurements.
	This graphs the probability of any given time measurement occurring. This function is estimated from a <a href="https://en.wikipedia.org/wiki/Log-normal_distribution">Log-normal
	distribution</a> fitted to the measurement data. The peak corresponds to the most likely execution time (point estimate).
	</li>

	<li>
	The chart on the right is the raw data from which the kernel density estimate is built.
	The <em>x</em>-axis indicates the number of loop iterations, while the <em>y</em>-axis shows measured execution time for the given number of loop iterations.
	The line behind the values is the linear regression estimate of execution time for a given number of iterations.
	Ideally, all measurements will be on (or very near) this line.
	The transparent area behind it shows the 95% confidence interval for the execution time estimate.
	Note that, for technical reasons, when <code>bench_run()</code> is called multiple times in the same
	function, only the last call's measurements are shown in this chart.
	</li>
	</ul>

	<p>
	Under the charts is a small table. They show the results of a log-space linear regression
	run on the measurements displayed in the right-hand chart.
	</p>

	<ul>
	<li>
	<em>Adjusted cycles</em> is the estimated number of CPU cycles taken by a single function
	call, after subtracting the overhead of the measurement loop and function call itself.
	</li>
	<li>
	<em>Adjusted time</em> is the estimated time taken by a single function call, after
	normalization by the estimated hardware timer tick frequency, and includes extra variance
	from the uncertainty of the tick rate estimation.
	</li>
	<li>
	<em>Raw cycles</em> and <em>Raw time</em> reflect the raw data collected from the underlying
	hardware counter, before subtracting the measurement overhead.
	</li>
	</ul>
	</aside>

	<footer>
	<div class="content">
	<h1 class="colophon-header">colophon</h1>
	<p>
	This report was created using the <a id="checkasm-version" href="https://code.videolan.org/videolan/checkasm">checkasm</a>
	assembly testing and performance analysis tool. The data analysis and report template was copied from
	the <a href=" http://hackage.haskell.org/package/criterion">Haskell criterion</a> library.
	</p>
	</div>
	</footer>
	</body>