Add framerate baselines to the splay-detail test
[WebKit-https.git] / PerformanceTests / ChangeLog
1 2017-01-16  Filip Pizlo  <fpizlo@apple.com>
2
3         Add framerate baselines to the splay-detail test
4         https://bugs.webkit.org/show_bug.cgi?id=167659
5
6         Reviewed by Geoffrey Garen.
7
8         * JetStream/Octane2/splay-detail.html:
9
10 2017-01-31  Carlos Alberto Lopez Perez  <clopez@igalia.com>
11
12         [EFL][GTK] Skip some IndexedDB tests that time out on the Perf bots.
13
14         Unreviewed performance test gardening.
15
16         * Skipped:
17
18 2017-01-30  Zalan Bujtas  <zalan@apple.com>
19
20         Simple line layout: Small tweaks to improve performance.
21         https://bugs.webkit.org/show_bug.cgi?id=167611
22         <rdar://problem/30274294>
23
24         Reviewed by Simon Fraser.
25
26         * Layout/simple-line-layout-non-repeating-text.html: Added.
27
28 2017-01-13  Said Abou-Hallawa  <sabouhallawa@apple.com>
29
30         26 MotionMark performance tests failing
31         https://bugs.webkit.org/show_bug.cgi?id=166854
32
33         Reviewed by Ryosuke Niwa.
34         
35         Skip running MotionMark on WebKit perf bots.
36
37         * Skipped:
38
39 2017-01-06  Said Abou-Hallawa  <sabouhallawa@apple.com>
40
41         Rename the directory and the files of MotionMark from Animometer to MotionMark
42         https://bugs.webkit.org/show_bug.cgi?id=166659
43
44         Reviewed by Jon Lee.
45         
46         Rename the directory and the files of the benchmark to its new name.
47
48         * MotionMark/developer.html: Renamed from PerformanceTests/Animometer/developer.html.
49         * MotionMark/index.html: Renamed from PerformanceTests/Animometer/index.html.
50         * MotionMark/resources/debug-runner/d3.min.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/d3.min.js.
51         * MotionMark/resources/debug-runner/graph.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/graph.js.
52         * MotionMark/resources/debug-runner/motionmark.css: Renamed from PerformanceTests/Animometer/resources/debug-runner/animometer.css.
53         * MotionMark/resources/debug-runner/motionmark.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/animometer.js.
54         * MotionMark/resources/debug-runner/tests.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/tests.js.
55         * MotionMark/resources/extensions.js: Renamed from PerformanceTests/Animometer/resources/extensions.js.
56         * MotionMark/resources/runner/benchmark-runner.js: Renamed from PerformanceTests/Animometer/resources/runner/benchmark-runner.js.
57         * MotionMark/resources/runner/crystal.svg: Renamed from PerformanceTests/Animometer/resources/runner/crystal.svg.
58         * MotionMark/resources/runner/lines.svg: Renamed from PerformanceTests/Animometer/resources/runner/lines.svg.
59         * MotionMark/resources/runner/logo.svg: Renamed from PerformanceTests/Animometer/resources/runner/logo.svg.
60         * MotionMark/resources/runner/motionmark.css: Renamed from PerformanceTests/Animometer/resources/runner/animometer.css.
61         * MotionMark/resources/runner/motionmark.js: Renamed from PerformanceTests/Animometer/resources/runner/animometer.js.
62         * MotionMark/resources/runner/tests.js: Renamed from PerformanceTests/Animometer/resources/runner/tests.js.
63         * MotionMark/resources/statistics.js: Renamed from PerformanceTests/Animometer/resources/statistics.js.
64         * MotionMark/resources/strings.js: Renamed from PerformanceTests/Animometer/resources/strings.js.
65         * MotionMark/tests/3d/resources/webgl.js: Renamed from PerformanceTests/Animometer/tests/3d/resources/webgl.js.
66         * MotionMark/tests/3d/webgl.html: Renamed from PerformanceTests/Animometer/tests/3d/webgl.html.
67         * MotionMark/tests/bouncing-particles/bouncing-canvas-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-canvas-images.html.
