[EFL][GTK] Skip some IndexedDB tests that time out on the Perf bots.
[WebKit-https.git] / PerformanceTests / ChangeLog
1 2017-01-31  Carlos Alberto Lopez Perez  <clopez@igalia.com>
2
3         [EFL][GTK] Skip some IndexedDB tests that time out on the Perf bots.
4
5         Unreviewed performance test gardening.
6
7         * Skipped:
8
9 2017-01-30  Zalan Bujtas  <zalan@apple.com>
10
11         Simple line layout: Small tweaks to improve performance.
12         https://bugs.webkit.org/show_bug.cgi?id=167611
13         <rdar://problem/30274294>
14
15         Reviewed by Simon Fraser.
16
17         * Layout/simple-line-layout-non-repeating-text.html: Added.
18
19 2017-01-13  Said Abou-Hallawa  <sabouhallawa@apple.com>
20
21         26 MotionMark performance tests failing
22         https://bugs.webkit.org/show_bug.cgi?id=166854
23
24         Reviewed by Ryosuke Niwa.
25         
26         Skip running MotionMark on WebKit perf bots.
27
28         * Skipped:
29
30 2017-01-06  Said Abou-Hallawa  <sabouhallawa@apple.com>
31
32         Rename the directory and the files of MotionMark from Animometer to MotionMark
33         https://bugs.webkit.org/show_bug.cgi?id=166659
34
35         Reviewed by Jon Lee.
36         
37         Rename the directory and the files of the benchmark to its new name.
38
39         * MotionMark/developer.html: Renamed from PerformanceTests/Animometer/developer.html.
40         * MotionMark/index.html: Renamed from PerformanceTests/Animometer/index.html.
41         * MotionMark/resources/debug-runner/d3.min.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/d3.min.js.
42         * MotionMark/resources/debug-runner/graph.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/graph.js.
43         * MotionMark/resources/debug-runner/motionmark.css: Renamed from PerformanceTests/Animometer/resources/debug-runner/animometer.css.
44         * MotionMark/resources/debug-runner/motionmark.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/animometer.js.
45         * MotionMark/resources/debug-runner/tests.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/tests.js.
46         * MotionMark/resources/extensions.js: Renamed from PerformanceTests/Animometer/resources/extensions.js.
47         * MotionMark/resources/runner/benchmark-runner.js: Renamed from PerformanceTests/Animometer/resources/runner/benchmark-runner.js.
48         * MotionMark/resources/runner/crystal.svg: Renamed from PerformanceTests/Animometer/resources/runner/crystal.svg.
49         * MotionMark/resources/runner/lines.svg: Renamed from PerformanceTests/Animometer/resources/runner/lines.svg.
50         * MotionMark/resources/runner/logo.svg: Renamed from PerformanceTests/Animometer/resources/runner/logo.svg.
51         * MotionMark/resources/runner/motionmark.css: Renamed from PerformanceTests/Animometer/resources/runner/animometer.css.
52         * MotionMark/resources/runner/motionmark.js: Renamed from PerformanceTests/Animometer/resources/runner/animometer.js.
53         * MotionMark/resources/runner/tests.js: Renamed from PerformanceTests/Animometer/resources/runner/tests.js.
54         * MotionMark/resources/statistics.js: Renamed from PerformanceTests/Animometer/resources/statistics.js.
55         * MotionMark/resources/strings.js: Renamed from PerformanceTests/Animometer/resources/strings.js.
56         * MotionMark/tests/3d/resources/webgl.js: Renamed from PerformanceTests/Animometer/tests/3d/resources/webgl.js.
57         * MotionMark/tests/3d/webgl.html: Renamed from PerformanceTests/Animometer/tests/3d/webgl.html.
58         * MotionMark/tests/bouncing-particles/bouncing-canvas-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-canvas-images.html.
59         * MotionMark/tests/bouncing-particles/bouncing-canvas-shapes.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html.
60         * MotionMark/tests/bouncing-particles/bouncing-css-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-css-images.html.
61         * MotionMark/tests/bouncing-particles/bouncing-css-shapes.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-css-shapes.html.
62         * MotionMark/tests/bouncing-particles/bouncing-svg-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-svg-images.html.
63         * MotionMark/tests/bouncing-particles/bouncing-svg-shapes.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-svg-shapes.html.
64         * MotionMark/tests/bouncing-particles/bouncing-tagged-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-tagged-images.html.
65         * MotionMark/tests/bouncing-particles/resources/bouncing-canvas-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-canvas-images.js.
66         * MotionMark/tests/bouncing-particles/resources/bouncing-canvas-particles.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-canvas-particles.js.
67         * MotionMark/tests/bouncing-particles/resources/bouncing-canvas-shapes.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-canvas-shapes.js.
68         * MotionMark/tests/bouncing-particles/resources/bouncing-css-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-css-images.js.
69         * MotionMark/tests/bouncing-particles/resources/bouncing-css-shapes.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-css-shapes.js.
70         * MotionMark/tests/bouncing-particles/resources/bouncing-svg-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-svg-images.js.
71         * MotionMark/tests/bouncing-particles/resources/bouncing-svg-particles.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-svg-particles.js.
72         * MotionMark/tests/bouncing-particles/resources/bouncing-svg-shapes.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js.
73         * MotionMark/tests/bouncing-particles/resources/bouncing-tagged-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-tagged-images.js.
74         * MotionMark/tests/bouncing-particles/resources/image1.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image1.jpg.
75         * MotionMark/tests/bouncing-particles/resources/image2.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image2.jpg.
76         * MotionMark/tests/bouncing-particles/resources/image3.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image3.jpg.
77         * MotionMark/tests/bouncing-particles/resources/image4.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image4.jpg.
78         * MotionMark/tests/bouncing-particles/resources/image5.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image5.jpg.
79         * MotionMark/tests/dom/compositing-transforms.html: Renamed from PerformanceTests/Animometer/tests/dom/compositing-transforms.html.
80         * MotionMark/tests/dom/focus.html: Renamed from PerformanceTests/Animometer/tests/dom/focus.html.
81         * MotionMark/tests/dom/leaves.html: Renamed from PerformanceTests/Animometer/tests/dom/leaves.html.
82         * MotionMark/tests/dom/particles.html: Renamed from PerformanceTests/Animometer/tests/dom/particles.html.
83         * MotionMark/tests/dom/resources/compositing-transforms.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/compositing-transforms.js.
84         * MotionMark/tests/dom/resources/dom-particles.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/dom-particles.js.
85         * MotionMark/tests/dom/resources/focus.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/focus.js.
86         * MotionMark/tests/dom/resources/leaves.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/leaves.js.
87         * MotionMark/tests/master/canvas-stage.html: Renamed from PerformanceTests/Animometer/tests/master/canvas-stage.html.
88         * MotionMark/tests/master/focus.html: Renamed from PerformanceTests/Animometer/tests/master/focus.html.
89         * MotionMark/tests/master/image-data.html: Renamed from PerformanceTests/Animometer/tests/master/image-data.html.
90         * MotionMark/tests/master/leaves.html: Renamed from PerformanceTests/Animometer/tests/master/leaves.html.
91         * MotionMark/tests/master/multiply.html: Renamed from PerformanceTests/Animometer/tests/master/multiply.html.
92         * MotionMark/tests/master/resources/canvas-stage.js: Renamed from PerformanceTests/Animometer/tests/master/resources/canvas-stage.js.
93         * MotionMark/tests/master/resources/canvas-tests.js: Renamed from PerformanceTests/Animometer/tests/master/resources/canvas-tests.js.
94         * MotionMark/tests/master/resources/compass.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/compass.svg.
95         * MotionMark/tests/master/resources/compass100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/compass100.png.
96         * MotionMark/tests/master/resources/console.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/console.svg.
97         * MotionMark/tests/master/resources/console100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/console100.png.
98         * MotionMark/tests/master/resources/contribute.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/contribute.svg.
99         * MotionMark/tests/master/resources/contribute100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/contribute100.png.
100         * MotionMark/tests/master/resources/debugger.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/debugger.svg.
101         * MotionMark/tests/master/resources/debugger100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/debugger100.png.
102         * MotionMark/tests/master/resources/focus.js: Renamed from PerformanceTests/Animometer/tests/master/resources/focus.js.
103         * MotionMark/tests/master/resources/image-data.js: Renamed from PerformanceTests/Animometer/tests/master/resources/image-data.js.
104         * MotionMark/tests/master/resources/inspector.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/inspector.svg.
105         * MotionMark/tests/master/resources/inspector100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/inspector100.png.
106         * MotionMark/tests/master/resources/layout.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/layout.svg.
107         * MotionMark/tests/master/resources/layout100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/layout100.png.
108         * MotionMark/tests/master/resources/leaves.js: Renamed from PerformanceTests/Animometer/tests/master/resources/leaves.js.
109         * MotionMark/tests/master/resources/multiply.js: Renamed from PerformanceTests/Animometer/tests/master/resources/multiply.js.
110         * MotionMark/tests/master/resources/particles.js: Renamed from PerformanceTests/Animometer/tests/master/resources/particles.js.
111         * MotionMark/tests/master/resources/performance.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/performance.svg.
112         * MotionMark/tests/master/resources/performance100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/performance100.png.
113         * MotionMark/tests/master/resources/script.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/script.svg.
114         * MotionMark/tests/master/resources/script100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/script100.png.
115         * MotionMark/tests/master/resources/shortcuts.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/shortcuts.svg.
116         * MotionMark/tests/master/resources/shortcuts100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/shortcuts100.png.
117         * MotionMark/tests/master/resources/standards.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/standards.svg.
118         * MotionMark/tests/master/resources/standards100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/standards100.png.
119         * MotionMark/tests/master/resources/storage.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/storage.svg.
120         * MotionMark/tests/master/resources/storage100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/storage100.png.
121         * MotionMark/tests/master/resources/styles.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/styles.svg.
122         * MotionMark/tests/master/resources/styles100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/styles100.png.
123         * MotionMark/tests/master/resources/svg-particles.js: Renamed from PerformanceTests/Animometer/tests/master/resources/svg-particles.js.
124         * MotionMark/tests/master/resources/text.js: Renamed from PerformanceTests/Animometer/tests/master/resources/text.js.
125         * MotionMark/tests/master/resources/timeline.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/timeline.svg.
126         * MotionMark/tests/master/resources/timeline100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/timeline100.png.
127         * MotionMark/tests/master/svg-particles.html: Renamed from PerformanceTests/Animometer/tests/master/svg-particles.html.
128         * MotionMark/tests/master/text.html: Renamed from PerformanceTests/Animometer/tests/master/text.html.
129         * MotionMark/tests/resources/main.js: Renamed from PerformanceTests/Animometer/tests/resources/main.js.
130         * MotionMark/tests/resources/math.js: Renamed from PerformanceTests/Animometer/tests/resources/math.js.
131         * MotionMark/tests/resources/stage.css: Renamed from PerformanceTests/Animometer/tests/resources/stage.css.
132         * MotionMark/tests/resources/star.svg: Renamed from PerformanceTests/Animometer/tests/resources/star.svg.
133         * MotionMark/tests/resources/yin-yang.png: Renamed from PerformanceTests/Animometer/tests/resources/yin-yang.png.
134         * MotionMark/tests/resources/yin-yang.svg: Renamed from PerformanceTests/Animometer/tests/resources/yin-yang.svg.
135         * MotionMark/tests/simple/resources/simple-canvas-paths.js: Renamed from PerformanceTests/Animometer/tests/simple/resources/simple-canvas-paths.js.
136         * MotionMark/tests/simple/resources/simple-canvas.js: Renamed from PerformanceTests/Animometer/tests/simple/resources/simple-canvas.js.
137         * MotionMark/tests/simple/resources/tiled-canvas-image.js: Renamed from PerformanceTests/Animometer/tests/simple/resources/tiled-canvas-image.js.
138         * MotionMark/tests/simple/simple-canvas-paths.html: Renamed from PerformanceTests/Animometer/tests/simple/simple-canvas-paths.html.
139         * MotionMark/tests/simple/tiled-canvas-image.html: Renamed from PerformanceTests/Animometer/tests/simple/tiled-canvas-image.html.
140         * MotionMark/tests/template/resources/template-canvas.js: Renamed from PerformanceTests/Animometer/tests/template/resources/template-canvas.js.
141         * MotionMark/tests/template/resources/template-css.js: Renamed from PerformanceTests/Animometer/tests/template/resources/template-css.js.
142         * MotionMark/tests/template/resources/template-svg.js: Renamed from PerformanceTests/Animometer/tests/template/resources/template-svg.js.
143         * MotionMark/tests/template/template-canvas.html: Renamed from PerformanceTests/Animometer/tests/template/template-canvas.html.
144         * MotionMark/tests/template/template-css.html: Renamed from PerformanceTests/Animometer/tests/template/template-css.html.
145         * MotionMark/tests/template/template-svg.html: Renamed from PerformanceTests/Animometer/tests/template/template-svg.html.
146
147 2016-12-13  Brady Eidson  <beidson@apple.com>
148
149         Add a PerfTest targeting IDBObjectStore.get().
150         https://bugs.webkit.org/show_bug.cgi?id=165816
151
152         Reviewed by Alex Christensen.
153
154         * IndexedDB/objectstore-get.html: Added.
155
156 2016-12-13  Brady Eidson  <beidson@apple.com>
157
158         Add a PerfTest targeting IDBIndex.get().
159         https://bugs.webkit.org/show_bug.cgi?id=165803
160
161         Reviewed by Alex Christensen.
162
163         * IndexedDB/index-get.html: Added.
164
165 2016-12-12  Brady Eidson  <beidson@apple.com>
166
167         More IndexedDB perf tests.
168         https://bugs.webkit.org/show_bug.cgi?id=165634
169
170         Reviewed by Sam Weinig.
171
172         * IndexedDB/index-multientry.html:
173         * IndexedDB/large-number-of-inserts-responsiveness.html: Added.
174         * IndexedDB/large-number-of-inserts.html: Added.
175         * IndexedDB/objectstore-cursor.html: Added. Runtime cut in 1/4th compared to r209672.
176
177         * resources/runner.js:
178
179 2016-12-12  Commit Queue  <commit-queue@webkit.org>
180
181         Unreviewed, rolling out r209672.
182         https://bugs.webkit.org/show_bug.cgi?id=165766
183
184         IndexedDB/objectstore-cursor.html perf test always times out
185         (Requested by ap on #webkit).
186
187         Reverted changeset:
188
189         "More IndexedDB perf tests."
190         https://bugs.webkit.org/show_bug.cgi?id=165634
191         http://trac.webkit.org/changeset/209672
192
193 2016-12-10  Brady Eidson  <beidson@apple.com>
194
195         More IndexedDB perf tests.
196         https://bugs.webkit.org/show_bug.cgi?id=165634
197
198         Reviewed by Sam Weinig.
199
200         * IndexedDB/index-multientry.html:
201         * IndexedDB/large-number-of-inserts-responsiveness.html: Added.
202         * IndexedDB/large-number-of-inserts.html: Added.
203         * IndexedDB/objectstore-cursor.html: Added.
204         
205         * resources/runner.js: Add "track responsiveness" functionality to PerfTestRunner.
206
207 2016-12-06  Filip Pizlo  <fpizlo@apple.com>
208
209         Concurrent GC should be stable enough to land enabled
210         https://bugs.webkit.org/show_bug.cgi?id=164990
211
212         Reviewed by Geoffrey Garen.
213         
214         Made CDjs more configurable and refined the "large.js" configuration. I was using that one and
215         the new "long.js" configuration to tune concurrent eden GCs.
216         
217         Added a new way of running Splay in browser, which using chartjs to plot the execution times of
218         2000 iterations. This includes the minified chartjs.
219
220         * JetStream/Octane2/splay-detail.html: Added.
221         * JetStream/cdjs/benchmark.js:
222         (benchmarkImpl):
223         (benchmark):
224         * JetStream/cdjs/long.js: Added.
225
226 2016-12-07  Brady Eidson  <beidson@apple.com>
227
228         Add IDB perf tests stressing key size.
229         https://bugs.webkit.org/show_bug.cgi?id=165567
230
231         Reviewed by Alex Christensen.
232
233         * IndexedDB/large-array-keys.html: Added.
234         * IndexedDB/large-binary-keys.html: Added.
235         * IndexedDB/large-string-keys.html: Added.
236
237 2016-12-05  Brady Eidson  <beidson@apple.com>
238
239         Add an IndexedDB perf test to PerformanceTests.
240         https://bugs.webkit.org/show_bug.cgi?id=165430
241
242         Reviewed by Alex Christensen.
243
244         * IndexedDB/index-multientry.html: Added.
245
246 2016-12-02  Filip Pizlo  <fpizlo@apple.com>
247
248         ES6SampleBench should report an average for Steady State so that all of the numbers are comparable
249         https://bugs.webkit.org/show_bug.cgi?id=165325
250
251         Reviewed by Saam Barati.
252
253         This makes all of the numbers that ES6SampleBench reports comparable to each other: they all speak of the time
254         it took to run an iteration of something.
255
256         * ES6SampleBench/results.js:
257         (Results.prototype.reportResult):
258
259 2016-11-18  Jer Noble  <jer.noble@apple.com>
260
261         Add some performance tests for various aspects of media loading and playback.
262         https://bugs.webkit.org/show_bug.cgi?id=164977
263
264         Reviewed by Darin Adler.
265
266         Add new tests for a few aspects of media playback; namely: how quickly  media elements fire
267         the "canplaythrough" event when all data is immediately available, how quickly HLS playback
268         switches resolutions, and at what playback rate MSE-backed video can play without dropping
269         frames. Skip these tests by default, as they're not runnable on all ports, and they require
270         a webserver (run-webkit-httpd) to complete on macOS.
271
272         * Media/HLSCanPlayThrough.html: Added.
273         * Media/HLSGearChange.html: Added.
274         * Media/MSECanPlayThrough.html: Added.
275         * Media/MSEPlaybackRate.html: Added.
276         * Media/StandardCanPlayThrough.html: Added.
277         * Media/hls/1080p/iframe_index.m3u8: Added.
278         * Media/hls/1080p/prog_index.m3u8: Added.
279         * Media/hls/1080p/test.ts: Added.
280         * Media/hls/480p/iframe_index.m3u8: Added.
281         * Media/hls/480p/prog_index.m3u8: Added.
282         * Media/hls/480p/test.ts: Added.
283         * Media/hls/720p/iframe_index.m3u8: Added.
284         * Media/hls/720p/prog_index.m3u8: Added.
285         * Media/hls/720p/test.ts: Added.
286         * Media/hls/720p/test.ts.back: Added.
287         * Media/hls/index.m3u8: Added.
288         * Media/media-source-loader.js: Added.
289         (MediaSourceLoader):
290         (MediaSourceLoader.prototype.loadManifest):
291         (MediaSourceLoader.prototype.loadManifestSucceeded):
292         (MediaSourceLoader.prototype.loadManifestFailed):
293         (MediaSourceLoader.prototype.loadMediaData):
294         (MediaSourceLoader.prototype.loadMediaDataSucceeded):
295         (MediaSourceLoader.prototype.loadMediaDataFailed):
296         (MediaSourceLoader.prototype.get type):
297         (MediaSourceLoader.prototype.get duration):
298         (MediaSourceLoader.prototype.get initSegment):
299         (MediaSourceLoader.prototype.get mediaSegmentsLength):
300         (MediaSourceLoader.prototype.mediaSegments):
301         (MediaSourceLoader.prototype.get everyMediaSegment):
302         * Media/test-fragmented-video.json: Added.
303         * Media/test-fragmented-video.mp4: Added.
304         * Media/test.mp4: Added.
305         * Skipped:
306
307 2016-11-18  Filip Pizlo  <fpizlo@apple.com>
308
309         Concurrent GC should be able to run splay in debug mode and earley/raytrace in release mode with no perf regression
310         https://bugs.webkit.org/show_bug.cgi?id=164282
311
312         Reviewed by Geoffrey Garen and Oliver Hunt.
313         
314         CDjs is a fun benchmark for stressing concurrent GCs, but to really give the GC a good
315         workout you need to increase the amount of work that the test does. This adds a second
316         configuration of the benchmark that has more aircraft. It uses much more memory and causes us
317         to do more GCs and those GCs take longer.
318
319         * JetStream/cdjs/benchmark.js:
320         (benchmarkImpl):
321         (benchmark):
322         * JetStream/cdjs/large.js: Added.
323
324 2016-11-14  Filip Pizlo  <fpizlo@apple.com>
325
326         Unreviewed, revert unintended change.
327
328         * ES6SampleBench/Air/benchmark.js:
329
330 2016-10-03  Saam Barati  <sbarati@apple.com>
331
332         MapHash should speculate on the type of its child node
333         https://bugs.webkit.org/show_bug.cgi?id=161922
334
335         Reviewed by Filip Pizlo.
336
337         I gave the main generator a name so it's easier to see what
338         it is when using the sampling profiler.
339
340         * ES6SampleBench/Basic/ast.js:
341         (Basic.Program):
342
343 2016-09-19  Sergio Villar Senin  <svillar@igalia.com>
344
345         [css-grid] Remove the x2 computation of row sizes with indefinite heights
346         https://bugs.webkit.org/show_bug.cgi?id=162150
347
348         Reviewed by Darin Adler.
349
350         Added a new test case which checks the layout performance of grids inside other grids, i.e,
351         grids acting both as grid container and grid item.
352
353         * Layout/nested-grid.html: Added.
354
355 2016-09-09  Simon Fraser  <simon.fraser@apple.com>
356
357         Perf test Animation/css-accelerated-animation.html failing
358         https://bugs.webkit.org/show_bug.cgi?id=161795
359
360         Skip Animation/css-accelerated-animation.html since this only works as a MobileSafari
361         perf test.
362
363         * Skipped:
364
365 2016-09-08  Simon Fraser  <simon.fraser@apple.com>
366
367         Add a content-animation test with accelerated CSS animations
368         https://bugs.webkit.org/show_bug.cgi?id=161776
369
370         Reviewed by Dean Jackson.
371
372         Add a test that measures the frame rate of accelerated CSS animations.
373
374         This is like css-animation.html (which animates 'left' and 'top') but animates
375         transforms instead. In order to get animation in X and Y we make two nested elements
376         and set transformX() on one, and transformY() on the other.
377
378         * Animation/css-accelerated-animation.html: Added.
379
380 2016-08-23  Saam Barati  <sbarati@apple.com>
381
382         It should be easy to run ES6SampleBench from the jsc shell
383         https://bugs.webkit.org/show_bug.cgi?id=161085
384
385         Reviewed by Yusuke Suzuki.
386
387         This patch makes ES6 sample bench runnable through the `jsc` shell.
388         To do that, you need to be in the PerformanceTests/ES6SampleBench
389         directory and run `jsc cli.js`. To make this work, the benchmark
390         is now aware of being run in two modes: via the browser, and via
391         the shell. Each entry point will set a variable `isInBrowser`,
392         and the benchmark will do different things based on if that
393         variable is true or false.
