Title: [212129] trunk/PerformanceTests
Revision
212129
Author
fpi...@apple.com
Date
2017-02-10 11:49:55 -0800 (Fri, 10 Feb 2017)

Log Message

Beef up the ARES-6 explainer text
https://bugs.webkit.org/show_bug.cgi?id=167867

Reviewed by Saam Barati.
        
Added appropriate links.

Added text that explains the subscores.

* ARES-6/index.html:

Modified Paths

Diff

Modified: trunk/PerformanceTests/ARES-6/index.html (212128 => 212129)


--- trunk/PerformanceTests/ARES-6/index.html	2017-02-10 19:26:36 UTC (rev 212128)
+++ trunk/PerformanceTests/ARES-6/index.html	2017-02-10 19:49:55 UTC (rev 212129)
@@ -104,9 +104,13 @@
         
         <p>ARES-6 measures the execution time of _javascript_&#8217;s newest features, including symbols, for-of, arrow functions, Map/Set/WeakMap, let/const, classes, proxies, string interpolation, destructuring, default arguments, spread, tail calls, and generators.</p>
 
-        <p>Air is an ES2015 port of the <a href="" B3 JIT</a>&#8217;s <a href="" phase</a>. This code is a heavy user of Map, Set, classes, spread, and for-of. The benchmark runs <code>allocateStack</code> on hot function bodies from other popular _javascript_ benchmarks: <code>executeIteration</code> from Octane/Gameboy, <code>gaussianBlur</code> from Kraken, and <code>scanIdentifier</code> from Octane/Typescript. Because metacircularity, the benchmark also runs <code>allocateStack</code> on the largest hot function in Air.</p>
+        <p>Air is an ES2015 port of the <a href="" B3 JIT</a>&#8217;s <a href="" phase</a>. This code is a heavy user of Map, Set, classes, spread, and for-of. The benchmark runs <code>allocateStack</code> on hot function bodies from other popular _javascript_ benchmarks: <code>executeIteration</code> from <a href="" <code>gaussianBlur</code> from <a href="" and <code>scanIdentifier</code> from Octane/Typescript. Because metacircularity, the benchmark also runs <code>allocateStack</code> on the largest hot function in Air. <a href=""
 rg/browser/trunk/PerformanceTests/ARES-6/Air?rev=211697">Browse the source.</a></p>
 
-        <p>Basic is an <a href="" implementation of the ECMA-55 BASIC standard</a>. BASIC was an interactive language from the start, so our implementation uses generator functions for the entire interpreter. When the BASIC program says <code>INPUT</code>, our interpreter says <code>yield</code>. This allows users of the interpreter to easily integrate it into a runloop. Basic also uses classes, Map, and WeakMap. The benchmark runs a handful of simple programs, the most complex of which finds prime numbers.</p>
+        <p>Basic is an <a href="" implementation</a> of the <a href="" BASIC standard</a>. BASIC was an interactive language from the start, so our implementation uses generator functions for the entire interpreter. When the BASIC program says <code>INPUT</code>, our interpreter says <code>yield</code>. This allows users of the interpreter to easily integrate it into a runloop. Basic also uses classes, Map, and WeakMap. The benchmark runs a handful of simple programs, the most complex of which finds prime numbers. <a href="" the source.</a></p>
+        
+        <p>ARES-6 rewards browsers that start up quickly and run smoothly. It's not enough to just measure the total running time of a workload. Browsers may perform differently for the same workload depending on how many times it has run. Garbage collection runs periodically, making some iterations take longer than others. Code that runs repeatedly gets optimized by the browser, so the first iteration of any workload is more expensive than the rest. ARES-6 runs these benchmarks for 200 iterations and reports the execution time of the first iteration, the average of the worst 2% iterations, and the overall geometric mean. Each of these values is given an equal weight when computing the ovarall time. ARES-6 equally rewards fast start-up, low jank, and sophisticated adaptive optimizations for long-running code.</p>
+        
+        <p>Each ARES-6 sample has 200 iterations of Air and 200 iterations of Basic. ARES-6 runs 10 samples, and reports the average with 95% confidence intervals. Each sample runs in a fresh <code>iframe</code> to simulate some of the effects of page navigation.</p>
 
         <p>We like making _javascript_ run faster, but we didn&#8217;t have any ES2015 code to optimize. ARES-6 is a collection of sample ES2015 code written by the WebKit _javascript_ team, which we used to bootstrap our ES2015 optimization efforts.</p>
         
@@ -122,4 +126,4 @@
     <script src=""
         
 </body>
-</html>
\ No newline at end of file
+</html>

Modified: trunk/PerformanceTests/ChangeLog (212128 => 212129)


--- trunk/PerformanceTests/ChangeLog	2017-02-10 19:26:36 UTC (rev 212128)
+++ trunk/PerformanceTests/ChangeLog	2017-02-10 19:49:55 UTC (rev 212129)
@@ -1,3 +1,16 @@
+2017-02-05  Filip Pizlo  <fpi...@apple.com>
+
+        Beef up the ARES-6 explainer text
+        https://bugs.webkit.org/show_bug.cgi?id=167867
+
+        Reviewed by Saam Barati.
+        
+        Added appropriate links.
+
+        Added text that explains the subscores.
+
+        * ARES-6/index.html:
+
 2017-02-06  Jer Noble  <jer.no...@apple.com>
 
         Playback stalls when a SourceBuffer append causes frame eviction
_______________________________________________
webkit-changes mailing list
webkit-changes@lists.webkit.org
https://lists.webkit.org/mailman/listinfo/webkit-changes

Reply via email to