68         * MotionMark/tests/bouncing-particles/bouncing-canvas-shapes.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html.
69         * MotionMark/tests/bouncing-particles/bouncing-css-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-css-images.html.
70         * MotionMark/tests/bouncing-particles/bouncing-css-shapes.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-css-shapes.html.
71         * MotionMark/tests/bouncing-particles/bouncing-svg-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-svg-images.html.
72         * MotionMark/tests/bouncing-particles/bouncing-svg-shapes.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-svg-shapes.html.
73         * MotionMark/tests/bouncing-particles/bouncing-tagged-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-tagged-images.html.
74         * MotionMark/tests/bouncing-particles/resources/bouncing-canvas-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-canvas-images.js.
75         * MotionMark/tests/bouncing-particles/resources/bouncing-canvas-particles.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-canvas-particles.js.
76         * MotionMark/tests/bouncing-particles/resources/bouncing-canvas-shapes.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-canvas-shapes.js.
77         * MotionMark/tests/bouncing-particles/resources/bouncing-css-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-css-images.js.
78         * MotionMark/tests/bouncing-particles/resources/bouncing-css-shapes.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-css-shapes.js.
79         * MotionMark/tests/bouncing-particles/resources/bouncing-svg-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-svg-images.js.
80         * MotionMark/tests/bouncing-particles/resources/bouncing-svg-particles.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-svg-particles.js.
81         * MotionMark/tests/bouncing-particles/resources/bouncing-svg-shapes.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js.
82         * MotionMark/tests/bouncing-particles/resources/bouncing-tagged-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-tagged-images.js.
83         * MotionMark/tests/bouncing-particles/resources/image1.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image1.jpg.
84         * MotionMark/tests/bouncing-particles/resources/image2.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image2.jpg.
85         * MotionMark/tests/bouncing-particles/resources/image3.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image3.jpg.
86         * MotionMark/tests/bouncing-particles/resources/image4.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image4.jpg.
87         * MotionMark/tests/bouncing-particles/resources/image5.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image5.jpg.
88         * MotionMark/tests/dom/compositing-transforms.html: Renamed from PerformanceTests/Animometer/tests/dom/compositing-transforms.html.
89         * MotionMark/tests/dom/focus.html: Renamed from PerformanceTests/Animometer/tests/dom/focus.html.
90         * MotionMark/tests/dom/leaves.html: Renamed from PerformanceTests/Animometer/tests/dom/leaves.html.
91         * MotionMark/tests/dom/particles.html: Renamed from PerformanceTests/Animometer/tests/dom/particles.html.
92         * MotionMark/tests/dom/resources/compositing-transforms.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/compositing-transforms.js.
93         * MotionMark/tests/dom/resources/dom-particles.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/dom-particles.js.
94         * MotionMark/tests/dom/resources/focus.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/focus.js.
95         * MotionMark/tests/dom/resources/leaves.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/leaves.js.
96         * MotionMark/tests/master/canvas-stage.html: Renamed from PerformanceTests/Animometer/tests/master/canvas-stage.html.
97         * MotionMark/tests/master/focus.html: Renamed from PerformanceTests/Animometer/tests/master/focus.html.
98         * MotionMark/tests/master/image-data.html: Renamed from PerformanceTests/Animometer/tests/master/image-data.html.
99         * MotionMark/tests/master/leaves.html: Renamed from PerformanceTests/Animometer/tests/master/leaves.html.
100         * MotionMark/tests/master/multiply.html: Renamed from PerformanceTests/Animometer/tests/master/multiply.html.
101         * MotionMark/tests/master/resources/canvas-stage.js: Renamed from PerformanceTests/Animometer/tests/master/resources/canvas-stage.js.
102         * MotionMark/tests/master/resources/canvas-tests.js: Renamed from PerformanceTests/Animometer/tests/master/resources/canvas-tests.js.
103         * MotionMark/tests/master/resources/compass.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/compass.svg.
104         * MotionMark/tests/master/resources/compass100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/compass100.png.
105         * MotionMark/tests/master/resources/console.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/console.svg.
106         * MotionMark/tests/master/resources/console100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/console100.png.
107         * MotionMark/tests/master/resources/contribute.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/contribute.svg.