394
395         * ES6SampleBench/Air/benchmark.js:
396         * ES6SampleBench/Air/stress-test.js:
397         * ES6SampleBench/Basic/benchmark.js:
398         (runBenchmark):
399         * ES6SampleBench/Basic/stress-test.js:
400         * ES6SampleBench/air_benchmark.js:
401         * ES6SampleBench/basic_benchmark.js:
402         * ES6SampleBench/cli.js: Added.
403         (return.doRun):
404         (makeBenchmarkRunner):
405         * ES6SampleBench/driver.js:
406         (Driver):
407         (Driver.prototype._recomputeSummary):
408         (Driver.prototype._iterate.window.setTimeout):
409         (Driver.prototype._iterate):
410         (Driver.prototype._updateIterations):
411         * ES6SampleBench/glue.js:
412         (reportResult):
413         * ES6SampleBench/index.html:
414         * ES6SampleBench/results.js:
415         (Results):
416         (Results.prototype.reportRunning):
417         (Results.prototype.reportDone):
418         (Results.prototype.reportError):
419         * ES6SampleBench/stats.js:
420         (Stats):
421         (Stats.prototype.toString):
422         (Stats.prototype._update):
423
424 2016-08-01  Said Abou-Hallawa  <sabouhallawa@apple.com>
425
426         Add an option to run the MotionMark tests with classic tile size (512x512)
427         https://bugs.webkit.org/show_bug.cgi?id=160371
428
429         Reviewed by Darin Adler.
430
431         Setting the body to large size (3000x3000) and overflow to scroll forces
432         the classic tile size (512x512). This option does not affect the canvas
433         tests because the canvas is displayed on one tile regardless of its size.
434
435         * Animometer/developer.html:
436         * Animometer/resources/debug-runner/animometer.css:
437         (body.showing-test-container.tiles-big):
438         (body.showing-test-container.tiles-classic):
439         * Animometer/resources/debug-runner/animometer.js:
440         * Animometer/resources/runner/animometer.js:
441         (window.benchmarkController.startBenchmark):
442
443 2016-06-29  Filip Pizlo  <fpizlo@apple.com>
444
445         Generators violate bytecode liveness validation
446         https://bugs.webkit.org/show_bug.cgi?id=159279
447
448         Reviewed by Yusuke Suzuki.
449
450         Add Basic to our test harness.
451
452         Also made some cosmetic changes to the benchmark harness.
453
454         * ES6SampleBench/Basic/basic-tests.yaml: Added.
455         * ES6SampleBench/Basic/stress-test.js: Added.
456         (preciseTime):
457         * ES6SampleBench/driver.js:
458         (Driver):
459         (Driver.prototype.start):
460         (Driver.prototype.reportError):
461         * ES6SampleBench/glue.js:
462         * ES6SampleBench/index.html:
463
464 2016-06-28  Filip Pizlo  <fpizlo@apple.com>
465
466         ES6SampleBench should have a harness
467         https://bugs.webkit.org/show_bug.cgi?id=159246
468
469         Reviewed by Saam Barati.
470         
471         This adds a simple web harness for ES6SampleBench. It runs Air and Basic 10 times for 200
472         iterations each and reports three metrics:
473         
474         First iteration: the time it takes for the first iteration to run. This is the first
475         iteration after the benchmark is loaded into the iframe, so it's representative of what
476         would happen if one of these workloads only ran for a short time.
477         
478         Worst 2%: of the last 199 iterations, the average of the worst 2% iterations. If code like
479         any of these workloads was used in an important event handler, you'd want that code to run
480         well in the worst case in addition to having great throughput.
481         
482         Steady state: the total of the last 199 iterations. This is representative of what would
483         happen if you ran code like this for a long time.
484         
485         The total score is the geomean of the firstIteration/worstCase/steadyState numbers of the
486         two benchmarks.
487         
488         The harness does statistics using Student's T-distribution confidence intervals.
489
490         * ES6SampleBench/Basic/benchmark.js:
491         (Benchmark):
492         * ES6SampleBench/air_benchmark.js: Added.
493         * ES6SampleBench/basic_benchmark.js: Added.
494         * ES6SampleBench/driver.js: Added.
495         (Driver):
496         (Driver.prototype.addBenchmark):
497         (Driver.prototype.start):
498         (Driver.prototype.reportResult):
499         (Driver.prototype.reportError):
500         (Driver.prototype._recomputeSummary.Geomean):
501         (Driver.prototype._recomputeSummary.Geomean.prototype.add):
502         (Driver.prototype._recomputeSummary.Geomean.prototype.get result):
503         (Driver.prototype._recomputeSummary):
504         (Driver.prototype._iterate):
505         (Driver.prototype._updateIterations):
506         * ES6SampleBench/glue.js: Added.
507         * ES6SampleBench/index.html: Added.
508         * ES6SampleBench/results.js: Added.
509         (Results):
510         (Results.prototype.get benchmark):
511         (Results.prototype.reset):
512         (Results.prototype.reportRunning):
513         (Results.prototype.reportDone):
514         (Results.prototype.reportResult.averageAbovePercentile):
515         (Results.prototype.reportResult):
516         (Results.prototype.reportError):
517         * ES6SampleBench/stats.js: Added.
518         (Stats):
519         (Stats.prototype.reset):
520         (Stats.prototype.add):
521         (Stats.prototype.get numIterations):
522         (Stats.prototype.valueForIteration):
523         (Stats.get result.tDist):
524         (Stats.prototype.get result):
525         (Stats.prototype.toString):
526         (Stats.prototype._update):
527         * ES6SampleBench/style.css: Added.
528         (body):
529         (body, th, tr):
530         (h1, h2, h3, h4):
531         (h1):
532         (h2):
533         (h3):
534         (hr):
535         (address):
536         (img):
537         (.underline):
538         (ol.loweralpha):
539         (ol.upperalpha):
540         (ol.lowerroman):
541         (ol.upperroman):
542         (ol.arabic):
543         (.banner-link:link, .banner-link:visited):
544         (:link, :visited):
545         (h4 :link, h4 :visited, h5 :link, h5 :visited, h6 :link, h6 :visited):
546         (.anchor:link, .anchor:visited):
547         (* > .anchor:link, * > .anchor:visited):
548         (span:hover .anchor):
549         (a.forbidden, span.forbidden):
550         (a.missing:hover):
551         (a.closed:link, a.closed:visited, span.closed):
552         (pre):
553         (div.code):
554         (div.code pre):
555         (dt):
556         (dd):
557         (dd:last-child):
558         (.site-logo):
559         (.site-logo .tagline):
560         (table):
561         (#contents):
562         (p):
563         (p:last-child):
564         (th):
565         (td):
566
567 2016-06-28  Jon Lee  <jonlee@apple.com>
568
569         Update focus test
570         https://bugs.webkit.org/show_bug.cgi?id=159242
571         rdar://problem/27070007
572
573         Reviewed by Dean Jackson.
574         Provisionally reviewed by Said Abou-Hallawa.
575
576         Move previous test to dom suite, and update the test for better reporting of frame rate, although
577         it uses a different rendering path.
578
579         * Animometer/resources/debug-runner/tests.js: Add to dom suite.
580         * Animometer/tests/dom/focus.html: Copied from PerformanceTests/Animometer/tests/master/focus.html.
581         * Animometer/tests/dom/resources/focus.js: Copied from PerformanceTests/Animometer/tests/master/resources/focus.js.
582         * Animometer/tests/master/focus.html: Remove center element.
583         * Animometer/tests/master/resources/focus.js: Use narrower size range with smaller particles. Remove the
584         container elements. Inline getBlurValue and getOpacityValue since they are only called once.
585
586 2016-06-28  Filip Pizlo  <fpizlo@apple.com>
587
588         Move Air.js and Basic into ES6SampleBench
589
590         Rubber stamped by Geoffrey Garen.
591         
592         We want to group our ES6 benchmarks together, and eventually, we want to give them a
593         common harness.
594
595         * Air.js: Removed.
596         * Air.js/README.md: Removed.
597         * Air.js/airjs-tests.yaml: Removed.
598         * Air.js/all.js: Removed.
599         * Air.js/allocate_stack.js: Removed.
600         * Air.js/arg.js: Removed.
601         * Air.js/basic_block.js: Removed.
602         * Air.js/benchmark.js: Removed.
603         * Air.js/code.js: Removed.
604         * Air.js/custom.js: Removed.
605         * Air.js/frequented_block.js: Removed.
606         * Air.js/insertion_set.js: Removed.
607         * Air.js/inst.js: Removed.
608         * Air.js/liveness.js: Removed.
609         * Air.js/make_dist.sh: Removed.
610         * Air.js/opcode.js: Removed.
611         * Air.js/payload-airjs-ACLj8C.js: Removed.
612         * Air.js/payload-gbemu-executeIteration.js: Removed.
613         * Air.js/payload-imaging-gaussian-blur-gaussianBlur.js: Removed.
614         * Air.js/payload-typescript-scanIdentifier.js: Removed.
615         * Air.js/reg.js: Removed.
616         * Air.js/stack_slot.js: Removed.
617         * Air.js/stress-test.js: Removed.
618         * Air.js/strip-hash.rb: Removed.
619         * Air.js/symbols.js: Removed.
620         * Air.js/test.html: Removed.
621         * Air.js/test.js: Removed.
622         * Air.js/tmp.js: Removed.
623         * Air.js/tmp_base.js: Removed.
624         * Air.js/util.js: Removed.
625         * Basic: Removed.
626         * Basic/ast.js: Removed.
627         * Basic/basic.js: Removed.
628         * Basic/benchmark.js: Removed.
629         * Basic/caseless_map.js: Removed.
630         * Basic/lexer.js: Removed.
631         * Basic/number.js: Removed.
632         * Basic/parser.js: Removed.
633         * Basic/random.js: Removed.
634         * Basic/state.js: Removed.
635         * Basic/test.html: Removed.
636         * Basic/test.js: Removed.
637         * Basic/util.js: Removed.
638         * ES6SampleBench: Added.
639         * ES6SampleBench/Air: Copied from PerformanceTests/Air.js.
640         * ES6SampleBench/Basic: Copied from PerformanceTests/Basic.
641         * Skipped:
642
643 2016-06-24  Filip Pizlo  <fpizlo@apple.com>
644
645         Add a ES6 generator benchmark
646         https://bugs.webkit.org/show_bug.cgi?id=159101
647
648         Rubber stamped by Keith Miller.
649         
650         This adds a Basic interpreter loosely based on ECMA-55:
651         
652         http://www.ecma-international.org/publications/files/ECMA-ST-WITHDRAWN/ECMA-55,%201st%20Edition,%20January%201978.pdf
653         
654         It includes a lexer that is a generator, a parser that uses regular expressions, and an AST
655         walk interpreter where the walking functions for statements are generators that call each
656         other with yield*. This enables the interpreter to look like an AST walk even though it can
657         yield at INPUT and PRINT statements.
658         
659         This also uses for-of, classes, Map, and WeakMap. It also uses deprecated-but-awesome RegExp
660         features like RegExp.lastMatch and RegExp.rightContext. I did it that way because this is
661         how I've always written lexers in dynamic languages; see offlineasm's lex() method in
662         parser.rb for example.
663         
664         The benchmark runs a handful of simple Basic programs. The longest-running one computes
665         prime numbers.
666         
667         Includes a command-line and web harness. On my machine it runs in 2-3 seconds.
668
669         * Basic: Added.
670         * Basic/ast.js: Added.
671         (Basic.NumberApply):
672         (Basic.Variable):
673         (Basic.Const):
674         (Basic.NumberPow):
675         (Basic.NumberMul):
676         (Basic.NumberDiv):
677         (Basic.NumberNeg):
678         (Basic.NumberAdd):
679         (Basic.NumberSub):
680         (Basic.StringVar):
681         (Basic.Equals):
682         (Basic.NotEquals):
683         (Basic.LessThan):
684         (Basic.GreaterThan):
685         (Basic.LessEqual):
686         (Basic.GreaterEqual):
687         (Basic.GoTo):
688         (Basic.GoSub):
689         (Basic.Def):
690         (Basic.Let):
691         (Basic.If):
692         (Basic.Return):
693         (Basic.Stop):
694         (Basic.On):
695         (sideState.shouldStop):
696         (Basic.For):
697         (Basic.Next):
698         (Basic.Print):
699         (Basic.Input):
700         (Basic.Read):
701         (Basic.Restore):
702         (Basic.Dim):
703         (Basic.Randomize):
704         (Basic.End):
705         (Basic.Program):
706         * Basic/basic.js: Added.
707         (prepare):
708         (simulate):
709         * Basic/benchmark.js: Added.
710         (Benchmark):
711         (Benchmark.prototype.runIteration.expect):
712         (Benchmark.prototype.runIteration):
713         (runBenchmark):
714         * Basic/caseless_map.js: Added.
715         (CaselessMap):
716         * Basic/lexer.js: Added.
717         (lex.consumeWhitespace):
718         (lex.consume):
719         (lex):
720         * Basic/lexer_test.js: Added.
721         * Basic/number.js: Added.
722         (NumberValue):
723         (NumberValue.prototype.get value):
724         (NumberValue.prototype.apply):
725         (NumberValue.prototype.leftApply):
726         (NumberValue.prototype.assign):
727         (NumberArray.):
728         (NumberArray):
729         (NumberArray.prototype.apply):
730         (NumberArray.prototype.leftApply):
731         (NumberFunction):
732         (NumberFunction.prototype.apply):
733         (NumberFunction.prototype.leftApply):
734         (NativeFunction):
735         (NativeFunction.prototype.apply):
736         (NativeFunction.prototype.leftApply):
737         * Basic/parser.js: Added.
738         (parse):
739         (parse.pushToken):
740         (parse.peekToken):
741         (parse.consumeKind):
742         (parse.consumeToken):
743         (parse.parseVariable):
744         (parse.parseNumericExpression.parsePrimary):
745         (parse.parseNumericExpression.parseFactor):
746         (parse.parseNumericExpression.parseTerm):
747         (parse.parseNumericExpression):
748         (parse.parseConstant):
749         (parse.parseStringExpression):
750         (parse.isStringExpression):
751         (parse.parseRelationalExpression):
752         (parse.parseNonNegativeInteger):
753         (parse.parseGoToStatement):
754         (parse.parseGoSubStatement):
755         (parse.parseStatement):
756         (parse.parseStatements):
757         * Basic/random.js: Added.
758         (createRNG):
759         (createRNGWithFixedSeed):
760         (createRNGWithRandomSeed):
761         * Basic/state.js: Added.
762         (State):
763         (State.prototype.getValue):
764         (State.prototype.getSideState):
765         (State.prototype.abort):
766         (State.prototype.validate):
767         * Basic/test.html: Added.
768         * Basic/test.js: Added.
769         * Basic/util.js: Added.
770         (this.performance.performance.now.currentTime):
771         (else.this.preciseTime.currentTime):
772         (else.currentTime):
773         * Skipped: Make sure that we don't run Basic yet.
774
775 2016-06-21  Jon Lee  <jonlee@apple.com>
776
777         Update canvas size when benchmark begins
778         https://bugs.webkit.org/show_bug.cgi?id=159010
779
780         Reviewed by Dean Jackson.
781
782         Previously we would evaluate the media queries and assign the canvas size to the test
783         on body.onload. Instead, do it once the user starts the benchmark.
784
785         * Animometer/resources/debug-runner/animometer.js: Move benchmarkController.determineCanvasSize()
786         to benchmarkController._startBenchmark.
787         * Animometer/resources/runner/animometer.js: Ditto.
788
789 2016-06-21  Jon Lee  <jonlee@apple.com>
790
791         Add new timestamp option
792         https://bugs.webkit.org/show_bug.cgi?id=159006
793
794         Reviewed by Dean Jackson.
795
796         Add a new option to take timestamps from the rAF callback.
797
798         * Animometer/developer.html: Added option.
799         * Animometer/tests/resources/main.js:
800         (Benchmark): Fall back to using the rAF timestamp if performance.now()
801         is not available.
802         (Benchmark._animateLoop): Update how timestamp is set. Prefer to use
803         local var instead of accessing private var.
804
805 2016-06-21  Jon Lee  <jonlee@apple.com>
806
807         Improvements to Animometer benchmark
808         https://bugs.webkit.org/show_bug.cgi?id=157738
809
810         Reviewed by Dean Jackson.
811         Provisionally reviewed by Said Abou-Hallawa.
812
813         Update tests.
814
815         * Animometer/tests/master/text.html: Ensure only three text sizes for
816         the three canvases.
817         * Animometer/tests/master/focus.html: Reduce the text size to fit with smaller
818         particle sizes.
819         * Animometer/tests/master/resources/focus.js: Remove the quadratic distribution for
820         particle sizes, and make it linear. Reduce the size variance. Shuffle the math to
821         reduce some calculations per frame. Fix the placement of the particles which might
822         otherwise be culled.
823         * Animometer/tests/master/resources/image-data.js: Reduce the particle size
824         to encourage larger scores.
825
826 2016-06-21  Jon Lee  <jonlee@apple.com>
827
828         Improvements to Animometer benchmark
829         https://bugs.webkit.org/show_bug.cgi?id=157738
830
831         Reviewed by Dean Jackson.
832         Provisionally reviewed by Said Abou-Hallawa.
833
834         Include confidence interval for the final score, and store the canvas
835         size in the serialization so that it is accurately shown in results.
836
837         * Animometer/developer.html: Add a "confidence" div.
838         * Animometer/index.html: Ditto. Convert "mean" to "confidence".
839         * Animometer/resources/debug-runner/animometer.js: Look at options, and
840         if the configuration is included, update the body class based on it
841         (similar to what we do when we first load the page). That way, if you
842         drag-and-drop previous results in, that configuration is reflected in
843         the dashboard. Show the full confidence interval.
844
845         * Animometer/resources/debug-runner/animometer.css:
846         * Animometer/resources/debug-runner/animometer.js: Style update.
847         * Animometer/resources/runner/animometer.css:
848
849         * Animometer/resources/runner/animometer.js:
850         (_processData): Propagate the confidence interval values out and calculate
851         the lower and upper bounds. For now, shortcut the aggregate calculation,
852         since we only go through one iteration.
853         (this._processData.calculateScore): Propagate the confidence interval out
854         to be next to the score. Depending on the controller these values are
855         calculated differently.
856         (this._processData._getResultsProperty): Convenience function.
857         (this._processData.get score): Refactor.
858         (this._processData.get scoreLowerBound): Get the aggregate lower bound.
859         (this._processData.get scoreUpperBound): Get the aggregate upper bound.
860         (window.sectionsManager.setSectionScore):
861         (window.benchmarkController._startBenchmark): When the benchmark starts, note
862         the canvas size and add it to options. That way it will be automatically be
863         serialized.
864         (window.benchmarkController.showResults): Include the maximum deviation
865         percentage.
866         * Animometer/resources/runner/lines.svg: Make the background line up with the
867         skew.
868         * Animometer/resources/runner/tests.js:
869         (Headers.details.text): Refactor.
870         * Animometer/resources/statistics.js:
871         (Statistics.largestDeviationPercentage): Convenience function to calculate
872         the deviation percentage on either end and return the largest deviation.
873         * Animometer/resources/strings.js:
874
875         Allow specifying a regression profile to use instead of taking the one
876         with the lowest error.
877
878         Address an issue in the focus test when the regression calculated ends
879         up overestimating the change point, causing a cascade of tougher
880         ramps. The reason behind this is that at max complexity of an initial
881         ramp, the frame length is very high, and it influences the second
882         segment of the piecewise regression strongly, causing it to be very
883         steep. As a result, the first segment, expected to be flat, ends up
884         covering a higher range of complexity. That makes the change point
885         much higher than it should be. To avoid this, we will add a sanity
886         check on the maximum value of the ramp. If the regression's projected
887         value at the maximum complexity of the current ramp is very slow (less
888         than 20 fps), 1) reduce the maximum complexity by 20%, and 2) do not
889         include the regression's change point in the change point estimator.
890         That estimator is used as the midpoint of the next ramp, and including
891         the change point from a poor regression can bake in the error. The
892         controller already knows how to adjust for ramps that are too easy for
893         the system.
894
895         * Animometer/resources/runner/animometer.js:
896         (this._processData.findRegression): Takes a preferred profile and gives that to
897         Regression.
898         (this._processData.calculateScore): With the ramp controller, take the profile
899         of the ramp that was used the most when calculating the ramp's regression. That
900         profile is what is used for the test's score.
901         * Animometer/resources/statistics.js:
902         (Regression.Utilities.createClass): Update to take an options object which can
903         specify a profile to calculate with. Otherwise it will continue to use both and
904         select the one with the lower error.
905         (_calculateRegression): Fix an issue where we claim 0 error if the regression
906         calculation fails due to divide-by-zero. Instead reject that regression calculation
907         by giving it Number.MAX_VALUE.
908         * Animometer/resources/strings.js: New strings for marking a regression as flat
909         or slope.
910         * Animometer/tests/resources/main.js:
911         (RampController): Rename the thresholds for clarity. Add a threshold that, if
912         exceeded, will lower the maximum complexity of the next ramp.
913         (tune): Relax the cdf check to consider whether the interval definitely falls in
914         the desired frame length threshold.
915         (processSamples): Include the profile in the ramp.
916
917         Update ramp controller test. Increase the length of the test to 30 seconds, and extend
918         the interval to 120 ms during sampling. Improve the estimation of the ramp parameters.
919
920         * Animometer/developer.html: Change default to 30 seconds, and don't show the progress bar
921         by default.
922         * Animometer/resources/runner/animometer.js: Change default to 30 seconds.
923         * Animometer/tests/resources/main.js: A number of improvements to the ramp controller, in
924         the order in which they appear in the patch:
925
926         - With a longer test length use longer ramps with longer intervals to get more data at each
927         complexity. Keep the 100 ms interval length during the ramp up phase since we don't need to
928         spend more time there to find the right order of magnitude, but increase it during the
929         ramps to 120 ms.
930         - The ramp linearly interpolates the complexity to render based on its timestamp, but it would
931         never sample the minimum complexity. Instead of lerping max to min complexity from time
932         0 to t where t is the ramp length, instead lerp from 0 to (t - intervalSampleLength) so that
933         we can have at least one interval sample at the min complexity for that ramp.
934         - Some regression calculations only come out with one line segment rather than the two
935         we expect. This could be due to a noisy ramp or the ramp's range is too narrow. If that's the
936         case, influence the minimum complexity of the next ramp towards the lowest bound of 1, so that
937         we ensure that at least part of the ramp is covering a complexity range that the system can
938         handle at full 60.
939         - Remove an assignment to interpolatedFrameLength since that is never subsequently used.
940
941         Update the format used to serialize the results for analysis.
942
943         Each data point used to have named properties for fields like complexity and frame rate.
944         In addition the serialized numbers had rounding errors that took up many characters.
945         Update the format by introducing a new data container called SampleData, which contains a
946         field map. The map maps a string to an array index. Each data point is an array, so, to
947         get a stat, use the field map to get the array index into the data point. This allows future
948         versions to track other data, and reduces the size of the output string by two-thirds.
949
950         * Animometer/resources/extensions.js:
951         (Utilities.toFixedNumber): Add convenience function that truncates the number to a fixed
952         precision, and converts it back to a number.
953         (SampleData): New class that contains sample data and a field map that maps properties to
954         an array index.
955         (get length): Number of data points.
956         (addField): Add a field to the field map.
957         (push): Add a data point.
958         (sort): Sort the data.
959         (slice): Return new SampleData object with sliced data.
960         (forEach): Iterate over the data with provided function.
961         (createDatum):
962         (getFieldInDatum): Returns the data point associated with the field name by looking it up
963         in the field map in the datum provided, which can be the datum object itself (an array) or
964         an index into the data member variable.
965         (setFieldInDatum): Sets the data point associated with the field name.
966         (at): Returns the data point at the provided index.
967         (toArray): Serializes the data where the field map serves as property names for each point.
968
969         * Animometer/resources/debug-runner/graph.js:
970         (updateGraphData): Remove unused _testData. Convert the data to the old array format for the
971         graph to use, since the old format was much easier to work with when displaying the graphs.
972         (onGraphTypeChanged): For some controllers, no alternative score or mean is provided.
973         * Animometer/resources/runner/animometer.css:
974         * Animometer/resources/runner/animometer.js: Refactor to use SampleData. Update JSON output
975         to only go to 3 digits of precision for purposes of reducing the data size.
976         * Animometer/resources/strings.js: Add new strings to put into the field maps.
977         * Animometer/tests/resources/main.js: Refactor to use SampleData.
978
979         * Animometer/developer.html:
980         * Animometer/index.html: Restructure results table for both pages. Add charset attribute to
981         tests.js include.
982         * Animometer/resources/debug-runner/animometer.css: Clear out styles from release runner.
983         * Animometer/resources/debug-runner/graph.js:
984         (onGraphTypeChanged): Update score and mean if bootstrap results are available from the
985         controller, since not all controllers do bootstrapping.
986         * Animometer/resources/debug-runner/tests.js: Update header text.
987         * Animometer/resources/runner/animometer.css: Include confidence interval in results.
988         * Animometer/resources/runner/animometer.js:
989         (ResultsTable._addHeader): Header contents can be HTML, so use innerHTML instead.
990         (ResultsTable._addBody): Add tbody element.
991         (ResultsTable._addTest): Allow a data cell to invoke a JS function to get its contents.
992         (window.benchmarkController.showResults): Add table that includes tests' confidence intervals.
993         * Animometer/resources/runner/tests.js:
994         (Headers.details.text): Add new details table that includes bootstrap confidence interval.
995         The interval can be asymmetric, but for simplicity, report the maximum deviation percentage
996         on either side of the bootstrap median.
997         * Animometer/resources/statistics.js:
998         (bootstrap): Include the confidence percentage in the return object.
999
1000         Report canvas size in results.
1001
1002         * Animometer/developer.html: Add markup to indicate whether a small, medium, or large
1003         canvas was used.
1004         * Animometer/index.html: Ditto.
1005         * Animometer/resources/debug-runner/animometer.js: Call determineCanvasSize().
1006         * Animometer/resources/runner/animometer.css: Update styles to set the canvas based on the
1007         body class size.
1008         * Animometer/resources/runner/animometer.js:
1009         (window.benchmarkController.initialize): Update styles to set the canvas based on the
1010         body class size.
1011         (window.benchmarkController.determineCanvasSize): Run various media queries and set the body
1012         class based on the size of the device.
1013
1014         * Animometer/developer.html: Refactor to include the main CSS file, and redo
1015         the layout so that it doesn't rely on flexbox.
1016         * Animometer/resources/debug-runner/animometer.css:
1017         * Animometer/resources/debug-runner/animometer.js:
1018         (updateDisplay): Since various parts of the script alter the body class, we can't
1019         replace the className directly. Instead, remove all display-based values and then add
1020         the one that was selected.
1021         * Animometer/resources/debug-runner/graph.js:
1022         (updateGraphData): To set the size of the graph, use window.innerHeight.
1023         * Animometer/resources/runner/animometer.js:
1024         (window.sectionsManager.showSection): Since various parts of the script alter the body
1025         class, we can't replace the className directly. Remove all of the section classes
1026         individually and then add the one desired.
1027         * Animometer/tests/resources/stage.css: Remove -apple-system as a font to use in the
1028         stage.
1029
1030 2016-06-12  Filip Pizlo  <fpizlo@apple.com>
1031
1032         Fix round-down goof in Air.js's ShuffleCustom.forEachArg
1033         https://bugs.webkit.org/show_bug.cgi?id=158674
1034
1035         Reviewed by Michael Saboff.
1036         
1037         x / 3 * 3 is not how you round down to multiples of 3 in JavaScript. You need to do
1038         Math.floor(x / 3) * 3 instead.
1039         
1040         This is a benign change, because having extra arguments to a Shuffle is not something we
1041         actually take advantage of yet. But I think it's best to match the original C++ code's
1042         intent.
1043
1044         * Air.js/custom.js:
1045         (const.ShuffleCustom.forEachArg):
1046
1047 2016-06-11  Filip Pizlo  <fpizlo@apple.com>
1048
1049         Unreviewed, skip Air.js. It got unskipped when I renamed it from JSAir.
1050
1051         * Skipped:
1052
1053 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1054
1055         Make it easy to package Air.js
1056         https://bugs.webkit.org/show_bug.cgi?id=158652
1057
1058         Reviewed by Benjamin Poulain.
1059         
1060         If you want to give Air.js to someone, you can now do:
1061         
1062         cd PerformanceTests/Air.js
1063         ./make_dist.sh
1064         
1065         This will create Air.js.tar.gz and a directory called Air.js. If you want to send someone
1066         the benchmark, just send either of those things and tell them to look at index.html.
1067         
1068         You may have to edit make_dist.sh for wherever you have Markdown.pl. You can get that from
1069         https://daringfireball.net/projects/markdown/
1070
1071         * Air.js/make_dist.sh: Added.
1072         * Air.js/test.html:
1073
1074 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1075
1076         Unreviewed, make the link to test.html into an actual link.
1077
1078         * Air.js/README.md:
1079
1080 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1081
1082         Fix some text in Air.js/README.md
1083         https://bugs.webkit.org/show_bug.cgi?id=158650
1084
1085         Reviewed by Benjamin Poulain.
1086         
1087         I read the text again and found bugs:
1088         
1089         - We never actually say how to run the benchmark. This change adds a blurb about how to run
1090           it.
1091
1092         - We both say that allocateStack is responsible for the bulk of the running time and that
1093           we haven't measured where the bulk of the time is spent. This changes the text to say that
1094           it was a goal to make allocateStack be the hottest part of the benchmark, but that we did
1095           not measure this.
1096
1097         * Air.js/README.md:
1098
1099 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1100
1101         Air.js should have some documentation
1102         https://bugs.webkit.org/show_bug.cgi?id=158648
1103
1104         Reviewed by Keith Miller.
1105         
1106         I want to be able to point people at a document if they want to know more about this
1107         benchmark.
1108
1109         * Air.js/README.md: Added.
1110
1111 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1112
1113         Rename JSAir to Air.js.
1114
1115         Rubber stamped by Mark Lam.
1116
1117         * Air.js: Copied from PerformanceTests/JSAir.
1118         * Air.js/airjs-tests.yaml: Copied from PerformanceTests/JSAir/jsair-tests.yaml.
1119         * Air.js/benchmark.js: Copied from PerformanceTests/JSAir/benchmark.js.
1120         (Benchmark):
1121         * Air.js/jsair-tests.yaml: Removed.
1122         * Air.js/payload-airjs-ACLj8C.js: Copied from PerformanceTests/JSAir/payload-jsair-ACLj8C.js.
1123         * Air.js/payload-jsair-ACLj8C.js: Removed.
1124         * Air.js/stress-test.js: Copied from PerformanceTests/JSAir/stress-test.js.
1125         * Air.js/test.html: Copied from PerformanceTests/JSAir/test.html.
1126         * Air.js/test.js: Copied from PerformanceTests/JSAir/test.js.
1127         * JSAir: Removed.
1128         * JSAir/all.js: Removed.
1129         * JSAir/allocate_stack.js: Removed.
1130         * JSAir/arg.js: Removed.
1131         * JSAir/basic_block.js: Removed.
1132         * JSAir/benchmark.js: Removed.
1133         * JSAir/code.js: Removed.
1134         * JSAir/custom.js: Removed.
1135         * JSAir/frequented_block.js: Removed.
1136         * JSAir/insertion_set.js: Removed.
1137         * JSAir/inst.js: Removed.
1138         * JSAir/jsair-tests.yaml: Removed.
1139         * JSAir/liveness.js: Removed.
1140         * JSAir/opcode.js: Removed.
1141         * JSAir/payload-gbemu-executeIteration.js: Removed.
1142         * JSAir/payload-imaging-gaussian-blur-gaussianBlur.js: Removed.
1143         * JSAir/payload-jsair-ACLj8C.js: Removed.
1144         * JSAir/payload-typescript-scanIdentifier.js: Removed.
1145         * JSAir/reg.js: Removed.
1146         * JSAir/stack_slot.js: Removed.
1147         * JSAir/stress-test.js: Removed.
1148         * JSAir/strip-hash.rb: Removed.
1149         * JSAir/symbols.js: Removed.
1150         * JSAir/test.html: Removed.
1151         * JSAir/test.js: Removed.
1152         * JSAir/tmp.js: Removed.
1153         * JSAir/tmp_base.js: Removed.
1154         * JSAir/util.js: Removed.
1155
1156 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1157
1158         JSC Stress Test failing: jsair-tests.yaml/test.js.ftl-eager-no-cjit
1159         https://bugs.webkit.org/show_bug.cgi?id=158571
1160
1161         Reviewed by Keith Miller.
1162         
1163         Introduce a different harness for run-jsc-stress-tests, which runs for a minimum of 10
1164         iterations and then tries to do any number of "bonus" iterations until it's been running for
1165         two seconds. Since this is the sort of test that isn't really meant to stress anything in
1166         particular, I think it's OK if it is time-limited in this way. The worst case is that some
1167         of its failures will be flaky, but I think that they would have been flaky anyway given the
1168         complexity of the test.
1169
1170         * JSAir/benchmark.js:
1171         (Benchmark):
1172         (Benchmark.prototype.runIteration):
1173         (benchmark): Deleted.
1174         * JSAir/jsair-tests.yaml:
1175         * JSAir/stress-test.js: Added.
1176         (preciseTime):
1177         * JSAir/test.html:
1178         * JSAir/test.js:
1179
1180 2016-06-09  Filip Pizlo  <fpizlo@apple.com>
1181
1182         Unreviewed, teach the perf bots not to run JSAir.
1183
1184         * Skipped:
1185
1186 2016-06-08  Filip Pizlo  <fpizlo@apple.com>
1187
1188         Use more ES6 features in JSAir
1189         https://bugs.webkit.org/show_bug.cgi?id=158497
1190
1191         Reviewed by Keith Miller.
1192         
1193         This improves JSAir with the following ES6 features suggested by JoePeck:
1194         
1195         - String interpolation.
1196         - Destructuring inside PatchCustom.
1197         - Default arguments.
1198         
1199         All of these things are on hot paths.
1200         
1201         Note that I didn't use string interpolation everywhere that I could, only in those places
1202         where it made the code more readable. In Ruby, I used the style that if the interpolation
1203         expression has any non-trivial stuff (like a ternary operator, a chain of calls, or embedded
1204         strings) then it's better to use regular strcat. I think that's what I carried over to here.
1205         
1206         Note that the previous change (Add result validation to JSAir) also made the Proxy code not
1207         dead, though it's not necessarily on the hot path. The Proxy isn't called into frequently
1208         but it's used from a function that is otherwise hot, so if calling into the Proxy prevents
1209         that function from being optimized then it will hurt so good.
1210         
1211         I also reenabled tail calls in a few places.
1212         
1213         This change doesn't seem to change the performance of the benchmark for us. That's expected
1214         since these ES6 features are cheap. Note that this claim doesn't include Proxy, which was
1215         added in a separate change and that change did make the benchmark overall more expensive.
1216
1217         * JSAir/allocate_stack.js:
1218         (allocateStack):
1219         * JSAir/arg.js:
1220         (Arg.createBitImm64):
1221         (Arg.createAddr):
1222         (Arg.createStack):
1223         (Arg.logScale):
1224         (Arg.createIndex):
1225         * JSAir/basic_block.js:
1226         (BasicBlock.get headerString):
1227         (BasicBlock.prototype.get if):
1228         (BasicBlock):
1229         * JSAir/benchmark.js:
1230         (benchmark):
1231         * JSAir/code.js:
1232         (Code):
1233         (Code.prototype.addBlock):
1234         (Code.prototype.addStackSlot):
1235         (Code.prototype.newTmp):
1236         (Code.prototype.get size):
1237         (Code.prototype.get blocks):
1238         (Code.prototype.get stackSlots):
1239         (Code.prototype.tmps):
1240         (Code.prototype.get callArgAreaSize):
1241         (Code.prototype.toString):
1242         * JSAir/custom.js:
1243         (const.PatchCustom.forEachArg):
1244         * JSAir/inst.js:
1245         (Inst):
1246         * JSAir/reg.js:
1247         (Reg.prototype.toString):
1248         * JSAir/util.js:
1249         (symbolName):
1250         (lowerSymbolName):
1251
1252 2016-06-07  Filip Pizlo  <fpizlo@apple.com>
1253
1254         Add result validation to JSAir
1255         https://bugs.webkit.org/show_bug.cgi?id=158493
1256
1257         Reviewed by Saam Barati.
1258         
1259         This adds the ability to hash a Code in a way that matches the C++ code's hashing of Code.
1260         This allows us to check if the Code that JSAir sees is the code that C++ saw. We use this to
1261         check the Code before and after allocateStack, and compare against hashes we got from C++.
1262         
1263         Doing this uncovered bugs. roundUpToMultipleOf wasn't doing anything. allocateStack was not
1264         allocating things correctly because I was concatting a Set to an Array, which doesn't really
1265         work. Now these bugs are fixed.
1266         
1267         The checking step adds to the running time so I reduced the number of iterations. The
1268         benchmark spends a decent amount of its time computing Code hashes; I think it's around 1/3
1269         total. This is probably OK. It's better to verify the results even if the running time is
1270         not all in the "core" of the algorithm.
1271         
1272         Also add a run-jsc-stress-tests yaml file to allow this to run as a test.
1273
1274         * JSAir/allocate_stack.js:
1275         * JSAir/arg.js:
1276         (Arg.createImm):
1277         (Arg.createBigImm):
1278         (Arg.createBitImm):
1279         (Arg.createBitImm64):
1280         (Arg.createWidth):
1281         (Arg.createSpecial):
1282         (Arg.prototype.get kind):
1283         (Arg.prototype.get isTmp):
1284         (Arg.prototype.get isImm):
1285         (Arg.prototype.get isSomeImm):
1286         (Arg.prototype.get isSomeBigImm):
1287         (Arg.prototype.get isCondition):
1288         (Arg.prototype.get isWidth):
1289         (Arg.prototype.get isSpecial):
1290         (Arg.prototype.get isAlive):
1291         (Arg.prototype.get tmp):
1292         (Arg.prototype.get value):
1293         (Arg.prototype.get lowValue):
1294         (Arg.prototype.get highValue):
1295         (Arg.prototype.get base):
1296         (Arg.prototype.get isGP):
1297         (Arg.prototype.get isFP):
1298         (Arg.prototype.isValidForm):
1299         (Arg.prototype.get isInvertible):
1300         (Arg.kindCode):
1301         (Arg.prototype.hash):
1302         (Arg.prototype.toString):
1303         (Arg):
1304         * JSAir/basic_block.js:
1305         (BasicBlock.get successorBlocks):
1306         * JSAir/benchmark.js:
1307         (benchmark):
1308         * JSAir/code.js:
1309         (Code.prototype.setFrameSize):
1310         (Code.prototype.hash):
1311         (Code.prototype.toString):
1312         (Code):
1313         * JSAir/inst.js:
1314         (Inst.prototype.get hasNonArgEffects):
1315         (Inst.prototype.hash):
1316         (Inst.prototype.toString):
1317         (Inst):
1318         * JSAir/jsair-tests.yaml: Added.
1319         * JSAir/opcode.js:
1320         (Inst_forEachArg):
1321         (Inst_hasNonArgEffects):
1322         (opcodeCode):
1323         * JSAir/payload-gbemu-executeIteration.js:
1324         (createPayloadGbemuExecuteIteration):
1325         * JSAir/payload-imaging-gaussian-blur-gaussianBlur.js:
1326         (createPayloadImagingGaussianBlurGaussianBlur):
1327         * JSAir/payload-jsair-ACLj8C.js:
1328         (createPayloadJSAirACLj8C):
1329         * JSAir/payload-typescript-scanIdentifier.js:
1330         (createPayloadTypescriptScanIdentifier):
1331         * JSAir/reg.js:
1332         (Reg.prototype.get isReg):
1333         (Reg.prototype.hash):
1334         (Reg.prototype.toString):
1335         * JSAir/stack_slot.js:
1336         (StackSlot.prototype.setOffsetFromFP):
1337         (StackSlot.prototype.hash):
1338         (StackSlot.prototype.toString):
1339         * JSAir/symbols.js:
1340         (relCondCode):
1341         (resCondCode):
1342         (doubleCondCode):
1343         * JSAir/test.html:
1344         * JSAir/tmp.js:
1345         (Tmp.prototype.get isReg):
1346         (Tmp.prototype.hash):
1347         (Tmp.prototype.toString):
1348         * JSAir/util.js:
1349         (roundUpToMultipleOf):
1350         (symbolName):
1351
1352 2016-06-07  Filip Pizlo  <fpizlo@apple.com>
1353
1354         Implement Air::allocateStack() in ES6 to see how much of a bad idea that is
1355         https://bugs.webkit.org/show_bug.cgi?id=158318
1356
1357         Reviewed by Saam Barati.
1358         
1359         This adds a new benchmark for us to play with called JSAir. It's a complete ES6
1360         implementation of Air's allocateStack() phase along with all of Air needed to run that
1361         phase. This includes things like stack slots, registers, temporaries, basic blocks,
1362         instructions, and all of the code for iterating over, inspecting, and modifying those
1363         things.
1364         
1365         To make this work, JSC can now dump Air just before allocateStack() in the form of JS code
1366         that creates a Code object that matches exactly what C++ Air saw. This benchmark comprises
1367         four Air IRs:
1368         
1369         - Octane/gbemu's largest function, executeIteration.
1370         - Kraken/imaging-gaussian-blur's largest function in OSR entry mode, gaussuanBlur.
1371         - Octane/typescript's largest function that is actually hot, scanIdentifier.
1372         - JSAir's largest hot function, which is anonymous, so we call it by its hash (ACLj8C).
1373         
1374         This runs in about 2 seconds on my machine in JSC trunk. It includes both a commandline
1375         harness and a web harness. JSAir is almost exactly 100x slower in ES6 in WebKit than the
1376         C++ Air::allocateStack() phase on which it is based.
1377         
1378         JSAir uses the following ES6 features:
1379         
1380         - Symbol.
1381         - for-of.
1382         - arrow functions.
1383         - Map/Set.
1384         - let/const.
1385         - classes.
1386         
1387         All of these things are used in anger and should end up on the hot path. There is also code
1388         that uses Proxies, but it ends up being dead.
1389         
1390         We can improve this even more:
1391         I still need to add result validation: https://bugs.webkit.org/show_bug.cgi?id=158493
1392         I want to make it use more ES6 features: https://bugs.webkit.org/show_bug.cgi?id=158497
1393         
1394         * JSAir: Added.
1395         * JSAir/all.js: Added.
1396         * JSAir/allocate_stack.js: Added.
1397         (allocateStack.attemptAssignment):
1398         (allocateStack.assign):
1399         (allocateStack.interfere):
1400         (allocateStack.):
1401         (allocateStack):
1402         * JSAir/arg.js: Added.
1403         (Arg):
1404         (Arg.isAnyUse):
1405         (Arg.isColdUse):
1406         (Arg.isWarmUse):
1407         (Arg.cooled):
1408         (Arg.isEarlyUse):
1409         (Arg.isLateUse):
1410         (Arg.isAnyDef):
1411         (Arg.isEarlyDef):
1412         (Arg.isLateDef):
1413         (Arg.isZDef):
1414         (Arg.typeForB3Type):
1415         (Arg.widthForB3Type):
1416         (Arg.conservativeWidth):
1417         (Arg.minimumWidth):
1418         (Arg.bytes):
1419         (Arg.widthForBytes):
1420         (Arg.createTmp):
1421         (Arg.fromReg):
1422         (Arg.createImm):
1423         (Arg.createBigImm):
1424         (Arg.createBitImm):
1425         (Arg.createBitImm64):
1426         (Arg.createAddr):
1427         (Arg.createStack):
1428         (Arg.createCallArg):
1429         (Arg.createStackAddr):
1430         (Arg.isValidScale):
1431         (Arg.logScale):
1432         (Arg.createIndex):
1433         (Arg.createRelCond):
1434         (Arg.createResCond):
1435         (Arg.createDoubleCond):
1436         (Arg.createWidth):
1437         (Arg.prototype.get kind):
1438         (Arg.prototype.get isTmp):
1439         (Arg.prototype.get isImm):
1440         (Arg.prototype.get isBigImm):
1441         (Arg.prototype.get isBitImm):
1442         (Arg.prototype.get isBitImm64):
1443         (Arg.prototype.get isSomeImm):
1444         (Arg.prototype.get isAddr):
1445         (Arg.prototype.get isStack):
1446         (Arg.prototype.get isCallArg):
1447         (Arg.prototype.get isIndex):
1448         (Arg.prototype.get isMemory):
1449         (Arg.prototype.get isStackMemory):
1450         (Arg.prototype.get isRelCond):
1451         (Arg.prototype.get isResCond):
1452         (Arg.prototype.get isDoubleCond):
1453         (Arg.prototype.get isCondition):
1454         (Arg.prototype.get isWidth):
1455         (Arg.prototype.get isAlive):
1456         (Arg.prototype.get tmp):
1457         (Arg.prototype.get value):
1458         (Arg.prototype.get base):
1459         (Arg.prototype.get hasOffset):
1460         (Arg.prototype.get offset):
1461         (Arg.prototype.get stackSlot):
1462         (Arg.prototype.get index):
1463         (Arg.prototype.get scale):
1464         (Arg.prototype.get logScale):
1465         (Arg.prototype.get width):
1466         (Arg.prototype.get isGPTmp):
1467         (Arg.prototype.get isFPTmp):
1468         (Arg.prototype.get isGP):
1469         (Arg.prototype.get isFP):
1470         (Arg.prototype.get hasType):
1471         (Arg.prototype.get type):
1472         (Arg.prototype.isType):
1473         (Arg.prototype.isCompatibleType):
1474         (Arg.prototype.get isGPR):
1475         (Arg.prototype.get gpr):
1476         (Arg.prototype.get isFPR):
1477         (Arg.prototype.get fpr):
1478         (Arg.prototype.get isReg):
1479         (Arg.prototype.get reg):
1480         (Arg.isValidImmForm):
1481         (Arg.isValidBitImmForm):
1482         (Arg.isValidBitImm64Form):
1483         (Arg.isValidAddrForm):
1484         (Arg.isValidIndexForm):
1485         (Arg.prototype.isValidForm):
1486         (Arg.prototype.forEachTmpFast):
1487         (Arg.prototype.usesTmp):
1488         (Arg.prototype.forEachTmp):
1489         (Arg.prototype.is):
1490         (Arg.prototype.as):
1491         (Arg.prototype.forEachFast):
1492         (Arg.prototype.forEach):
1493         (Arg.extract):
1494         (Arg.forEachFast):
1495         (Arg.forEach):
1496         (Arg.prototype.get condition):
1497         (Arg.prototype.get isInvertible):
1498         (Arg.prototype.toString):
1499         * JSAir/basic_block.js: Added.
1500         (BasicBlock):
1501         (BasicBlock.prototype.get index):
1502         (BasicBlock.prototype.get size):
1503         (BasicBlock.prototype.Symbol.iterator):
1504         (BasicBlock.prototype.at):
1505         (BasicBlock.get last):
1506         (BasicBlock.get insts):
1507         (BasicBlock.get numSuccessors):
1508         (BasicBlock.get successors):
1509         (BasicBlock.get successorBlocks.):
1510         (BasicBlock.get successorBlocks):
1511         (BasicBlock.set get numPredecessors):
1512         (BasicBlock.get predecessors):
1513         (BasicBlock.get frequency):
1514         (BasicBlock.get headerString):
1515         (BasicBlock.get footerString):
1516         * JSAir/benchmark.js: Added.
1517         (benchmark):
1518         * JSAir/code.js: Added.
1519         (Code):
1520         (Code.prototype.addBlock):
1521         (Code.prototype.addStackSlot):
1522         (Code.prototype.newTmp):
1523         (Code.prototype.get size):
1524         (Code.prototype.at):
1525         (Code.prototype.Symbol.iterator):
1526         (Code.prototype.get blocks):
1527         (Code.prototype.get stackSlots):
1528         (Code.prototype.tmps):
1529         (Code.prototype.get callArgAreaSize):
1530         (Code.prototype.requestCallArgAreaSize):
1531         (Code.prototype.get frameSize):
1532         (Code.prototype.setFrameSize):
1533         (Code.prototype.toString):
1534         * JSAir/custom.js: Added.
1535         (const.ShuffleCustom.forEachArg):
1536         (const.ShuffleCustom.hasNonArgNonControlEffects):
1537         (const.PatchCustom.forEachArg):
1538         (const.PatchCustom.hasNonArgNonControlEffects):
1539         (const.CCallCustom.forEachArg):
1540         (const.CCallCustom.hasNonArgNonControlEffects):
1541         (const.ColdCCallCustom.forEachArg):
1542         (const.ColdCCallCustom.hasNonArgNonControlEffects):
1543         * JSAir/frequented_block.js: Added.
1544         (FrequentedBlock):
1545         (FrequentedBlock.prototype.toString):
1546         * JSAir/insertion_set.js: Added.
1547         (Insertion):
1548         (Insertion.prototype.get index):
1549         (Insertion.prototype.get element):
1550         (Insertion.prototype.lessThan):
1551         (InsertionSet):
1552         (InsertionSet.prototype.appendInsertion):
1553         (InsertionSet.prototype.append):
1554         (InsertionSet.prototype.execute):
1555         * JSAir/inst.js: Added.
1556         (Inst):
1557         (Inst.prototype.append):
1558         (Inst.prototype.clear):
1559         (Inst.prototype.get opcode):
1560         (Inst.prototype.get args):
1561         (Inst.prototype.visitArg):
1562         (Inst.prototype.forEachTmpFast):
1563         (Inst.prototype.forEachArg):
1564         (Inst.prototype.forEachTmp):
1565         (Inst.prototype.forEach):
1566         (Inst.forEachDef):
1567         (Inst.forEachDefWithExtraClobberedRegs):
1568         (Inst.prototype.get hasNonArgEffects):
1569         (Inst.prototype.toString):
1570         * JSAir/liveness.js: Added.
1571         (Liveness):
1572         (Liveness.prototype.get thing):
1573         (Liveness.prototype.get code):
1574         (Liveness.prototype.get liveAtHead):
1575         (Liveness.prototype.get liveAtTail):
1576         (Liveness.prototype.localCalc.LocalCalc):
1577         (Liveness.prototype.localCalc.LocalCalc.prototype.get liveSet):
1578         (Liveness.prototype.localCalc.LocalCalc.prototype.execute):
1579         (Liveness.prototype.localCalc):
1580         * JSAir/opcode.js: Added.
1581         (Inst_forEachArg):
1582         (Inst_hasNonArgEffects):
1583         * JSAir/payload-gbemu-executeIteration.js: Added.
1584         (createPayloadGbemuExecuteIteration):
1585         * JSAir/payload-imaging-gaussian-blur-gaussianBlur.js: Added.
1586         (createPayloadImagingGaussianBlurGaussianBlur):
1587         * JSAir/payload-jsair-ACLj8C.js: Added.
1588         (createPayloadJSAirACLj8C):
1589         * JSAir/payload-typescript-scanIdentifier.js: Added.
1590         (createPayloadTypescriptScanIdentifier):
1591         * JSAir/reg.js: Added.
1592         (Reg):
1593         (Reg.fromReg):
1594         (Reg.prototype.get index):
1595         (Reg.prototype.get type):
1596         (Reg.prototype.get name):
1597         (Reg.prototype.get isCalleeSave):
1598         (Reg.prototype.get isReg):
1599         (Reg.prototype.toString):
1600         (Reg.extract):
1601         (Reg.forEachFast):
1602         (Reg.forEach):
1603         (newGPR):
1604         (Reg.gprs.Reg.fprs.Reg.calleeSaveGPRs.Reg.calleeSaveFPRs.Reg.calleeSaves):
1605         * JSAir/stack_slot.js: Added.
1606         (StackSlot):
1607         (StackSlot.prototype.get byteSize):
1608         (StackSlot.prototype.get kind):
1609         (StackSlot.prototype.get isLocked):
1610         (StackSlot.prototype.get isSpill):
1611         (StackSlot.prototype.get index):
1612         (StackSlot.prototype.ensureSize):
1613         (StackSlot.prototype.get alignment):
1614         (StackSlot.prototype.get offsetFromFP):
1615         (StackSlot.prototype.setOffsetFromFP):
1616         (StackSlot.prototype.toString):
1617         (StackSlot.extract):
1618         (StackSlot.forEachFast):
1619         (StackSlot.forEach):
1620         * JSAir/symbols.js: Added.
1621         * JSAir/test.html: Added.
1622         * JSAir/test.js: Added.
1623         * JSAir/tmp.js: Added.
1624         (Tmp):
1625         (Tmp.fromReg):
1626         (Tmp.prototype.get index):
1627         (Tmp.prototype.get type):
1628         (Tmp.prototype.get isReg):
1629         (Tmp.prototype.toString):
1630         (Tmp.extract):
1631         (Tmp.forEachFast):
1632         (Tmp.forEach):
1633         * JSAir/tmp_base.js: Added.
1634         (TmpBase.prototype.get isGP):
1635         (TmpBase.prototype.get isFP):
1636         (TmpBase.prototype.get isGPR):
1637         (TmpBase.prototype.get isFPR):
1638         (TmpBase.prototype.get reg):
1639         (TmpBase.prototype.get gpr):
1640         (TmpBase.prototype.get fpr):
1641         (TmpBase):
1642         * JSAir/util.js: Added.
1643         (isRepresentableAsInt32):
1644         (addIndexed):
1645         (roundUpToMultipleOf):
1646         (symbolName):
1647         (mergeIntoSet):
1648         (nonEmptyRangesOverlap):
1649         (rangesOverlap):
1650         (removeAllMatching):
1651         (swap):
1652         (bubble):
1653         (bubbleSort):
1654         (currentTime):
1655         (else.currentTime):
1656
1657 2016-05-26  Yusuke Suzuki  <utatane.tea@gmail.com>
1658
1659         [JSC] Allow JSBench to use precise time
1660         https://bugs.webkit.org/show_bug.cgi?id=158050
1661
1662         Reviewed by Geoffrey Garen.
1663
1664         * JSBench/amazon-chrome-win/urem.js:
1665         (else.window.performance.window.performance.now.currentTimeInMS):
1666         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1667         (else.else.currentTimeInMS):
1668         (onload.cb):
1669         (onload):
1670         * JSBench/amazon-chrome/urem.js:
1671         (else.window.performance.window.performance.now.currentTimeInMS):
1672         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1673         (else.else.currentTimeInMS):
1674         (onload.cb):
1675         (onload):
1676         * JSBench/amazon-firefox-win/urm.js:
1677         (else.window.performance.window.performance.now.currentTimeInMS):
1678         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1679         (else.else.currentTimeInMS):
1680         (onload.cb):
1681         (onload):
1682         * JSBench/amazon-firefox/urm.js:
1683         (else.window.performance.window.performance.now.currentTimeInMS):
1684         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1685         (else.else.currentTimeInMS):
1686         (onload.cb):
1687         (onload):
1688         * JSBench/amazon-safari/urem.js:
1689         (else.window.performance.window.performance.now.currentTimeInMS):
1690         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1691         (else.else.currentTimeInMS):
1692         (onload.cb):
1693         (onload):
1694         * JSBench/facebook-chrome-win/urem.js:
1695         (else.window.performance.window.performance.now.currentTimeInMS):
1696         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1697         (else.else.currentTimeInMS):
1698         (onload.cb):
1699         (onload):
1700         * JSBench/facebook-chrome/urem.js:
1701         (else.window.performance.window.performance.now.currentTimeInMS):
1702         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1703         (else.else.currentTimeInMS):
1704         (onload.cb):
1705         (onload):
1706         * JSBench/facebook-firefox-win/urem.js:
1707         (else.window.performance.window.performance.now.currentTimeInMS):
1708         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1709         (else.else.currentTimeInMS):
1710         (onload.cb):
1711         (onload):
1712         * JSBench/facebook-firefox/urem.js:
1713         (else.window.performance.window.performance.now.currentTimeInMS):
1714         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1715         (else.else.currentTimeInMS):
1716         (onload.cb):
1717         (onload):
1718         * JSBench/facebook-safari/urem.js:
1719         (else.window.performance.window.performance.now.currentTimeInMS):
1720         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1721         (else.else.currentTimeInMS):
1722         (onload.cb):
1723         (onload):
1724         * JSBench/google-chrome-win/urem.js:
1725         (else.window.performance.window.performance.now.currentTimeInMS):
1726         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1727         (else.else.currentTimeInMS):
1728         (onload.cb):
1729         (onload):
1730         * JSBench/google-chrome/urem.js:
1731         (else.window.performance.window.performance.now.currentTimeInMS):
1732         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1733         (else.else.currentTimeInMS):
1734         (onload.cb):
1735         (onload):
1736         * JSBench/google-firefox-win/urem.js:
1737         (else.window.performance.window.performance.now.currentTimeInMS):
1738         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1739         (else.else.currentTimeInMS):
1740         (onload.cb):
1741         (onload):
1742         * JSBench/google-firefox/uem.js:
1743         (else.window.performance.window.performance.now.currentTimeInMS):
1744         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1745         (else.else.currentTimeInMS):
1746         (onload.cb):
1747         (onload):
1748         * JSBench/google-safari/urem.js:
1749         (else.window.performance.window.performance.now.currentTimeInMS):
1750         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1751         (else.else.currentTimeInMS):
1752         (onload.cb):
1753         (onload):
1754         * JSBench/harness.js:
1755         (runBenchmark.window.currentTimeInMS):
1756         (runBenchmark.else.window.currentTimeInMS):
1757         * JSBench/twitter-chrome-win/rem.js:
1758         (else.window.performance.window.performance.now.currentTimeInMS):
1759         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1760         (else.else.currentTimeInMS):
1761         (onload.cb):
1762         (onload):
1763         * JSBench/twitter-chrome/urem.js:
1764         (else.window.performance.window.performance.now.currentTimeInMS):
1765         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1766         (else.else.currentTimeInMS):
1767         (onload.cb):
1768         (onload):
1769         * JSBench/twitter-firefox-win/urem.js:
1770         (else.window.performance.window.performance.now.currentTimeInMS):
1771         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1772         (else.else.currentTimeInMS):
1773         (onload.cb):
1774         (onload):
1775         * JSBench/twitter-firefox/urem.js:
1776         (else.window.performance.window.performance.now.currentTimeInMS):
1777         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1778         (else.else.currentTimeInMS):
1779         (onload.cb):
1780         (onload):
1781         * JSBench/twitter-safari/urem.js:
1782         (else.window.performance.window.performance.now.currentTimeInMS):
1783         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1784         (else.else.currentTimeInMS):
1785         (onload.cb):
1786         (onload):
1787         * JSBench/yahoo-chrome-win/urem.js:
1788         (else.window.performance.window.performance.now.currentTimeInMS):
1789         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1790         (else.else.currentTimeInMS):
1791         (onload.cb):
1792         (onload):
1793         * JSBench/yahoo-chrome/urem.js:
1794         (else.window.performance.window.performance.now.currentTimeInMS):
1795         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1796         (else.else.currentTimeInMS):
1797         (onload.cb):
1798         (onload):
1799         * JSBench/yahoo-firefox-win/urem.js:
1800         (else.window.performance.window.performance.now.currentTimeInMS):
1801         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1802         (else.else.currentTimeInMS):
1803         (onload.cb):
1804         (onload):
1805         * JSBench/yahoo-firefox/urem.js:
1806         (else.window.performance.window.performance.now.currentTimeInMS):
1807         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1808         (else.else.currentTimeInMS):
1809         (onload.cb):
1810         (onload):
1811         * JSBench/yahoo-safari/urem.js:
1812         (else.window.performance.window.performance.now.currentTimeInMS):
1813         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1814         (else.else.currentTimeInMS):
1815         (onload.cb):
1816         (onload):
1817
1818 2016-05-25  Geoffrey Garen  <ggaren@apple.com>
1819
1820         replaceable own properties seem to ignore replacement after property caching
1821         https://bugs.webkit.org/show_bug.cgi?id=158091
1822
1823         Reviewed by Darin Adler.
1824
1825         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
1826         * MallocBench/MallocBench/Benchmark.cpp:
1827         * MallocBench/MallocBench/Interpreter.cpp:
1828         (Interpreter::doMallocOp):
1829         * MallocBench/MallocBench/Interpreter.h:
1830         * MallocBench/MallocBench/fastMallocLog.63316.ops: Added.
1831         * MallocBench/MallocBench/jetstream.cpp: Added.
1832         (benchmark_jetstream):
1833         * MallocBench/MallocBench/jetstream.h: Added.
1834
1835 2016-05-25  Keith Miller  <keith_miller@apple.com>
1836
1837         Unreviewed, add JSBench to the skipped list for now since it doesn't
1838         work currently.
1839
1840         * Skipped:
1841
1842 2016-05-17  Keith Miller  <keith_miller@apple.com>
1843         We should have JSBench in PerformanceTests
1844         https://bugs.webkit.org/show_bug.cgi?id=157952
1845
1846         Rubber-stamped by Saam Barati.
1847
1848         There are some slight changes to the layout of the test directory
1849         to make it work nicely with run-jsc-benchmarks. Before JSBench
1850         had each of the browser specific sub-tests in a sub-directory.
1851         These have been flattened e.g. amazon/safari/ has become
1852         amazon-safari/.
1853
1854         * JSBench/amazon-chrome-win/urem.html: Added.
1855         * JSBench/amazon-chrome-win/urem.js: Added.
1856         * JSBench/amazon-chrome/urem.html: Added.
1857         * JSBench/amazon-chrome/urem.js: Added.
1858         * JSBench/amazon-firefox-win/urm.html: Added.
1859         * JSBench/amazon-firefox-win/urm.js: Added.
1860         * JSBench/amazon-firefox/urm.html: Added.
1861         * JSBench/amazon-firefox/urm.js: Added.
1862         * JSBench/amazon-safari/urem.html: Added.
1863         * JSBench/amazon-safari/urem.js: Added.
1864         * JSBench/browsercheck.js: Added.
1865         * JSBench/facebook-chrome-win/urem.html: Added.
1866         * JSBench/facebook-chrome-win/urem.js: Added.
1867         * JSBench/facebook-chrome/urem.html: Added.
1868         * JSBench/facebook-chrome/urem.js: Added.
1869         * JSBench/facebook-firefox-win/urem.html: Added.
1870         * JSBench/facebook-firefox-win/urem.js: Added.
1871         * JSBench/facebook-firefox/urem.html: Added.
1872         * JSBench/facebook-firefox/urem.js: Added.
1873         * JSBench/facebook-safari/urem.html: Added.
1874         * JSBench/facebook-safari/urem.js: Added.
1875         * JSBench/google-chrome-win/urem.html: Added.
1876         * JSBench/google-chrome-win/urem.js: Added.
1877         * JSBench/google-chrome/urem.html: Added.
1878         * JSBench/google-chrome/urem.js: Added.
1879         * JSBench/google-firefox-win/urem.html: Added.
1880         * JSBench/google-firefox-win/urem.js: Added.
1881         * JSBench/google-firefox/uem.html: Added.
1882         * JSBench/google-firefox/uem.js: Added.
1883         * JSBench/google-safari/urem.html: Added.
1884         * JSBench/google-safari/urem.js: Added.
1885         * JSBench/harness.html: Added.
1886         * JSBench/harness.js: Added.
1887         * JSBench/harness.py: Added.
1888         * JSBench/index.html: Added.
1889         * JSBench/reload.html: Added.
1890         * JSBench/twitter-chrome-win/rem.html: Added.
1891         * JSBench/twitter-chrome-win/rem.js: Added.
1892         * JSBench/twitter-chrome/urem.html: Added.
1893         * JSBench/twitter-chrome/urem.js: Added.
1894         * JSBench/twitter-firefox-win/urem.html: Added.
1895         * JSBench/twitter-firefox-win/urem.js: Added.
1896         * JSBench/twitter-firefox/urem.html: Added.
1897         * JSBench/twitter-firefox/urem.js: Added.
1898         * JSBench/twitter-safari/urem.html: Added.
1899         * JSBench/twitter-safari/urem.js: Added.
1900         * JSBench/yahoo-chrome-win/urem.html: Added.
1901         * JSBench/yahoo-chrome-win/urem.js: Added.
1902         * JSBench/yahoo-chrome/urem.html: Added.
1903         * JSBench/yahoo-chrome/urem.js: Added.
1904         * JSBench/yahoo-firefox-win/urem.html: Added.
1905         * JSBench/yahoo-firefox-win/urem.js: Added.
1906         * JSBench/yahoo-firefox/urem.html: Added.
1907         * JSBench/yahoo-firefox/urem.js: Added.
1908         * JSBench/yahoo-safari/urem.html: Added.
1909         * JSBench/yahoo-safari/urem.js: Added.
1910
1911 2016-05-20  Ryosuke Niwa  <rniwa@webkit.org>
1912
1913         run-benchmark's results should contain Animometer's debug output
1914         https://bugs.webkit.org/show_bug.cgi?id=157941
1915
1916         Reviewed by Stephanie Lewis.
1917
1918         Made developer.html support the JSON generated by run-benchmark which stores everything under debugOutput.
1919
1920         * Animometer/resources/debug-runner/animometer.js:
1921         (Utilities.initialize): Unwrap debugOutput in the case run-benchmark's result JSON is used.
1922
1923 2016-05-18  Timothy Hatcher  <timothy@apple.com>
1924
1925         Make Animometer work in all browsers
1926         https://bugs.webkit.org/show_bug.cgi?id=157855
1927         rdar://problem/26338521
1928
1929         Reviewed by Jon Lee.
1930
1931         * Animometer/resources/runner/logo.svg: Use a symbol instead
1932         of putting the id on the root svg. This works in all browsers.
1933
1934 2016-05-07  Timothy Hatcher  <timothy@apple.com>
1935
1936         Update Animometer design
1937         https://bugs.webkit.org/show_bug.cgi?id=157449
1938
1939         Reviewed by Darin Adler.
1940
1941         * Animometer/developer.html: Stop including runner/animometer.css.
1942         * Animometer/index.html: Udpated elements to fit new styles.
1943
1944         * Animometer/resources/debug-runner/animometer.css:
1945         Copy the old runner/animometer.css to here so the debug runner is unaffected.
1946
1947         * Animometer/resources/runner/animometer.css:
1948         New styles for the design. Drop flex box for broad browser support.
1949
1950         * Animometer/resources/runner/animometer.js:
1951         (ResultsTable.clear): Use textContent.
1952         (window.sectionsManager.showSection): Add class to the body.
1953         (window.sectionsManager.setSectionScore): Use textContent, not innerHTML.
1954
1955         * Animometer/resources/runner/crystal.svg: Added.
1956         * Animometer/resources/runner/lines.svg: Added.
1957         * Animometer/resources/runner/logo.svg: Added.
1958
1959         * Animometer/resources/runner/tests.js: Shorten test name and title case them.
1960
1961 2016-05-06  Manuel Rego Casasnovas  <rego@igalia.com>
1962
1963         [css-grid] Unprefix CSS Grid Layout properties
1964         https://bugs.webkit.org/show_bug.cgi?id=157137
1965
1966         Reviewed by Simon Fraser.
1967
1968         Remove "-webkit" prefix from all the grid layout properties,
1969         including the display value.
1970         Update the source code to remove the prefix where it was used too.
1971
1972         * Layout/auto-grid-lots-of-data.html:
1973         * Layout/fixed-grid-lots-of-data.html:
1974         * Layout/fixed-grid-lots-of-stretched-data.html:
1975
1976 2016-04-25  Geoffrey Garen  <ggaren@apple.com>
1977
1978         bmalloc: Misc improvements to MallocBench
1979         https://bugs.webkit.org/show_bug.cgi?id=157004
1980
1981         Reviewed by Darin Adler.
1982
1983         * MallocBench/run-malloc-benchmarks: Added --memory and --memory_warning
1984         modes for focused memory testing.
1985
1986         * MallocBench/MallocBench/Benchmark.cpp:
1987         (Benchmark::printReport): Clarified output.
1988
1989         (Benchmark::currentMemoryBytes): Added compressed memory because top
1990         does the same. (It always happens to zero in the benchmarks we run. But
1991         this is good for sanity.)
1992
1993         * MallocBench/MallocBench/CommandLine.cpp: Moved up to 8 runs to reduce
1994         variance.
1995
1996         * MallocBench/MallocBench/alloc_free.cpp:
1997         (benchmark_alloc_free): Cycle a single allocation in order to stress
1998         the effect of merging on calls to madvise.
1999
2000         * MallocBench/MallocBench/big.cpp:
2001         (benchmark_big): Graduated to 8kB-128kB because medium tests up to 8 and
2002         our large allocator doesn't kick in until 64kB.
2003
2004         * MallocBench/MallocBench/medium.cpp:
2005         (benchmark_medium): Test all the way down to 1kB because our large
2006         allocator used to service 1kB allocations and 1kB is an interesting
2007         middle size where memory is unusually large but allocation throughput
2008         still matters.
2009
2010         * MallocBench/MallocBench/stress.cpp:
2011         (benchmark_stress): Reduced the churn count to match stress_aligned
2012         because this test was taking too long to complete.
2013
2014         * MallocBench/MallocBench/stress_aligned.cpp:
2015         (benchmark_stress_aligned): Our new large allocator can handle even
2016         more absurdly large values.
2017
2018 2016-04-25  Simon Fraser  <simon.fraser@apple.com>
2019
2020         Add a content animation test that uses SVG animation.
2021         https://bugs.webkit.org/show_bug.cgi?id=156827
2022
2023         * Animation/svg-animation.html: Added.
2024         * Skipped:
2025
2026 2016-04-22  Simon Fraser  <simon.fraser@apple.com>
2027
2028         Skip two content animation tests which are only meant for iOS testing.
2029
2030         * Animation/css-animation.html: Added.
2031         * Animation/raf-animation.html: Added.
2032
2033         * Skipped:
2034
2035 2016-04-20  Simon Fraser  <simon.fraser@apple.com>
2036
2037         Add content animation tests to benchmark_runner, and allow the runner to collect device data as part of the results
2038         https://bugs.webkit.org/show_bug.cgi?id=156827
2039
2040         Add two files missing from the previous commit.
2041
2042         * Animation/css-animation.html: Added.
2043         * Animation/raf-animation.html: Added.
2044
2045 2016-04-20  Simon Fraser  <simon.fraser@apple.com>
2046
2047         Add content animation tests to benchmark_runner, and allow the runner to collect device data as part of the results
2048         https://bugs.webkit.org/show_bug.cgi?id=156827
2049
2050         Reviewed by Stephanie Lewis.
2051
2052         Two new files for the content-animation suite.
2053
2054         After patching for use in the benchmark, these tests are a little different to other benchmarks.
2055         The test content itself does not collect performance data; the tests spit out a boilerplate JSON
2056         result with a placeholder for framerate, which is collected in native code.
2057
2058         When run as a benchmark, test completion involves a couple of bounces between the test and MobileSafari.
2059         Test completion is initiated by a setTimeout() in the test, at which point location.hash is set to "#done".
2060         MobileSafari detects that, and asynchronously dumps performance data. Once this is complete, MobileSafari
2061         changes the URL has to "#submit", which triggers the XHR to the test relay that indicates the test is
2062         complete.
2063
2064         * Animation/css-animation.html: Added.
2065         * Animation/raf-animation.html: Added.
2066
2067 2016-04-15  Jon Lee  <jonlee@apple.com>
2068
2069         Animometer test could report a NaN
2070         https://bugs.webkit.org/show_bug.cgi?id=156646
2071
2072         Reviewed by Darin Adler.
2073         Provisionally reviewed by Said Abou-Hallawa.
2074
2075         * Animometer/tests/resources/main.js:
2076         (didFinishInterval): The ramp controller has a first phase where it ramps up the complexity and
2077         reacts based on how well the system handles the load. The assumption was that it would handle at least
2078         1 particle easily. That is not always the case. As a result, an interpolation calculation could
2079         end up setting an upper bound of NaN.
2080
2081         This occurs because we never get out of the first tier, so this._lastTierComplexity is undefined.
2082         Now that we guarantee a minimum complexity of 1, modify the conditional to check for this._lastTierComplexity
2083         before interpolating the upper bound of the first ramp. In the case where the system struggles with
2084         1 particle, set it to currentComplexity.
2085
2086 2016-04-09  Jon Lee  <jonlee@apple.com>
2087
2088         Update Animometer to accommodate different screens
2089         https://bugs.webkit.org/show_bug.cgi?id=156449
2090
2091         Reviewed by Darin Adler.
2092         Provisionally reviewed by Said Abou-Hallawa.
2093
2094         * Animometer/index.html: Wrap button in a container to add padding at the bottom.
2095         * Animometer/resources/debug-runner/animometer.css:
2096         (@media screen and (min-device-width: 1800px)): Deleted.
2097         (@media screen and (min-width: 1800px)): Cannot use min-device-width since it may match incorrectly.
2098         (screen and (max-device-height: 414px) and (orientation: landscape)): Some devices swap device width
2099         and height with orientation change.
2100         * Animometer/resources/runner/animometer.css: Similar.
2101         (screen and (min-device-width: 1024px) and (orientation: landscape)):
2102         (screen and (max-device-height: 414px) and (orientation: landscape)):
2103         (.frame-container): On smaller iPhones, adding 1px prevents the navigation bars from appearing.
2104         (@media screen and (min-device-width: 768px) and (max-device-width: 1024px)): Deleted.
2105         (@media (min-device-height: 768px) and (max-device-height: 1024px)): Target iPad Airs and similar.
2106         (@media screen and (min-device-width: 1024px) and (max-device-width: 1366px)): Deleted.
2107         (@media screen and (max-device-width: 1024px) and (min-device-height: 1366px)): Target iPad Pro.
2108         (#results footer): Add padding below the button for testing again.
2109         * Animometer/tests/master/multiply.html: Remove the center text.
2110         * Animometer/tests/master/resources/text.js: Update the test so that in every frame the text moves.
2111         * Animometer/tests/master/text.html: Update the text sizing depending on the size of the device.
2112
2113 2016-04-08  Jon Lee  <jonlee@apple.com>
2114
2115         Have Animometer benchmark always start with complexity of 1
2116         https://bugs.webkit.org/show_bug.cgi?id=156432
2117
2118         Reviewed by Ryosuke Niwa.
2119
2120         * Animometer/tests/resources/main.js: Update the default Controller and RampController to
2121         set its minimum complexities to 1 instead of 0.
2122
2123 2016-04-08  Jon Lee  <jonlee@apple.com>
2124
2125         Fix SVG benchmark test
2126         https://bugs.webkit.org/show_bug.cgi?id=156410
2127
2128         Reviewed by Dean Jackson.
2129
2130         * Animometer/resources/extensions.js: Update Point.zero to be a static Point.
2131         * Animometer/tests/simple/resources/tiled-canvas-image.js:
2132         (Stage.call._setupTiles): Refactor.
2133         * Animometer/tests/master/resources/particles.js:
2134         (Particle.prototype.reset): Use Point.center.
2135         (complexity): We are not using a gradient background anymore, so remove the +1.
2136         * Animometer/tests/master/resources/svg-particles.js: Update to use SVG transform
2137         instead of CSS transform.
2138
2139 2016-04-06  Jon Lee  <jonlee@apple.com>
2140
2141         Update master benchmark with SVG test
2142         https://bugs.webkit.org/show_bug.cgi?id=156273
2143
2144         Reviewed by Dean Jackson.
2145         Provisionally reviewed by Said Abou-Hallawa.
2146
2147         Switch masks tests for SVG path test.
2148
2149         * Animometer/resources/debug-runner/tests.js: Move mask test here.
2150         * Animometer/resources/runner/tests.js: Add SVG test here.
2151         * Animometer/tests/dom/particles.html: Renamed from PerformanceTests/Animometer/tests/master/particles.html.
2152         * Animometer/tests/dom/resources/dom-particles.js: Renamed from PerformanceTests/Animometer/tests/master/resources/dom-particles.js.
2153
2154         * Animometer/tests/master/resources/particles.js: Add minPosition for bounds checking. Prevents particle from being
2155         partially obscured.
2156         * Animometer/tests/master/resources/svg-particles.js: Added.
2157         (Particle): The particle is either a path object or a rect using a path as a clip. The
2158         same path is used either way. For each particle create a linear gradient with a random
2159         rotation.
2160         (SVGParticlesStage): Look in #shapes to see how many different kinds of paths are available.
2161         This makes the test more generic in case other shapes need to be tested.
2162         * Animometer/tests/master/svg-particles.html: Added. Have two defs, one that houses each
2163         particle's gradient, and one that holds the shape templates.
2164
2165 2016-03-27  Hunseop Jeong  <hs85.jeong@samsung.com>
2166
2167         [EFL] REGRESSION(r188793): It made 200 layout tests and Bindings/event-target-wrapper.html performance test fail
2168         https://bugs.webkit.org/show_bug.cgi?id=148470 
2169
2170         Reviewed by Darin Adler.
2171
2172         * Skipped: Unskip the Bindings/event-target-wrapper test.
2173
2174 2016-03-23  Jon Lee  <jonlee@apple.com>
2175
2176         Update focus and leaves tests
2177         https://bugs.webkit.org/show_bug.cgi?id=155825
2178
2179         Reviewed by Simon Fraser.
2180
2181         Particles would improperly disappear, before they were fully off stage. Update the calculation.
2182         * Animometer/tests/dom/resources/leaves.js:
2183         * Animometer/tests/master/resources/leaves.js:
2184
2185         * Animometer/tests/master/focus.html: Update the center element styling.
2186         * Animometer/tests/master/resources/focus.js:
2187         (FocusElement.Utilities.createClass): Give enough space in the container for each particle for the blur
2188         to spread. It should be about 3x more space than the blur radius on each side.
2189         (Stage.call.initialize): Ditto for the center element.
2190
2191 2016-03-23  Jon Lee  <jonlee@apple.com>
2192
2193         Update support for other platforms
2194         https://bugs.webkit.org/show_bug.cgi?id=155824
2195
2196         Reviewed by Simon Fraser.
2197
2198         * Animometer/resources/debug-runner/animometer.css: Remove extraneous media query. All of it should
2199         be handled in the release stylesheet.
2200         * Animometer/resources/debug-runner/animometer.js: Add support for browsers that return an HTMLCollection
2201         when retrieving radio elements with a name, instead of the selected value.
2202         * Animometer/resources/runner/animometer.css: Adjust the queries for devices of varying widths.
2203         (.frame-container): Move overflow:hidden to #stage.
2204         * Animometer/tests/master/resources/stage.css: Consolidate into tests/resources/stage.css.
2205         * Animometer/tests/resources/stage.css:
2206         (body): Remove overflow:hidden.
2207         (#stage): Set overflow:hidden here.
2208         (#center-text): Moved from master/resources/stage.css.
2209
2210         Update stylesheet hrefs.
2211         * Animometer/tests/master/canvas-stage.html:
2212         * Animometer/tests/master/focus.html:
2213         * Animometer/tests/master/leaves.html:
2214         * Animometer/tests/master/multiply.html:
2215         * Animometer/tests/master/particles.html:
2216         * Animometer/tests/master/text.html: Adjust text size to fit in the device screen.
2217
2218 2016-03-23  Jon Lee  <jonlee@apple.com>
2219
2220         Make the benchmark require device to be in landscape orientation
2221         https://bugs.webkit.org/show_bug.cgi?id=155822
2222         rdar://problem/25258650
2223
2224         Reviewed by Ryosuke Niwa.
2225
2226         Check that the device is in landscape orientation. If not, disable the button to start it,
2227         and show a warning message.
2228
2229         * Animometer/developer.html: Add a message about how the browser should be set up. Include a
2230         warning message that appears if the orientation is incorrect on mobile devices.
2231         * Animometer/index.html: Ditto.
2232
2233         * Animometer/resources/debug-runner/animometer.css: Migrate the .hidden rule into the release
2234         stylesheet. Update the style to accommodate the new UI.
2235         * Animometer/resources/runner/animometer.css:
2236
2237         * Animometer/resources/runner/animometer.js:
2238         (window.benchmarkController.initialize): Add the orientation listener is needed.
2239         (window.benchmarkController.addOrientationListenerIfNecessary): Only mobile devices need this. Check
2240         to see for support of window.orientation.
2241         (window.benchmarkController._orientationChanged): Toggle the warning based on whether we match the
2242         landscape query. Set a state variable, which is needed for the debug harness. Call updateStartButtonState.
2243         (window.benchmarkController.updateStartButtonState):
2244         * Animometer/resources/debug-runner/animometer.js: Change _updateStartButtonState to return
2245         a boolean about whether at least one test is selected. That will be used in
2246         benchmarkController.updateStartButtonState(). Move the load event listener to the release version.
2247         (window.benchmarkController.updateStartButtonState): Override the release version, and also check
2248         that at least one test is selected.
2249
2250 2016-03-22  Geoffrey Garen  <ggaren@apple.com>
2251
2252         MallocBench: consolidate regression testing for aligned allocation
2253         https://bugs.webkit.org/show_bug.cgi?id=155762
2254
2255         Reviewed by Michael Saboff.
2256
2257         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
2258         * MallocBench/MallocBench/Benchmark.cpp:
2259         * MallocBench/MallocBench/memalign.cpp: Removed.
2260         * MallocBench/MallocBench/memalign.h: Removed. The stress_aligned test
2261         covers this and much more.
2262
2263         * MallocBench/MallocBench/stress_aligned.cpp:
2264         (benchmark_stress_aligned): Include specific tests for extreme sizes
2265         and alignments.
2266
2267 2016-03-21  Jon Lee  <jonlee@apple.com>
2268
2269         Update benchmark tests
2270         https://bugs.webkit.org/show_bug.cgi?id=155723
2271
2272         Reviewed by Darin Adler.
2273         Provisionally reviewed by Said Abou-Hallawa.
2274
2275         Add quadratic and bezier segments to the canvas path test.
2276
2277         * Animometer/resources/runner/tests.js: Some of the query strings are unnecessary.
2278         Rename the test.
2279         * Animometer/tests/master/resources/canvas-stage.js:
2280         (tune): Update to be able to take an array of possible constructors. Choose one randomly.
2281         * Animometer/tests/master/resources/canvas-tests.js:
2282         (CanvasLinePoint.Utilities.createClass): Move the point selection out to a separate
2283         function called randomPoint() for reuse. Move X_LOOPS, Y_LOOPS, and offsets into the
2284         class definition.
2285         (randomPoint): Scale the grid down a little bit so that the lines along the edge of the
2286         canvas are not cut off when the stroke size is thick.
2287         (CanvasQuadraticSegment): Added.
2288         (CanvasBezierSegment): Added.
2289         (SimpleCanvasStage): Pass in an array of the different segment types. Since line segments
2290         are short compared to the curved ones, make it twice as likely to render a line segment.
2291         (SimpleCanvasStage.animate): Update the drawing code so that we render all line segments.
2292
2293         Add a helper method that selects a random element from an array.
2294
2295         * Animometer/tests/resources/main.js:
2296         (Stage.randomElementInArray): Select a random element from the provided array.
2297
2298         * Animometer/tests/bouncing-particles/resources/bouncing-tagged-images.js: Refactor.
2299         * Animometer/tests/dom/resources/leaves.js: Ditto.
2300         * Animometer/tests/master/resources/dom-particles.js: Ditto.
2301         * Animometer/tests/master/resources/image-data.js: Ditto.
2302         * Animometer/tests/master/resources/leaves.js: Ditto.
2303         * Animometer/tests/simple/resources/simple-canvas-paths.js: Ditto.
2304
2305         Add canvas tests that includes all stroke and fill paths. This makes it possible to avoid having to
2306         include the full simple canvas suite for perf testing.
2307
2308         * Animometer/resources/debug-runner/tests.js: Add new tests. Move the canvas test into the Canvas
2309         suite.
2310         * Animometer/tests/simple/resources/simple-canvas-paths.js: Add a CanvasStroke and CanvasFill particle
2311         that random selects an object to render.
2312
2313         * Animometer/resources/debug-runner/tests.js: Move 3D suite before basic canvas suite.
2314         When updating the perf bot script, we will include the suites up to this one, but exclude
2315         the basic canvas suite.
2316
2317         Move compositing transforms test to HTML suite and remove the empty Miscellaneous suite.
2318
2319         * Animometer/resources/debug-runner/tests.js:
2320         * Animometer/tests/dom/compositing-transforms.html: Renamed from PerformanceTests/Animometer/tests/misc/compositing-transforms.html.
2321         * Animometer/tests/dom/resources/compositing-transforms.js: Renamed from PerformanceTests/Animometer/tests/misc/resources/compositing-transforms.js.
2322
2323         Clean up miscellaneous test suite. Add a canvas ellipse test, and remove the other
2324         canvas tests.
2325
2326         * Animometer/resources/debug-runner/tests.js: Add ellipse tests to the simple suite.
2327         * Animometer/tests/misc/canvas-electrons.html: Removed.
2328         * Animometer/tests/misc/canvas-stars.html: Removed.
2329         * Animometer/tests/misc/resources/canvas-electrons.js: Removed.
2330         * Animometer/tests/misc/resources/canvas-stars.js: Removed.
2331         * Animometer/tests/simple/resources/simple-canvas-paths.js: Add ellipse primitives.
2332
2333         Merge text tests together into one.
2334
2335         * Animometer/resources/runner/tests.js: Remove international.html.
2336         * Animometer/tests/master/international.html: Removed.
2337         * Animometer/tests/master/resources/text.js:
2338         (animate): Update styling. Manually calculate gradients for the shadow particles.
2339         Reduce the step size for y direction to avoid cutting text off at the margins. Increase
2340         step for x since there will be a little more room.
2341         * Animometer/tests/master/text.html: Add more translations and lay it out in a table.
2342
2343 2016-03-20  Jon Lee  <jonlee@apple.com>
2344
2345         Add a link to show debug data
2346         https://bugs.webkit.org/show_bug.cgi?id=155724
2347
2348         Reviewed by Simon Fraser.
2349
2350         * Animometer/developer.html: Attach the onclick handler to the score.
2351         * Animometer/index.html: Ditto.
2352
2353 2016-03-18  Dean Jackson  <dino@apple.com>
2354
2355         Add a basic WebGL test to Animometer
2356         https://bugs.webkit.org/show_bug.cgi?id=155475
2357         <rdar://problem/25156860>
2358
2359         Reviewed by Simon Fraser and Jon Lee.
2360
2361         Add a pretty simple WebGL test to Animometer
2362         which tries to draw a number of animated triangles.
2363         The result still needs a bit of tuning.
2364
2365         * Animometer/resources/debug-runner/tests.js: Add a new
2366         Suite for "3D" tests.
2367         * Animometer/tests/3d/resources/webgl.js: Added.
2368         * Animometer/tests/3d/webgl.html: Added.
2369
2370 2016-03-18  Jon Lee  <jonlee@apple.com>
2371
2372         Add support for statically linking to a specific test
2373         https://bugs.webkit.org/show_bug.cgi?id=155631
2374
2375         Rubber-stamped by Darin Adler.
2376
2377         * Animometer/developer.html: Update to call restartBenchmark instead so that it works for
2378         both the preset and the manually set test sessions.
2379         * Animometer/resources/debug-runner/animometer.css:
2380         (.tree .link): Style the "link" UI.
2381         * Animometer/resources/debug-runner/animometer.js: Move updateDisplay from suitesManager
2382         to optionsManager since this is a visual option and has nothing to do with the suites
2383         settings. Update the construction of the test UI by including a "link" after each test.
2384         Clicking on that link brings up a JS prompt with a URL and a query string with the current
2385         parameters of the controller, and selected test. Pasting this URL into the location bar will
2386         automatically start running the selected test.
2387         (suitesManager.suitesFromQueryString): Iterate through the Suites and tests and find the one
2388         that matches the provided parameters. Returns an object similar to the form in
2389         suitesManager.updateLocalStorageFromUI.
2390         (benchmarkController.initialize): After settings up the events and options, try parsing the
2391         query string and running the benchmark immediately. Otherwise, fall back to the form.
2392         (benchmarkController.startBenchmark): Store the options and suites into member variables for
2393         reuse in restartBenchmark.
2394         (benchmarkController.startBenchmarkImmediatelyIfEncoded): Convert the query string to an object.
2395         If that's successful, find the suite and test referenced in the query string. Start the benchmark if
2396         the search for the test succeeded.
2397         * Animometer/resources/extensions.js:
2398         (Utilities.stripNonASCIICharacters): Helper method to convert the name of the suite and test
2399         into a query-string-friendly version.
2400         (Utilities.convertObjectToQueryString): Helper method to convert an object to query string
2401         format.
2402         (Utilities.convertQueryStringToObject): Helper method to convert query string into an object
2403         with properties and values.
2404
2405 2016-03-16  Jon Lee  <jonlee@apple.com>
2406
2407         Add a new benchmark test
2408         https://bugs.webkit.org/show_bug.cgi?id=155570
2409
2410         Reviewed by Simon Fraser.
2411
2412         New Leaves test includes various image sizes and opacity.
2413
2414         * Animometer/resources/debug-runner/tests.js: Add it to the HTML test suite.
2415         * Animometer/tests/dom/leaves.html: Added.
2416         * Animometer/tests/dom/resources/leaves.js: Added. Override the
2417         (Particle.call.reset): Uses a range of sizes, and opacity.
2418         (Particle.call.animate): Opacity goes up then down. When it hits 0, reset the particle.
2419         (Particle.call.move): Set transform and opacity.
2420         * Animometer/tests/master/resources/leaves.js: Get rid of the closure so that it
2421         can be used in this test. Update the relative path so that it works in both the master
2422         and dom test suite.
2423
2424 2016-03-15  Simon Fraser  <simon.fraser@apple.com>
2425
2426         Add developer Animometer test that bounces P3-tagged images
2427         https://bugs.webkit.org/show_bug.cgi?id=155511
2428
2429         Reviewed by Tim Horton.
2430         
2431         Add a test for rendering performance of tagged images. The 5 images are tagged
2432         with the Display P3 colorspace.
2433
2434         * Animometer/resources/debug-runner/tests.js:
2435         * Animometer/tests/bouncing-particles/bouncing-tagged-images.html: Added.
2436         * Animometer/tests/bouncing-particles/resources/bouncing-tagged-images.js: Added.
2437         * Animometer/tests/bouncing-particles/resources/image1.jpg: Added.
2438         * Animometer/tests/bouncing-particles/resources/image2.jpg: Added.
2439         * Animometer/tests/bouncing-particles/resources/image3.jpg: Added.
2440         * Animometer/tests/bouncing-particles/resources/image4.jpg: Added.
2441         * Animometer/tests/bouncing-particles/resources/image5.jpg: Added.
2442
2443 2016-03-09  Jon Lee  <jonlee@apple.com>
2444
2445         Enhance existing Animometer tests
2446         https://bugs.webkit.org/show_bug.cgi?id=155261
2447
2448         Reviewed by Simon Fraser.
2449
2450         * Animometer/tests/master/resources/canvas-tests.js: Add a gradient to
2451         the filled circles. To expose more of the gradient, add another circle.
2452         (SimpleCanvasStage.call.animate): For each frame, create a gradient with
2453         undulating stop points and colors. Fill the circles twice; once with
2454         the solid color, and once with the gradient.
2455
2456         * Animometer/tests/master/resources/dom-particles.js: Refactor the
2457         emission variables into a separate stage for this test. Add a colorOffset
2458         variable to make the colors of each particle slightly different, since
2459         the ramp controller can add large numbers of particles all at once, which
2460         would otherwise get all the same color.
2461         * Animometer/tests/master/resources/particles.js:
2462         (initialize): Remove the code specific to the SVG mask test.
2463
2464         * Animometer/tests/master/resources/image-data.js:
2465         (initialize): Remove unused local variable.
2466         * Animometer/tests/master/resources/multiply.js:
2467         (initialize): Make the test harder by adding more total particles.
2468
2469 2016-03-09  Jon Lee  <jonlee@apple.com>
2470
2471         Add text tests
2472         https://bugs.webkit.org/show_bug.cgi?id=155257
2473
2474         Reviewed by Simon Fraser.
2475
2476         * Animometer/resources/extensions.js:
2477         (UnitBezier.Utilities.createClass): Add a class that computes Bezier points
2478         assuming that two of the control points are at (0,0) and (1,1). Taken from
2479         WebCore/platform/graphics/UnitBezier.h
2480         * Animometer/resources/runner/tests.js: Add a test for English text and one
2481         for international text.
2482         * Animometer/tests/master/international.html: Added.
2483         * Animometer/tests/master/resources/text.js: Added. The test assumes there is
2484         #template div which it will copy. The copies are placed behind the template, and
2485         are set with different colors each frame. They are moved around with CSS transform.
2486         * Animometer/tests/master/text.html: Added.
2487
2488         Remove the other text tests, since these ones cover the same techniques.
2489         * Animometer/tests/text/layering-text.html: Removed.
2490         * Animometer/tests/text/resources/layering-text.js: Removed.
2491         * Animometer/tests/text/resources/text-boxes.js: Removed.
2492         * Animometer/tests/text/text-boxes.html: Removed.
2493         * Animometer/resources/debug-runner/tests.js:
2494
2495 2016-03-09  Jon Lee  <jonlee@apple.com>
2496
2497         Add a new image test
2498         https://bugs.webkit.org/show_bug.cgi?id=155232
2499
2500         Reviewed by Dean Jackson.
2501         Provisionally reviewed by Said Abou-Hallawa.
2502
2503         The image test renders PNGs and moves them with translate and rotate
2504         transforms. Each particle has a lifetime, and when the lifetime ends or
2505         the particle goes offscreen, it resets itself somewhere on the stage.
2506
2507         * Animometer/resources/debug-runner/tests.js: Remove the CSS bouncing PNG
2508         images test, because this one tests the same technique.
2509         * Animometer/resources/runner/tests.js: Added here as "Leaves".
2510         * Animometer/tests/master/leaves.html: Added.
2511         * Animometer/tests/master/resources/particles.js: Refactor out the parts
2512         specific to the DOM particles test. Consequently make velocity a public
2513         member.
2514         (initialize): The options parameter is never used, so remove it.
2515         (animate): Remove unused local variable.
2516         * Animometer/tests/master/resources/leaves.js: Maintains a focal point that
2517         moves back and forth across the canvas. That point affects the velocity
2518         of the particles. When the focal point is on the leftmost side, it is
2519         between [-6, -2], and set in reset(). When the focal point is on the rightmost
2520         side, it will be from [2, 6].
2521         * Animometer/tests/master/resources/dom-particles.js: Move JS specific to
2522         this test here from particles.js.
2523
2524         New images.
2525         * Animometer/tests/master/resources/compass100.png: Added.
2526         * Animometer/tests/master/resources/console100.png: Added.
2527         * Animometer/tests/master/resources/contribute100.png: Added.
2528         * Animometer/tests/master/resources/debugger100.png: Added.
2529         * Animometer/tests/master/resources/inspector100.png: Added.
2530         * Animometer/tests/master/resources/layout100.png: Added.
2531         * Animometer/tests/master/resources/performance100.png: Added.
2532         * Animometer/tests/master/resources/script100.png: Added.
2533         * Animometer/tests/master/resources/shortcuts100.png: Added.
2534         * Animometer/tests/master/resources/standards100.png: Added.
2535         * Animometer/tests/master/resources/storage100.png: Added.
2536         * Animometer/tests/master/resources/styles100.png: Added.
2537         * Animometer/tests/master/resources/timeline100.png: Added.
2538
2539 2016-03-03  Jon Lee  <jonlee@apple.com>
2540
2541         Add ability to retrieve raw data from release harness
2542         https://bugs.webkit.org/show_bug.cgi?id=155026
2543
2544         Reviewed by Simon Fraser.
2545
2546         * Animometer/developer.html: Remove the special UI in the debug
2547         harness.
2548         * Animometer/resources/runner/animometer.css: Add styles for the
2549         overlay.
2550         * Animometer/resources/runner/animometer.js: Let 'j' show the JSON
2551         results, but only if the overlay doesn't exist. Add 'esc' key to dismiss
2552         the overlay.
2553         (window.benchmarkController.selectResults): Cycle the cases around
2554         so that the first 's' press selects both the benchmark score and
2555         the individual test scores.
2556         * Animometer/resources/debug-runner/animometer.css: Remove
2557         unneeded rules.
2558         * Animometer/resources/debug-runner/animometer.js: Make the same
2559         call to handleKeyPress.
2560
2561 2016-03-03  Jon Lee  <jonlee@apple.com>
2562
2563         Make sure multiply test particles have at least some opacity
2564         https://bugs.webkit.org/show_bug.cgi?id=155027
2565
2566         Reviewed by Simon Fraser.
2567
2568         * Animometer/tests/master/resources/multiply.js:
2569         (initialize): Have the elements spin a little faster to make it
2570         more obvious when the system is being stressed.
2571         (_addTile):
2572         (animate): Make sure each element being animated has at least 1%
2573         opacity.
2574
2575 2016-03-03  Jon Lee  <jonlee@apple.com>
2576
2577         Update image test
2578         https://bugs.webkit.org/show_bug.cgi?id=154962
2579
2580         Rubber-stamped by Darin Adler.
2581
2582         * Animometer/tests/master/image-data.html: Make each canvas have a compositing layer.
2583         * Animometer/tests/master/resources/image-data.js:
2584         (initialize): Update to load multiple images serially.
2585         (_loadImage):
2586         (tune): Use display instead of visibility. When showing an element anew, move it somewhere else.
2587         (_createTestElement): Set up the element, refactor out placement of element to _refreshElement.
2588         (_refreshElement): Place element in a tile grid.
2589         (animate):
2590         (_getRandomNeighboringPixelIndex): Use a more conservative distribution to make the effect last a little longer.
2591         * Animometer/tests/master/resources/compass.svg: Added.
2592         * Animometer/tests/master/resources/console.svg: Added.
2593         * Animometer/tests/master/resources/contribute.svg: Added.
2594         * Animometer/tests/master/resources/debugger.svg: Added.
2595         * Animometer/tests/master/resources/inspector.svg: Added.
2596         * Animometer/tests/master/resources/layout.svg: Added.
2597         * Animometer/tests/master/resources/performance.svg: Added.
2598         * Animometer/tests/master/resources/script.svg: Added.
2599         * Animometer/tests/master/resources/shortcuts.svg: Added.
2600         * Animometer/tests/master/resources/standards.svg: Added.
2601         * Animometer/tests/master/resources/storage.svg: Added.
2602         * Animometer/tests/master/resources/styles.svg: Added.
2603         * Animometer/tests/master/resources/timeline.svg: Added.
2604
2605 2016-03-02  Jon Lee  <jonlee@apple.com>
2606
2607         Add some new controllers, and refine tests
2608         https://bugs.webkit.org/show_bug.cgi?id=154914
2609
2610         Reviewed by Simon Fraser.
2611
2612         Improve tests.
2613
2614         * Animometer/tests/master/focus.html: Move each particle into a layer. Put the particle
2615         in a container element. Expand the container element by the max blur radius, and clip
2616         overflow. This way, when the blur is applied, it avoids causing layer resizes.
2617         * Animometer/tests/master/resources/focus.js:
2618         (FocusElement.Utilities.createClass): Create a container element, and put the particle
2619         inside.
2620         (hide): Set display:none.
2621         (show): Set display:block.
2622         (animate): Apply filters to the container element.
2623         (FocusStage.call.initialize): Instead of inserting and removing elements from the DOM,
2624         keep them in the stage, but set the display style instead. Use this._offsetIndex to
2625         keep track of which elements are displayed.
2626         (FocusStage.call.animate): Move some calculations around to avoid doing unneeded math.
2627         (FocusStage.call.getBlurValue): Make sure elements are always blurred.
2628         (FocusStage.call.getOpacityValue): Make sure elements have some opacity.
2629         * Animometer/tests/master/resources/multiply.js: Have the particles on the edge of the
2630         stage appear less black when the complexity gets large enough.
2631         (tune): Have this._distanceFactor calculate the factor. Avoid calculating square root each
2632         frame.
2633
2634 2016-03-02  Jon Lee  <jonlee@apple.com>
2635
2636         Add some new controllers, and refine tests
2637         https://bugs.webkit.org/show_bug.cgi?id=154914
2638
2639         Reviewed by Simon Fraser.
2640
2641         Add a controller that centers around 30 fps instead of 60 fps.
2642
2643         * Animometer/developer.html: Add a new option.
2644         * Animometer/resources/debug-runner/animometer.js:
2645         * Animometer/resources/runner/animometer.js:
2646         (this._processData.findRegression): When calculating the complexity-frameLength regression,
2647         check the controller, and use a 30 fps baseline if needed.
2648         * Animometer/resources/statistics.js:
2649         (Regression.Utilities.createClass): Update to allow clients to specify the baseline
2650         frame length.
2651         * Animometer/tests/resources/main.js:
2652         (tune): Override some of the constants in RampController. Move those constants out for
2653         easier reading.
2654
2655         Add a fixed controller, with no step.
2656
2657         * Animometer/developer.html: Add a controller that takes no step.
2658         * Animometer/resources/debug-runner/animometer.js:
2659         * Animometer/tests/resources/main.js:
2660         (Rotater.Utilities.createClass):
2661
2662         Switch to ramp controller as default.
2663
2664         * Animometer/developer.html: Increase the test length to 20 seconds.
2665         * Animometer/resources/debug-runner/animometer.js: Rename the "adjustment" field to "controller"
2666         since that is a more accurate description.
2667         * Animometer/resources/debug-runner/graph.js:
2668         * Animometer/resources/runner/animometer.js: Update preferences for release suite.
2669
2670 2016-02-26  Jon Lee  <jonlee@apple.com>
2671
2672         Address Dean's comments in 154673.
2673
2674         * Animometer/developer.html:
2675         * Animometer/resources/debug-runner/animometer.js:
2676         * Animometer/resources/debug-runner/graph.js:
2677
2678 2016-02-25  Jon Lee  <jonlee@apple.com>
2679
2680         Update animation benchmark and tests
2681         https://bugs.webkit.org/show_bug.cgi?id=154673
2682
2683         Reviewed by Dean Jackson.
2684
2685         Update test visuals.
2686
2687         * Animometer/tests/master/resources/canvas-tests.js:
2688         (CanvasLineSegment.Utilities.createClass): Line segments near the edge of the stage get clipped.
2689         Update circle position and radius to minimize impact.
2690         * Animometer/tests/master/resources/particles.js: Get rid of rotating gradient background, and
2691         have 3 locations for emitting particles.
2692         * Animometer/tests/master/resources/multiply.js: Update distance metric so that fringe tiles get
2693         more color.
2694         (initialize): Fix some of the math for laying out the tiles.
2695         (animate): When a tile isn't used, set visibility: hidden.
2696
2697 2016-02-24  Jon Lee  <jonlee@apple.com>
2698
2699         Update animation benchmark and tests
2700         https://bugs.webkit.org/show_bug.cgi?id=154673
2701
2702         Reviewed by Dean Jackson.
2703
2704         Update the ramp controller.
2705
2706         The controller refines the complexity interval to test across.
2707
2708         * Animometer/resources/statistics.js: Add functions that estimate cumulative distribution function.
2709         (Regression): For the flat regression, force the first segment to be at 60 fps.
2710         (valueAt): Add convenience function to return interpolated value based on the regression used.
2711         (_calculateRegression): Include the number of points included for both segments, and the piecewise
2712         errors.
2713         * Animometer/tests/resources/math.js: Make the Kalman estimator subclass Experiment, and allow it
2714         to be reset.
2715
2716         * Animometer/tests/resources/main.js: Initialize the tier such that it starts at 10^0 = 1.
2717         Increase the number of ramps. Maintain three FPS thresholds-- the frame rate of interest, a limit
2718         on the lowest FPS we care to go for later interpolation, and a minimum FPS threshold we want to
2719         aim for each ramp. Also keep three estimators: a running average of the change point, a minimum
2720         boundary for each ramp, and an estimator for all the frames within an interval. The first two
2721         are used to determine the parameters of the next ramp, and the latter allows us to refine the
2722         parameters.
2723         (update): During the tier phase, it is possible that the highest complexity possible for a test
2724         won't stress the system enough to trigger stopping the tier phase and transitioning to the ramps.
2725         If the complexity doesn't change when going to the next tier, we've maxed the test out, and move
2726         on. When the tier phase completed, turn off Controller.frameLengthEstimator, which estimates the
2727         FPS at each tier.
2728         (tune): At each interval, look at the confidence distribution of being on the 60 FPS side or the
2729         slow side. If the slowest FPS we achieve at the ramp's maximum complexity is not at least
2730         _fpsRampSlowThreshold, then increase the maximum complexity. If we ever achieve 60 FPS, increase
2731         the ramp's minimum complexity to that level. If, at an even lower complexity, a glitch causes the
2732         FPS to drop, we reset the minimum complexity.
2733
2734         Have the bootstrap calculation occur between tests. Clean up harness.
2735
2736         * Animometer/resources/debug-runner/animometer.js: Run bootstrap after a test has
2737         completed to avoid doing all of it at the end before showing the results. Clean up
2738         parameters being passed around.
2739         * Animometer/resources/debug-runner/tests.js:
2740         (text):
2741         * Animometer/resources/runner/animometer.js:
2742         (this._processData.calculateScore): Save the results to the same object holding the data.
2743         (this._processData._processData): In the case where a file is dragged, calculate the score
2744         serially. Grab the results object and move it to the results variable and remove it from
2745         the data object. This avoids serializing the results into the JSON.
2746         (this._processData.findRegression): Include the samples used for bootstrapping. Reduce the
2747         resample size to shorten the wait.
2748         * Animometer/resources/runner/benchmark-runner.js:
2749         * Animometer/resources/statistics.js:
2750         (bootstrap): Update how bootstrapData is sorted. In some regression results the mix of
2751         floats and integers causes an alphabetical sort to occur.
2752         * Animometer/resources/strings.js:
2753
2754         Add meta charset so that encodings between harness and test match.
2755
2756         * Animometer/tests/bouncing-particles/bouncing-canvas-images.html:
2757         * Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html:
2758         * Animometer/tests/bouncing-particles/bouncing-css-images.html:
2759         * Animometer/tests/bouncing-particles/bouncing-css-shapes.html:
2760         * Animometer/tests/bouncing-particles/bouncing-svg-images.html:
2761         * Animometer/tests/bouncing-particles/bouncing-svg-shapes.html:
2762         * Animometer/tests/master/canvas-stage.html:
2763         * Animometer/tests/master/focus.html:
2764         * Animometer/tests/master/image-data.html:
2765         * Animometer/tests/master/multiply.html:
2766         * Animometer/tests/master/particles.html:
2767         * Animometer/tests/misc/canvas-electrons.html:
2768         * Animometer/tests/misc/canvas-stars.html:
2769         * Animometer/tests/misc/compositing-transforms.html:
2770         * Animometer/tests/simple/simple-canvas-paths.html:
2771         * Animometer/tests/simple/tiled-canvas-image.html:
2772         * Animometer/tests/template/template-canvas.html:
2773         * Animometer/tests/template/template-css.html:
2774         * Animometer/tests/template/template-svg.html:
2775         * Animometer/tests/text/layering-text.html:
2776         * Animometer/tests/text/text-boxes.html:
2777
2778         Update test harness reporting.
2779
2780         * Animometer/developer.html: Add missing meta charset.
2781         * Animometer/index.html: Remove unnecessary utf-8 declaration.
2782         * Animometer/resources/debug-runner/animometer.css: Add convenience classes for
2783         formatting the results table.
2784         * Animometer/resources/debug-runner/animometer.js: Adjust which stats are shown.
2785         * Animometer/resources/debug-runner/tests.js: Display bootstrapping statistics.
2786         * Animometer/resources/strings.js: Move strings not used by the release harness.
2787
2788         Switch to a pseudo-random number generator.
2789
2790         * Animometer/resources/statistics.js: Add a Pseudo class, with a simple
2791         pseudo-random number generator.
2792         (_calculateRegression): Reset the generator before running bootstrap.
2793         (bootstrap): Deleted.
2794
2795         Replace Math.random with Pseudo.random.
2796         * Animometer/tests/master/resources/canvas-tests.js:
2797         * Animometer/tests/master/resources/focus.js:
2798         * Animometer/tests/master/resources/particles.js:
2799         * Animometer/tests/resources/main.js:
2800
2801         Use bootstrapping to get confidence interval in the breakpoint.
2802
2803         For the ramp controller, calculate the piecewise regression, and then use
2804         bootstrapping in order to find the 95% confidence interval. Use the raw data.
2805
2806         * Animometer/developer.html: Default to the complexity graph. Add a legend
2807         checkbox to toggle visibility of the bootstrap score and histogram.
2808         * Animometer/resources/debug-runner/animometer.css: Make some more space to show
2809         the old raw and average scores in the legend. Add new styles for the data.
2810         * Animometer/resources/debug-runner/graph.js:
2811         (_addRegressionLine): Allow passing an array for the variance bar tied to the
2812         regression line. Now |stdev| is |range|.
2813         (createComplexityGraph): Add bootstrap median, and overlay a histogram of
2814         the bootstrap samples. Switch raw samples from circles to X's.
2815         (onComplexityGraphOptionsChanged): Allow toggling of the bootstrap data.
2816         (onGraphTypeChanged): Move the regressions for the raw and average samples to the
2817         legend. In the subtitle use the bootstrap median, and include the 95% confidence
2818         interval.
2819         * Animometer/resources/runner/animometer.js:
2820         (this._processData.findRegression): Factor out the code that determines which
2821         samples to include when calculating the piecewise regression. For series that have
2822         many samples, or a wider range of recorded complexities, throw away the 2.5%
2823         lowest and highest samples before calculating the regression. Keep all samples
2824         if the number of samples to regress is small or the range of complexities is
2825         narrow.
2826         (this._processData._calculateScore): Factor out regression calculation to
2827         findRegression(). Bootstrap the change point of the regression. The score is the
2828         median.
2829         * Animometer/resources/statistics.js:
2830         (_calculateRegression): Correct an issue in the calculation of the regression, where
2831         the denominator can be 0.
2832         (bootstrap): Template for bootstrapping. Create a bootstrap sample array, Create
2833         re-samples by random selection with replacement. Return the 95% confidence samples,
2834         the bootstrap median, mean, and the data itself.
2835         * Animometer/resources/strings.js: Add bootstrap.
2836         * Animometer/tests/resources/main.js:
2837         (processSamples): Don't prematurely cut the sample data.
2838
2839         Fix graph drawing.
2840
2841         * Animometer/resources/debug-runner/animometer.js: Add spacing in the JSON output.
2842         Multiple tests output a lot of JSON and can hang when selecting JSON with no whitespace.
2843         * Animometer/resources/debug-runner/animometer.css:
2844         (#complexity-graph .series.raw circle): Update the color.
2845         * Animometer/resources/debug-runner/graph.js: Use the FPS axis instead of the
2846         complexity axis, which can vary in domain. For determining the complexity domain,
2847         only use samples after samplingTimeOffset.
2848
2849         Allow dropping results JSON.
2850
2851         * Animometer/developer.html: Add a button.
2852         * Animometer/resources/debug-runner/animometer.css:
2853         * Animometer/resources/debug-runner/animometer.js: Read the data and go straight
2854         to the dashboard. With JSON output, write out only the options and the raw data.
2855
2856         Teach the harness to evaluate the samples and determine the test score.
2857
2858         This will allow us to update how the score is calculated separately from the samples recorded.
2859         This also prepares the harness to be able to accept JSON of prior runs.
2860
2861         * Animometer/resources/strings.js: Clean up and remove unneeded strings and reduce some of the
2862         hierarchy.
2863         * Animometer/resources/debug-runner/tests.js: Update to use the new strings.
2864
2865         * Animometer/tests/resources/main.js: Allow all controllers to show a complexity-FPS graph.
2866         (_processComplexitySamples): Factor out some of the sample processing done in the ramp
2867         controller for the benefit of the other controllers. |complexitySamples| contains a list of
2868         samples. Sort the samples by complexity. Optionally remove the top x% of samples.
2869         Group them, and calculate distribution of samples within the same complexity, and add those as
2870         new entries into |complexityAverageSamples|.
2871         (Controller.processSamples): Move the code responsible for determining the complexity and FPS
2872         scores out to ResultsDashboard. The structure of the data returned by the controller is:
2873
2874         {
2875             controller: [time-regression, time-regression, ...], // optional, data specific to controller
2876             marks: [...],
2877             samples: {                    // all of the sample data
2878                 controller: [...],
2879                 complexity: [...],        // processed from controller samples
2880                 complexityAverage: [...], // processed from complexity samples
2881             }
2882         }
2883
2884         (AdaptiveController.processSamples): Adding the target frame length is no longer necessary; we
2885         now pass the test options to the graph.
2886         (Regression): Move to statistics.js.
2887         * Animometer/resources/statistics.js: Move Regression to here. Add a check if the sampling range
2888         only contains one sample, since we cannot calculate a regression from one sample point.
2889
2890         Teach the test harness to evaluate the data.
2891         * Animometer/resources/runner/animometer.js:
2892         (ResultsDashboard): Store the options used to run the test and the computed results/score separately
2893         from the data. The results are stored as:
2894
2895         {
2896             score: /* geomean of iteration score */,
2897             iterationsResults: [
2898                 {
2899                     score: /* geomean of tests */,
2900                     testsResults: {
2901                         suiteName: {
2902                             testName: {
2903                                 controller: {
2904                                     average:
2905                                     concern:
2906                                     stdev:
2907                                     percent:
2908                                 },
2909                                 frameLength: { ... },
2910                                 complexity: {
2911                                     complexity:
2912                                     stdev:
2913                                     segment1:
2914                                     segment2:
2915                                 },
2916                                 complexityAverage: { ... }
2917                             },
2918                             testName: { ... },
2919                         },
2920                         ... next suite ...
2921                     }
2922                 },
2923                 { ...next iteration... }
2924             ]
2925         }
2926
2927         * Animometer/resources/debug-runner/animometer.js: Pass options around instead of relying
2928         on what was selected in the form. This will later allow for dropping previous results, and
2929         using those runs' options when calculating scores.
2930         (ResultsTable._addGraphButton): Simplify button action by using attached test data.
2931         * Animometer/resources/debug-runner/graph.js: Refactor to use the data.
2932
2933         Consolidate JS files, and move statistics out to a separate JS.
2934
2935         Preparation for having the Controller only handle recording and storage of the samples,
2936         and leave the evaluation of the test score out to the harness. Move Experiment to
2937         a new statistics.js, where Regression will also eventually go. Get rid of algorithm.js
2938         and move it to utilities.js since the Heap is used only for Experiments.
2939
2940         * Animometer/tests/resources/algorithm.js: Removed. Heap is in utilities.js.
2941         * Animometer/tests/resources/sampler.js: Removed. Experiment is in statistics.js,
2942         Sampler in main.js.
2943         * Animometer/tests/resources/main.js: Move Sampler here.
2944         * Animometer/resources/statistics.js: Added. Move Statistics and Experiment here.
2945         * Animometer/resources/extensions.js: Move Heap here. Attach static method to create
2946         a max or min heap to Heap, instead of a new Algorithm object.
2947
2948         Update JS files.
2949         * Animometer/developer.html:
2950         * Animometer/index.html:
2951         * Animometer/tests/bouncing-particles/bouncing-canvas-images.html:
2952         * Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html:
2953         * Animometer/tests/bouncing-particles/bouncing-css-images.html:
2954         * Animometer/tests/bouncing-particles/bouncing-css-shapes.html:
2955         * Animometer/tests/bouncing-particles/bouncing-svg-images.html:
2956         * Animometer/tests/bouncing-particles/bouncing-svg-shapes.html:
2957         * Animometer/tests/master/canvas-stage.html:
2958         * Animometer/tests/master/focus.html:
2959         * Animometer/tests/master/image-data.html:
2960         * Animometer/tests/master/multiply.html:
2961         * Animometer/tests/master/particles.html:
2962         * Animometer/tests/misc/canvas-electrons.html:
2963         * Animometer/tests/misc/canvas-stars.html:
2964         * Animometer/tests/misc/compositing-transforms.html:
2965         * Animometer/tests/simple/simple-canvas-paths.html:
2966         * Animometer/tests/simple/tiled-canvas-image.html:
2967         * Animometer/tests/template/template-canvas.html:
2968         * Animometer/tests/template/template-css.html:
2969         * Animometer/tests/template/template-svg.html:
2970         * Animometer/tests/text/layering-text.html:
2971         * Animometer/tests/text/text-boxes.html:
2972
2973         Fix the cursor in the graph analysis when the min
2974         complexity is not 0.
2975
2976         * Animometer/resources/debug-runner/graph.js:
2977         (_addRegression):
2978         (createComplexityGraph):
2979
2980 2016-02-23  Geoffrey Garen  <ggaren@apple.com>
2981
2982         Fix some issues in MallocBench
2983         https://bugs.webkit.org/show_bug.cgi?id=154600
2984
2985         Reviewed by Oliver Hunt.
2986
2987         * MallocBench/MallocBench.xcodeproj/project.pbxproj: Make nimlang.ops
2988         a part of the build so that it copies to the right place for execution.
2989
2990         * MallocBench/MallocBench/Interpreter.cpp:
2991         (Interpreter::Interpreter): Don't try to open .ops files for writing
2992         because we're only going to read and we might need extra permissions to
2993         write.
2994
2995         (Interpreter::~Interpreter): Give more context when opening a file fails
2996         to help with debugging.
2997
2998         * MallocBench/MallocBench/stress.cpp:
2999         (benchmark_stress): Reduce iterations to complete in less than 10 seconds.
3000
3001 2016-02-23  Michael Saboff  <msaboff@apple.com>
3002
3003         Unreviewed change to revert extraneous changes made part of change set 196955.
3004
3005         * MallocBench/MallocBench/Interpreter.cpp:
3006         (Interpreter::doMallocOp):
3007         (Interpreter::Thread::switchTo):
3008         (writeData): Deleted.
3009
3010 2016-02-19  Michael Saboff  <msaboff@apple.com>
3011
3012         MallocBench: Added recording for nimlang website, new recording details and added new options
3013         https://bugs.webkit.org/show_bug.cgi?id=154485
3014
3015         Reviewed by Geoff Garen.
3016
3017         Added new capabilities to MallocBench.  These include:
3018             Added a recording of http://nim-lang.org/docs/lib.html.
3019             Added thread id to the recording and the ability to playback switching threads in MallocBench
3020             Added aligned allocations to recordings and the ability to playback
3021             Added --use-thread-id option to honor recorded thread ids
3022             Added --detailed-report to output remaining allocations by size after playback
3023             Added --no-warmup to not run the warm up iteration
3024
3025         Changed the way that options are passed down to the benchmarks.  Instead of passing individual
3026         boolean or numeric option values, just pass a reference the CommandLine itself.  Each benchmark
3027         can access the options that are appropriate.  The Benchmark class also uses the options for
3028         is parallel, run counts and warm up.
3029
3030         Added thread id and aligned malloc to the Op by noticing that structure padding and Opcode allowed
3031         for another 32 bits of data.  Breaking that unused 32 bits into a 16 bit thread id value and a
3032         16 bit log base 2 of the alignment for aligned malloc allowed for existing recordings to playback
3033         without any incompatibilities.
3034
3035         Threaded operation is simulated by creating threads as needed.  As long as the next Op's thread id
3036         is the same as the last, operation continues as normal.  When the next Op has a different thread id,
3037         we switch to that thread using the shared Op stream to continue playing back.  There is a mutex to
3038         assure that only one thread is really running at a time and a condition variable used to wait
3039         that the current thread id matches each block thread's thread id.  This doesn't simulate true
3040         concurrent threading, but is instead plays back Ops recorded for multiple thread faithfully.
3041
3042         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
3043         * MallocBench/MallocBench/Benchmark.cpp:
3044         (deallocateHeap):
3045         (Benchmark::Benchmark):
3046         (Benchmark::runOnce):
3047         (Benchmark::run):
3048         * MallocBench/MallocBench/Benchmark.h:
3049         (Benchmark::isValid):
3050         * MallocBench/MallocBench/CommandLine.cpp:
3051         (CommandLine::printUsage):
3052         * MallocBench/MallocBench/CommandLine.h:
3053         (CommandLine::isValid):
3054         (CommandLine::benchmarkName):
3055         (CommandLine::isParallel):
3056         (CommandLine::useThreadID):
3057         (CommandLine::detailedReport):
3058         (CommandLine::warmUp):
3059         (CommandLine::heapSize):
3060         (CommandLine::runs):
3061         * MallocBench/MallocBench/Interpreter.cpp:
3062         (Interpreter::Interpreter):
3063         (Interpreter::run):
3064         (Interpreter::readOps):
3065         (Interpreter::doOnSameThread):
3066         (Interpreter::switchToThread):
3067         (Interpreter::detailedReport):
3068         (compute2toPower):
3069         (writeData):
3070         (Interpreter::doMallocOp):
3071         (Interpreter::Thread::Thread):
3072         (Interpreter::Thread::stop):
3073         (Interpreter::Thread::~Thread):
3074         (Interpreter::Thread::runThread):
3075         (Interpreter::Thread::waitToRun):
3076         (Interpreter::Thread::switchTo):
3077         * MallocBench/MallocBench/Interpreter.h:
3078         (Interpreter::Thread::isMainThread):
3079         * MallocBench/MallocBench/alloc_free.cpp: Added.
3080         (benchmark_alloc_free):
3081         * MallocBench/MallocBench/alloc_free.h: Added.
3082         * MallocBench/MallocBench/balloon.cpp:
3083         (benchmark_balloon):
3084         * MallocBench/MallocBench/balloon.h:
3085         * MallocBench/MallocBench/big.cpp:
3086         (benchmark_big):
3087         * MallocBench/MallocBench/big.h:
3088         * MallocBench/MallocBench/churn.cpp:
3089         (benchmark_churn):
3090         * MallocBench/MallocBench/churn.h:
3091         * MallocBench/MallocBench/facebook.cpp:
3092         (benchmark_facebook):
3093         * MallocBench/MallocBench/facebook.h:
3094         * MallocBench/MallocBench/flickr.cpp:
3095         (benchmark_flickr):
3096         (benchmark_flickr_memory_warning):
3097         * MallocBench/MallocBench/flickr.h:
3098         * MallocBench/MallocBench/fragment.cpp:
3099         (validate):
3100         (benchmark_fragment):
3101         (benchmark_fragment_iterate):
3102         * MallocBench/MallocBench/fragment.h:
3103         * MallocBench/MallocBench/list.cpp:
3104         (benchmark_list_allocate):
3105         (benchmark_list_traverse):
3106         * MallocBench/MallocBench/list.h:
3107         * MallocBench/MallocBench/main.cpp:
3108         (main):
3109         * MallocBench/MallocBench/medium.cpp:
3110         (benchmark_medium):
3111         * MallocBench/MallocBench/medium.h:
3112         * MallocBench/MallocBench/memalign.cpp:
3113         (test):
3114         (benchmark_memalign):
3115         * MallocBench/MallocBench/memalign.h:
3116         * MallocBench/MallocBench/message.cpp:
3117         (benchmark_message_one):
3118         (benchmark_message_many):
3119         * MallocBench/MallocBench/message.h:
3120         * MallocBench/MallocBench/nimlang.cpp: Added.
3121         (benchmark_nimlang):
3122         * MallocBench/MallocBench/nimlang.h: Added.
3123         * MallocBench/MallocBench/nimlang.ops: Added.
3124         * MallocBench/MallocBench/realloc.cpp:
3125         (benchmark_realloc):
3126         * MallocBench/MallocBench/realloc.h:
3127         * MallocBench/MallocBench/reddit.cpp:
3128         (benchmark_reddit):
3129         (benchmark_reddit_memory_warning):
3130         * MallocBench/MallocBench/reddit.h:
3131         * MallocBench/MallocBench/stress.cpp:
3132         (deallocate):
3133         (benchmark_stress):
3134         * MallocBench/MallocBench/stress.h:
3135         * MallocBench/MallocBench/stress_aligned.cpp:
3136         (benchmark_stress_aligned):
3137         * MallocBench/MallocBench/stress_aligned.h:
3138         * MallocBench/MallocBench/theverge.cpp:
3139         (benchmark_theverge):
3140         (benchmark_theverge_memory_warning):
3141         * MallocBench/MallocBench/theverge.h:
3142         * MallocBench/MallocBench/tree.cpp:
3143         (benchmark_tree_allocate):
3144         (benchmark_tree_traverse):
3145         (benchmark_tree_churn):
3146         * MallocBench/MallocBench/tree.h:
3147         * MallocBench/run-malloc-benchmarks:
3148
3149 2016-02-11  Jon Lee  <jonlee@apple.com>
3150
3151         Fix a missing refactoring.
3152
3153         * Animometer/tests/master/resources/multiply.js:
3154         (animate): Move to Utilities.lerp.
3155         (_lerp): Deleted.
3156
3157 2016-02-10  Jon Lee  <jonlee@apple.com>
3158
3159         Add new benchmark tests.
3160         https://bugs.webkit.org/show_bug.cgi?id=154063
3161
3162         Provisionally reviewed by Said Abou-Hallawa.
3163
3164         Add tests for get/put image data, filters, opacity, and css transforms.
3165
3166         * Animometer/resources/runner/benchmark-runner.js:
3167         (_runBenchmarkAndRecordResults): Update the body background color to match that of
3168         the stage.
3169         (this._runNextIteration): Clear the background color style for the results page.
3170         * Animometer/resources/runner/tests.js:
3171         * Animometer/tests/master/focus.html: Added.
3172         * Animometer/tests/master/image-data.html: Added.
3173         * Animometer/tests/master/multiply.html: Added.
3174         * Animometer/tests/master/resources/focus.js: Added.
3175         * Animometer/tests/master/resources/image-data.js: Added.
3176         * Animometer/tests/master/resources/multiply.js: Added.
3177         * Animometer/tests/master/resources/stage.css: Move common styles out.
3178         * Animometer/tests/resources/main.js: Update Stage.randomBool to use Math.random.
3179         Add Stage.randomSign for randomly setting a direction. Add the notion of the
3180         current timestamp of the test to Benchmark, since some animations cycle through
3181         colors and rely on an incremental counter like the time.
3182
3183 2016-02-09  Said Abou-Hallawa  <sabouhallawa@apple.com>
3184
3185         Add internal benchmark tests for CSS mix-blend-modes and filters
3186         https://bugs.webkit.org/show_bug.cgi?id=154058
3187
3188         Provisionally reviewed by Jon Lee.
3189
3190         * Animometer/resources/debug-runner/tests.js: Include the new tests in the
3191         "HTML suite" of the debug runner.
3192
3193         * Animometer/resources/extensions.js:
3194         (Utilities.browserPrefix):
3195         (Utilities.setElementPrefixedProperty): Utility functions to allow setting
3196         prefixed style properties.
3197
3198         * Animometer/tests/bouncing-particles/resources/bouncing-css-shapes.js:
3199         Set the mix-blend-mode and the filter to some random values if the options
3200         of the test requested that.
3201
3202         * Animometer/tests/bouncing-particles/resources/bouncing-particles.js:
3203         (parseShapeParameters): Parse the url options "blend" and "filter" and set
3204         the corresponding flags.
3205
3206         * Animometer/tests/resources/main.js:
3207         (randomStyleMixBlendMode):
3208         (randomStyleFilter): Return random mix-blend-mode and filter.
3209
3210 2016-02-08  Jon Lee  <jonlee@apple.com>
3211
3212         Add a ramp controller
3213         https://bugs.webkit.org/show_bug.cgi?id=154028
3214
3215         Provisionally reviewed by Said Abou-Hallawa.
3216
3217         Enhance the graph to include a complexity-fps graph, in addition
3218         to the time graph.
3219
3220         * Animometer/developer.html: Add a ramp option.
3221         * Animometer/resources/debug-runner/animometer.css: Update the style.
3222         * Animometer/resources/strings.js: Flatten the Strings.text constants.
3223         * Animometer/resources/debug-runner/animometer.js:
3224         (ResultsTable.call._addGraphButton): Refactor.
3225         (ResultsTable.call._addTest): Add regression data.
3226         (benchmarkController): Add a form that allows the user to switch between the two forms,
3227         Add a form that allows the user to toggle different data. Hide certain header columns
3228         depending on the selected controller.
3229         * Animometer/resources/debug-runner/graph.js: Add the complexity regressions.
3230         * Animometer/resources/debug-runner/tests.js: Add headers for the ramp results.
3231         * Animometer/resources/runner/animometer.js:
3232         (ResultsTable): If a header is disabled don't include them in _flattenedHeaders.
3233         * Animometer/tests/resources/main.js:
3234         (Controller): Allow options to specify the capacity for sample arrays.
3235         (Regression): A piecewise regression that tries to fit a slope and a flat profile.
3236         (_calculateRegression): Options can fix the slope and bias when calculating the minimal
3237         error. Sweep across the samples in time (which could be backward depending on the controller)
3238         and calculate the intersection point.
3239         (RampController): This controller assumes that the target frame rate is below
3240         58 FPS. It runs in two stages. The first stage quickly determines the order of
3241         magnitude of objects needed to stress the system by the setting the complexity
3242         to increasingly difficult tiers. Perform a series of ramps descending from a
3243         high-water mark of complexity. The complexity needed to reach the target frame
3244         length is done by performing a piecewise regression on each ramp, and track a
3245         running average of these values. For the next ramp, make that running average
3246         the center of the ramp. With a minimum complexity of 0, the high-water mark is
3247         twice that average. The score is based on the highest complexity that can
3248         reach 60 fps.
3249
3250 2016-02-08  Jon Lee  <jonlee@apple.com>
3251
3252         Address Said's comments on the benchmark, and do some clean up.
3253
3254         * Animometer/developer.html:
3255         * Animometer/resources/debug-runner/animometer.css: Add styles for averages.
3256         * Animometer/resources/debug-runner/animometer.js: Use the right
3257         Strings constants.
3258         * Animometer/resources/debug-runner/graph.js:
3259         (_addRegressionLine): Add missing code to draw the line and standard
3260         deviation highlight.
3261         (onGraphTypeChanged): Remove unneeded variables
3262         (onTimeGraphOptionsChanged):
3263         * Animometer/resources/runner/benchmark-runner.js:
3264         (_runBenchmarkAndRecordResults): Rename samplers to suiteResults and
3265         _suitesSamplers to _suitesResults.
3266         * Animometer/tests/resources/main.js:
3267         (results): Call processSamples().
3268         (update): Change sampling timestamp comparison.
3269         (_animateLoop): Move shouldStop call to before the update.
3270         * Animometer/tests/resources/sampler.js:
3271         (process): Rename to processSamples().
3272
3273 2016-02-07  Jon Lee  <jonlee@apple.com>
3274
3275         Teach Controller to measure intervals, and turn off the frame length estimator.
3276
3277         * Animometer/tests/resources/main.js: Default interval length is 100 ms.
3278         (start): Set the first interval.
3279         (_measureAndResetInterval): Reports the average frame length of the interval that just
3280         completed, and sets up the next interval.
3281         (update): If there is no length, then just use the estimator per frame, otherwise the
3282         estimator measures per interval. Add a didFinishInterval for subclasses to process
3283         prior to recording the sample. Update tune() to include whether an interval had
3284         finished.
3285         (StepController): Step controllers don't measure on an interval basis.
3286
3287 2016-02-07  Jon Lee  <jonlee@apple.com>
3288
3289         Minor refactoring. Rename Controller._estimator to Controller._frameLengthEstimator
3290         and switch the parameters for start(), update(), and tune(), so that the timestamp
3291         is first and stage is second.
3292
3293         * Animometer/tests/resources/main.js:
3294
3295 2016-02-07  Jon Lee  <jonlee@apple.com>
3296
3297         Move ResultsTable functionality not needed for release tests out.
3298         Move reporting of score and mean to selection of the time-based graph.
3299
3300         * Animometer/developer.html: Rename graph-options to time-graph-options.
3301         * Animometer/resources/debug-runner/animometer.js:
3302         (DeveloperResultsTable): Moved from runner/animometer.js. Switch from mean
3303         values to "average" objects which can hold stdev. Move graph button and
3304         calculation of noisy measurements here. Sophisticated header processing
3305         is not needed in release suite.
3306         (populateTable): Use DeveloperResultsTable.
3307         * Animometer/resources/debug-runner/graph.js: Pull time graph creation to
3308         its own function, and add a new onGraphTypeChanged handler in preparation
3309         of a complexity graph to be added later.
3310         * Animometer/resources/runner/animometer.js:
3311         (ResultsTable): Simplify to just handle test names and scores.
3312
3313 2016-02-07  Jon Lee  <jonlee@apple.com>
3314
3315         Tests: reuse objects already made.
3316
3317         Avoid thrash of object creation and removal by maintaining an index that
3318         moves along the array as the adjust values change. If the tune value
3319         requires more objects than the maximum size of the object array, then create
3320         new objects. This means that the object array size never decreases.
3321
3322         * Animometer/tests/master/resources/canvas-stage.js: Maintain a separate
3323         offsetIndex. For these tests, we want to avoid drawing the oldest objects,
3324         so the scene will draw the object at offsetIndex to the end of the array.
3325         (tune): Reverse the logic since "removal" of objects is much simpler and
3326         involves simply changing the offsetIndex.
3327         (animate): Update the for loop to draw from offsetIndex to the end.
3328         (complexity): Update the definition.
3329         * Animometer/tests/master/resources/canvas-tests.js: Maintain a separate
3330         offsetIndex. For these tests, we want to avoid drawing the newest objects,
3331         so the scene will draw the object at index 0 to the object at offsetIndex.
3332         (SimpleCanvasStage.animate): Fly-by removal of local stage variable,
3333         which is unneeded. Update the for loop to draw from offsetIndex to the end.
3334         * Animometer/tests/simple/resources/simple-canvas-paths.js:
3335         (SimpleCanvasStage.animate): Update the for loop to draw from 0 to
3336         offsetIndex.
3337         * Animometer/tests/simple/resources/simple-canvas.js:
3338         (tune): Update logic. Here, offsetIndex represents the boundary of the last
3339         index to render.
3340         (animate): Update the for loop to draw from 0 to offsetIndex.
3341         (complexity): Update the definition.
3342
3343 2016-02-07  Jon Lee  <jonlee@apple.com>
3344
3345         Tests: refactor and update styles.
3346
3347         * Animometer/tests/resources/main.js: Add helper methods that return
3348         a color that hue rotates based on the date, and a counter value that
3349         increases based on the date. Fix randomInt() to not bias against the min and
3350         max values.
3351
3352         * Animometer/tests/master/resources/canvas-tests.js: Use new helper methods.
3353         * Animometer/tests/master/resources/dom-particles.js: Ditto.
3354         * Animometer/tests/master/resources/particles.js: Ditto.
3355         * Animometer/tests/simple/resources/simple-canvas-paths.js: Refactor to
3356         use a rotating color instead of a random color. The fast switching of color
3357         is too vivid to watch.
3358
3359         * Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js:
3360         (BouncingSvgParticlesStage.call.createGradient): Fix the gradient so
3361         that the last stop is located at the end.
3362
3363 2016-02-07  Jon Lee  <jonlee@apple.com>
3364
3365         Refactor tune() to not return the complexity of the scene.
3366
3367         We have stage.complexity() now, so returning the complexity through tune
3368         is unnecessary.
3369
3370         * Animometer/tests/bouncing-particles/resources/bouncing-particles.js:
3371         * Animometer/tests/master/resources/canvas-stage.js:
3372         * Animometer/tests/master/resources/particles.js:
3373         * Animometer/tests/misc/resources/canvas-electrons.js:
3374         * Animometer/tests/misc/resources/canvas-stars.js:
3375         * Animometer/tests/resources/main.js:
3376         * Animometer/tests/simple/resources/simple-canvas.js:
3377         * Animometer/tests/simple/resources/tiled-canvas-image.js:
3378         * Animometer/tests/template/resources/template-canvas.js:
3379         * Animometer/tests/template/resources/template-css.js:
3380         * Animometer/tests/template/resources/template-svg.js:
3381         * Animometer/tests/text/resources/layering-text.js:
3382
3383 2016-02-07  Jon Lee  <jonlee@apple.com>
3384
3385         Make the fixed controller a step controller instead. Halfway through the test
3386         it will bump up the complexity 4-fold. Calculate the step timestamp using options
3387         instead of a separate parameter to the Controller constructor.
3388
3389         * Animometer/developer.html: Change value to "step"
3390         * Animometer/resources/debug-runner/animometer.js:
3391         (window.suitesManager.updateEditsElementsState): Show number inputs when set to "step".
3392         * Animometer/tests/resources/main.js:
3393         (update): Provide a hook for subclasses to tune.
3394         (StepController): Maintain a flag determining whether we've stepped, and the time
3395         we should step.
3396         (Benchmark): Use the new StepController.
3397
3398 2016-02-07  Jon Lee  <jonlee@apple.com>
3399
3400         Adjust the FPS graph scale.
3401
3402         Instead of making the FPS graph linearly scale, scale it based on the frame length,
3403         but show the data in terms of FPS. Because it is inversely proportional, and most
3404         of the data never gets below 20, concentrate the axis from 20-60 FPS, since otherwise
3405         over half of the available graph space ends up blank.
3406
3407         This means we should convert all of the FPS data to frame length data.
3408
3409         * Animometer/resources/debug-runner/graph.js: Update the domain to be based on
3410         frame length in milliseconds instead of FPS. Update the cursor to consider all of the
3411         values being shown, and then pick the min and max values to represent the length of the
3412         cursor.
3413         * Animometer/resources/runner/animometer.js:
3414         * Animometer/resources/strings.js:
3415         * Animometer/tests/resources/main.js:
3416         (processSamples): Add the ability to only sample a range of the data instead of everything
3417         after an offset index. Update sampler to record the frame lengths instead of the frame
3418         rate.
3419
3420 2016-02-07  Jon Lee  <jonlee@apple.com>
3421
3422         Add option to use different methods for retrieving a timestamp.
3423
3424         * Animometer/developer.html: Add performance.now and Date.now options.
3425         * Animometer/resources/runner/animometer.js: Default to performance.now.
3426         (window.benchmarkController.startBenchmark):
3427         * Animometer/tests/resources/main.js: Tie the desired method to _getTimestamp.
3428         (run): Use _getTimestamp.
3429         (_animateLoop): Ditto.
3430
3431 2016-02-07  Jon Lee  <jonlee@apple.com>
3432
3433         Allow adding any number of markers to the graph. The markers can be labeled
3434         and contain timestamp and sample index data. Make it a part of the controller
3435         rather than keeping it in the sampler.
3436
3437         * Animometer/resources/debug-runner/animometer.css: Add styles for markers
3438         * Animometer/resources/debug-runner/graph.js: Create the markers and add
3439         text labels.
3440         * Animometer/resources/runner/animometer.js: Assume the samplingTimeOffset
3441         is just one of the marks provided.
3442         * Animometer/resources/strings.js: Add Strings.json.marks.
3443         * Animometer/tests/resources/main.js:
3444         (Controller): Keep marks here. They are keyed by the marker name, so no two
3445         markers should have the same name.
3446         (recordFirstSample): Refactor to use mark.
3447         (mark): Allows for arbitrary data if needed later. The timestamp maintained
3448         is relative to the absolute start timestamp.
3449         (containsMark): Checks whether a mark with a specific comment exists.
3450         (processSamples): Removes the _startTimestamp offset from the marks before
3451         setting it in results.
3452         * Animometer/tests/resources/sampler.js: Remove marks.
3453
3454 2016-02-07  Jon Lee  <jonlee@apple.com>
3455
3456         Get rid of options member variable in Benchmark.
3457
3458         Options are only needed when initializing the stage or benchmark, so there's no
3459         need to also keep a reference to it.
3460
3461         * Animometer/tests/resources/main.js: Get rid of options variable in Benchmark.
3462         Pass options to Controllers and Stages.
3463         (Controller.Utilities.createClass):
3464         (Benchmark.Utilities.createClass):
3465         (get options): Deleted.
3466
3467         * Animometer/tests/bouncing-particles/resources/bouncing-canvas-images.js:
3468         * Animometer/tests/bouncing-particles/resources/bouncing-canvas-particles.js:
3469         * Animometer/tests/bouncing-particles/resources/bouncing-canvas-shapes.js:
3470         * Animometer/tests/bouncing-particles/resources/bouncing-css-images.js:
3471         * Animometer/tests/bouncing-particles/resources/bouncing-css-shapes.js:
3472         * Animometer/tests/bouncing-particles/resources/bouncing-particles.js:
3473         * Animometer/tests/bouncing-particles/resources/bouncing-svg-images.js:
3474         * Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js:
3475         * Animometer/tests/master/resources/canvas-stage.js:
3476         * Animometer/tests/master/resources/canvas-tests.js:
3477         * Animometer/tests/master/resources/particles.js:
3478         * Animometer/tests/misc/resources/canvas-electrons.js:
3479         * Animometer/tests/misc/resources/canvas-stars.js:
3480         * Animometer/tests/misc/resources/compositing-transforms.js:
3481         * Animometer/tests/simple/resources/simple-canvas-paths.js:
3482         * Animometer/tests/simple/resources/tiled-canvas-image.js:
3483         * Animometer/tests/template/resources/template-canvas.js:
3484         * Animometer/tests/template/resources/template-css.js:
3485         * Animometer/tests/template/resources/template-svg.js:
3486         * Animometer/tests/text/resources/layering-text.js:
3487
3488 2016-02-07  Jon Lee  <jonlee@apple.com>
3489
3490         Update how the benchmark is run
3491         https://bugs.webkit.org/show_bug.cgi?id=153960
3492
3493         Provisionally reviewed by Said Abou-Hallawa.
3494
3495         Introduce the notion of a Controller. It is responsible for recording, updating,
3496         and processing the statistics and complexity of the benchmark. This allows
3497         plugging in different Controllers.
3498
3499         This strips most of the functionality from Animator and BenchmarkState, so fold
3500         what's left into Benchmark. Now, Benchmarks only own a stage and a controller, but
3501         are responsible for driving the animation loop.
3502
3503         Rewrite Animator._shouldRequestAnotherFrame into two different Controllers. One
3504         maintains a fixed complexity, and the other adapts the complexity to meet a
3505         fixed FPS.
3506
3507         Fix the Kalman estimator to be modeled on a scalar variable with no model.
3508
3509         * Animometer/tests/resources/main.js: Remove BenchmarkState and Animator, and
3510         replace it with a Controller. Add a FixedController and refactor the previous controller
3511         to an AdaptiveController.
3512
3513         (Controller): Controllers own the estimator and the sampler. When a new frame is
3514         displayed, the animation loop calls update(). The estimator and sampler record
3515         stats, then tune. Samplers can track multiple series of data. The basic controller
3516         tracks timestamp, complexity, and estimated frame rate.
3517                 The Kalman estimation is based on the frame length rather than the frame
3518         rate. Because FPS is inverse proportional to frame length, in the case where the measured
3519         frame length is very small, the FPS ends up being a wildly large number (in the order of
3520         600-1000 "FPS"), and it pulls the estimator up drastically enough that it takes a while
3521         for it to settle back down. Using frame length reduces the impact of these spikes.
3522                 Converging the estimation takes enough time to avoid initializing it immediately
3523         when the benchmark starts. Instead, the benchmark runs for a brief period of time (100ms)
3524         before running it in earnest. Allow controllers an opportunity to set the complexity
3525         before starting recording.
3526                 When the benchmark is complete, the controller has an opportunity to process
3527         the samples. The default implementation calculates the raw FPS based on the time
3528         difference of the samples, and calculates the complexity score. This is moved from
3529         Benchmark.processSamples.
3530
3531         (Controller): Initialize timestamps. These are at first relative to the start of the
3532         benchmark, but are offset by the absolute start time during start(). By default maintain
3533         3 data series, but subclasses can override.
3534         (start): Calls recordFirstSample() for subclasses to override if needed.
3535         (recordFirstSample): For basic controller, start sampling at the beginning.
3536         (update): Update the frame length estimator and sample.
3537         (shouldStop): Checks that the time is before _endTimestamp.
3538         (results): Returns the processed samples.
3539         (processSamples): Iterate through the sample data and collate them. Include scores.
3540
3541         (FixedComplexityController): Controller that tunes the stage to the desired complexity
3542         prior to starting, and keeps it at that complexity.
3543
3544         (AdaptiveController): Have the estimator estimate the interval frame rate instead of the
3545         raw frame rate.
3546                 The previous version of this controller ignored the frame that came after the
3547         adjustment. The raw FPS show that whatever noise the scene change adds is negligible
3548         compared to the noise of the system overall. Stop ignoring that frame and include all
3549         frames in the measurements.
3550
3551         (Benchmark): Remove dependency on animator, and instantiate a runner based on what is
3552         selected. Most of the loop's functionality is in Controller, so remove here.
3553         (Benchmark.run): Remove start() since it is only called from run(), and fold it in here.
3554         (Benchmark._animateLoop): Fold in from Animator.animateLoop. Let the benchmark run for
3555         a brief period before calling Controller.start().
3556
3557         * Animometer/tests/resources/math.js: Fix the Kalman estimator. The filter estimates
3558         a scalar variable, and makes basic assumptions regarding the model. As a result
3559         none of the linear algebra classes are needed, so remove Matrix, Vector3, and Matrix3.
3560         (SimpleKalmanEstimator): Calculate the gain based on the provided process and
3561         measurement errors.
3562         (KalmanEstimator): Deleted.
3563         (IdentityEstimator): Deleted.
3564         (PIDController): Refactor to use the Utilities.createClass() helper.
3565
3566         The Kalman filter algorithm is explained here http://greg.czerniak.info/guides/kalman1/.
3567         The state, represented by a scalar, is the estimated frame length. There is no user
3568         transition of the state, and the state is the same as the measurement. With this model,
3569         the estimation error converges, so calculate the gain ahead of time.
3570
3571         * Animometer/developer.html: Remove fixed-after-warmup since it is not useful.
3572         Replace the option to toggle the estimator, and make it possible to customize the
3573         estimator's error parameters. Show raw FPS by default, and remove interval FPS,
3574         which will be shown instead of the filtered raw FPS.
3575         * Animometer/resources/debug-runner/animometer.css: Put the header behind the graph.
3576         Remove #intervalFPS rules; move the color to #filteredFPS.
3577         * Animometer/resources/debug-runner/graph.js:
3578         (updateGraphData): Update the hr style to force the layout to be calculated
3579         correctly. Change the tick format to be in terms of seconds, since the timestamps
3580         are in milliseconds. Remove interval data.
3581         * Animometer/resources/runner/animometer.js:
3582         (window.benchmarkController.startBenchmark): Set Kalman parameters.
3583         * Animometer/resources/runner/benchmark-runner.js:
3584         (_runBenchmarkAndRecordResults): When a benchmark completes, expect it to return
3585         the final data, rather than passing a sampler from the controller. This avoids
3586         needing to expose the sampler variable in the benchmark.
3587         * Animometer/tests/resources/sampler.js:
3588         (process): Move the setting of the target frame rate to AdaptiveController.
3589
3590 2016-02-06  Jon Lee  <jonlee@apple.com>
3591
3592         Code clean up: Move Rotater function closer to Stage static methods.
3593         The Rotater is used together with those methods; keep them close.
3594
3595         * Animometer/tests/resources/main.js:
3596
3597 2016-02-06  Jon Lee  <jonlee@apple.com>
3598
3599         Update the JS includes due to ResultsTable move.
3600
3601         * Animometer/developer.html:
3602         * Animometer/index.html:
3603
3604 2016-02-06  Jon Lee  <jonlee@apple.com>
3605
3606         Move createElement and createSVGElement to Utilities.
3607
3608         * Animometer/resources/extensions.js:
3609         (Utilities.createElement): Added.
3610         (Utilities.createSVGElement): Added.
3611         (DocumentExtension.createElement): Deleted.
3612         (DocumentExtension.createSvgElement): Deleted.
3613
3614         * Animometer/resources/debug-runner/animometer.js:
3615         * Animometer/resources/runner/animometer.js:
3616         * Animometer/tests/bouncing-particles/resources/bouncing-svg-images.js:
3617         * Animometer/tests/bouncing-particles/resources/bouncing-svg-particles.js:
3618         * Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js:
3619
3620 2016-02-06  Jon Lee  <jonlee@apple.com>
3621
3622         Add a convenience function for creating a class.
3623
3624         The pattern for creating a class is common enough to add as a Utilities
3625         helper function. It also makes it easy to collapse class definitions when
3626         editing.
3627
3628         * Animometer/resources/debug-runner/animometer.js: Move ProgressBar definition,
3629         since it is only used here.
3630         * Animometer/resources/runner/animometer.js: Move ResultsDashboard and
3631         ResultsTable definition, since it is only used here.
3632         * Animometer/resources/extensions.js: Move Utilities definition to the top. Convert
3633         Point, Insets, SimplePromise.
3634         (ProgressBar): Moved to animometer.js.
3635         (ResultsDashboard): Moved to animometer.js.
3636         (ResultsTable): Moved to animometer.js.
3637         * Animometer/resources/runner/benchmark-runner.js: Convert BenchmarkRunnerState,
3638         BenchmarkRunner.
3639         * Animometer/tests/resources/main.js: Convert Rotater, Stage, Animator, Benchmark.
3640         * Animometer/tests/resources/sampler.js: Convert Experiment, Sampler.
3641
3642         Convert test primitives.
3643         * Animometer/tests/master/resources/canvas-tests.js: Convert CanvasLineSegment,
3644         CanvasArc, CanvasLinePoint.
3645         * Animometer/tests/simple/resources/simple-canvas-paths.js: Convert CanvasLineSegment,
3646         CanvasLinePoint, CanvasQuadraticSegment, CanvasQuadraticPoint, CanvasBezierSegment,
3647         CanvasBezierPoint, CanvasArcToSegment, CanvasArcToSegmentFill, CanvasArcSegment,
3648         CanvasArcSegmentFill, CanvasRect, CanvasRectFill.
3649         * Animometer/tests/simple/resources/tiled-canvas-image.js: Convert CanvasImageTile.
3650
3651 2016-02-06  Jon Lee  <jonlee@apple.com>
3652
3653         Minor improvements to debug harness.
3654
3655         * Animometer/developer.html:
3656         * Animometer/resources/debug-runner/animometer.css:
3657         (#suites): Put the complexity text boxes closer to the test names.
3658         (#options):
3659         (#rawFPS circle): Make the interval FPS appear as a separate data series, with a line.
3660         (#intervalFPS path):
3661         (#intervalFPS circle):
3662         * Animometer/resources/debug-runner/animometer.js:
3663         (window.optionsManager.updateLocalStorageFromUI): Convert number inputs from text.
3664         (window.suitesManager._onChangeTestCheckbox): Refactor to take a checkbox.
3665         (window.suitesManager._createTestElement): Enhance such that typing into the complexity
3666         input will automatically select that test for running.
3667         (window.suitesManager.updateLocalStorageFromJSON): Make the harness work for private
3668         browsing.
3669         * Animometer/resources/debug-runner/graph.js: Separate the intervalFPS data, and show
3670         more accuracy in timestamps.
3671
3672 2016-02-06  Jon Lee  <jonlee@apple.com>
3673
3674         Refactor helper methods for getting random values for a stage.
3675
3676         Instead of requiring a Stage instance, just attach it to the Stage object.
3677
3678         * Animometer/tests/bouncing-particles/resources/bouncing-canvas-shapes.js:
3679         * Animometer/tests/bouncing-particles/resources/bouncing-css-shapes.js:
3680         * Animometer/tests/bouncing-particles/resources/bouncing-particles.js:
3681         * Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js:
3682         * Animometer/tests/master/resources/canvas-tests.js:
3683         * Animometer/tests/master/resources/particles.js:
3684         * Animometer/tests/misc/resources/canvas-electrons.js:
3685         * Animometer/tests/misc/resources/canvas-stars.js:
3686         * Animometer/tests/misc/resources/compositing-transforms.js:
3687         * Animometer/tests/resources/main.js:
3688         * Animometer/tests/simple/resources/simple-canvas-paths.js:
3689
3690 2016-02-05  Said Abou-Hallawa  <sabouhallawa@apple.com>
3691
3692         Add a new graphics test for CanvasRenderingContext2D functions: getImageData and putImageData
3693         https://bugs.webkit.org/show_bug.cgi?id=151716
3694
3695         Reviewed by Darin Adler.
3696
3697         The purpose of this test is to measure the performance of getImageData
3698         and putImageData functions. This test draws a background on the canvas
3699         and then gets some random tiles from this background and draw them in
3700         destinations different from their original sources.
3701
3702         * Animometer/resources/debug-runner/tests.js: Adding the new test to the canvas simple tests suite.
3703
3704         * Animometer/resources/extensions.js:
3705         (Array.prototype.shuffle): Shuffles the elements of an array.
3706
3707         (Point.zero): Returns a new Point object whose x and y are equal zero.
3708         (Point.prototype.str): Used for debugging the Point object.
3709
3710         * Animometer/tests/simple/resources/tiled-canvas-image.js: Added.
3711         (CanvasImageTile):
3712         (CanvasImageTile.prototype.getImageData):
3713         (CanvasImageTile.prototype.putImageData):
3714         (Stage.call.initialize):
3715         (Stage.call._createTiles):
3716         (Stage.call._nextTilePosition):
3717         (Stage.call.tune):
3718         (Stage.call._drawBackground):
3719         (Stage.call.animate):
3720         (Stage.call.complexity):
3721         (Stage.call):
3722         * Animometer/tests/simple/tiled-canvas-image.html: Added.
3723
3724 2016-01-07  Jon Lee  <jonlee@apple.com>
3725
3726         Fix new test.
3727
3728         * Animometer/resources/runner/tests.js: Wrong URL from an
3729         old patch.
3730         * Animometer/tests/master/particles.html:
3731         * Animometer/tests/master/resources/dom-particles.js:
3732         (Particle.call.reset): Figured out a simpler way to set up
3733         the particles.
3734         (this.move.reset): Deleted.
3735         (this.move._applyAttributes): Deleted.
3736         * Animometer/tests/master/resources/particles.js:
3737         (Particle): Call move() after reset().
3738
3739 2016-01-07  Jon Lee  <jonlee@apple.com>
3740
3741         Update benchmark test suite
3742         https://bugs.webkit.org/show_bug.cgi?id=152679
3743
3744         Reviewed by Simon Fraser.
3745
3746         Add a new test. The test has a rotating background
3747         gradient, and does a better job physically simulating
3748         particles.
3749
3750         * Animometer/resources/extensions.js: Teach Point to take constants as well as other Points.
3751         (Point.prototype.length): Added.
3752         (Point.prototype.normalize): Added.
3753         * Animometer/resources/runner/tests.js: Add the test to the master suite.
3754         * Animometer/tests/master/particles.html: Added.
3755         * Animometer/tests/master/resources/particles.js: Added. Parent class for different kinds of particles.
3756         (Particle):
3757         (Particle.prototype.reset): If the particle starts slowing down in terms of its animation,
3758         reset it.
3759         (Particle.prototype.animate): Bounce off the walls elastically, and include gravity.
3760         (Particle.prototype.move): Subclasses should override.
3761         (ParticlesStage): Stage includes a rotating gradient background.
3762         * Animometer/tests/master/resources/dom-particles.js: Added. Creates a <div> and adds it to
3763         the stage.
3764         * Animometer/tests/resources/star.svg: Added.
3765
3766 2016-01-03  Jon Lee  <jonlee@apple.com>
3767
3768         Update benchmark test suite
3769         https://bugs.webkit.org/show_bug.cgi?id=152679
3770
3771         Reviewed by Simon Fraser.
3772
3773         Move algorithm.js and sampler.js to tests/ and benchmark-runner.js to runner/.
3774
3775         Needed by both harnesses.
3776         * Animometer/resources/runner/benchmark-runner.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/benchmark-runner.js.
3777         * Animometer/developer.html:
3778         * Animometer/index.html:
3779
3780         Needed only by the tests. Move to tests/. Statistics, in sampler.js, is used by ResultsDashboard, so move that
3781         into extensions.js.
3782         * Animometer/resources/extensions.js:
3783         * Animometer/tests/resources/algorithm.js: Renamed from PerformanceTests/Animometer/resources/algorithm.js.
3784         * Animometer/tests/resources/sampler.js: Renamed from PerformanceTests/Animometer/resources/sampler.js.
3785         * Animometer/tests/bouncing-particles/bouncing-canvas-images.html:
3786         * Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html:
3787         * Animometer/tests/bouncing-particles/bouncing-css-images.html:
3788         * Animometer/tests/bouncing-particles/bouncing-css-shapes.html:
3789         * Animometer/tests/bouncing-particles/bouncing-svg-images.html:
3790         * Animometer/tests/bouncing-particles/bouncing-svg-shapes.html:
3791         * Animometer/tests/master/canvas-stage.html:
3792         * Animometer/tests/misc/canvas-electrons.html:
3793         * Animometer/tests/misc/canvas-stars.html:
3794         * Animometer/tests/misc/compositing-transforms.html:
3795         * Animometer/tests/simple/simple-canvas-paths.html:
3796         * Animometer/tests/template/template-canvas.html:
3797         * Animometer/tests/template/template-css.html:
3798         * Animometer/tests/template/template-svg.html:
3799         * Animometer/tests/text/layering-text.html:
3800         * Animometer/tests/text/text-boxes.html:
3801
3802 2016-01-07  Jon Lee  <jonlee@apple.com>
3803
3804         Update benchmark test suite
3805         https://bugs.webkit.org/show_bug.cgi?id=152679
3806
3807         Reviewed by Simon Fraser.
3808
3809         Fix tests for other browsers.
3810
3811         * Animometer/resources/extensions.js:
3812         (Point.elementClientSize): Some browsers return 0 for SVG clientWidth and clientHeight.
3813         Use getBoundingClientRect() instead.
3814         * Animometer/tests/misc/resources/canvas-electrons.js:
3815         (CanvasElectron.prototype._draw): Some browsers don't support ellipse.
3816
3817 2016-01-07  Jon Lee  <jonlee@apple.com>
3818
3819         Add a waitUntilReady() step
3820         https://bugs.webkit.org/show_bug.cgi?id=152862
3821
3822         Reviewed by Simon Fraser.
3823
3824         Add a waitUntilReady() callback that lets the benchmark complete