108         * MotionMark/tests/master/resources/contribute100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/contribute100.png.
109         * MotionMark/tests/master/resources/debugger.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/debugger.svg.
110         * MotionMark/tests/master/resources/debugger100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/debugger100.png.
111         * MotionMark/tests/master/resources/focus.js: Renamed from PerformanceTests/Animometer/tests/master/resources/focus.js.
112         * MotionMark/tests/master/resources/image-data.js: Renamed from PerformanceTests/Animometer/tests/master/resources/image-data.js.
113         * MotionMark/tests/master/resources/inspector.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/inspector.svg.
114         * MotionMark/tests/master/resources/inspector100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/inspector100.png.
115         * MotionMark/tests/master/resources/layout.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/layout.svg.
116         * MotionMark/tests/master/resources/layout100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/layout100.png.
117         * MotionMark/tests/master/resources/leaves.js: Renamed from PerformanceTests/Animometer/tests/master/resources/leaves.js.
118         * MotionMark/tests/master/resources/multiply.js: Renamed from PerformanceTests/Animometer/tests/master/resources/multiply.js.
119         * MotionMark/tests/master/resources/particles.js: Renamed from PerformanceTests/Animometer/tests/master/resources/particles.js.
120         * MotionMark/tests/master/resources/performance.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/performance.svg.
121         * MotionMark/tests/master/resources/performance100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/performance100.png.
122         * MotionMark/tests/master/resources/script.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/script.svg.
123         * MotionMark/tests/master/resources/script100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/script100.png.
124         * MotionMark/tests/master/resources/shortcuts.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/shortcuts.svg.
125         * MotionMark/tests/master/resources/shortcuts100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/shortcuts100.png.
126         * MotionMark/tests/master/resources/standards.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/standards.svg.
127         * MotionMark/tests/master/resources/standards100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/standards100.png.
128         * MotionMark/tests/master/resources/storage.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/storage.svg.
129         * MotionMark/tests/master/resources/storage100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/storage100.png.
130         * MotionMark/tests/master/resources/styles.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/styles.svg.
131         * MotionMark/tests/master/resources/styles100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/styles100.png.
132         * MotionMark/tests/master/resources/svg-particles.js: Renamed from PerformanceTests/Animometer/tests/master/resources/svg-particles.js.
133         * MotionMark/tests/master/resources/text.js: Renamed from PerformanceTests/Animometer/tests/master/resources/text.js.
134         * MotionMark/tests/master/resources/timeline.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/timeline.svg.
135         * MotionMark/tests/master/resources/timeline100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/timeline100.png.
136         * MotionMark/tests/master/svg-particles.html: Renamed from PerformanceTests/Animometer/tests/master/svg-particles.html.
137         * MotionMark/tests/master/text.html: Renamed from PerformanceTests/Animometer/tests/master/text.html.
138         * MotionMark/tests/resources/main.js: Renamed from PerformanceTests/Animometer/tests/resources/main.js.
139         * MotionMark/tests/resources/math.js: Renamed from PerformanceTests/Animometer/tests/resources/math.js.
140         * MotionMark/tests/resources/stage.css: Renamed from PerformanceTests/Animometer/tests/resources/stage.css.
141         * MotionMark/tests/resources/star.svg: Renamed from PerformanceTests/Animometer/tests/resources/star.svg.
142         * MotionMark/tests/resources/yin-yang.png: Renamed from PerformanceTests/Animometer/tests/resources/yin-yang.png.
143         * MotionMark/tests/resources/yin-yang.svg: Renamed from PerformanceTests/Animometer/tests/resources/yin-yang.svg.
144         * MotionMark/tests/simple/resources/simple-canvas-paths.js: Renamed from PerformanceTests/Animometer/tests/simple/resources/simple-canvas-paths.js.
145         * MotionMark/tests/simple/resources/simple-canvas.js: Renamed from PerformanceTests/Animometer/tests/simple/resources/simple-canvas.js.
146         * MotionMark/tests/simple/resources/tiled-canvas-image.js: Renamed from PerformanceTests/Animometer/tests/simple/resources/tiled-canvas-image.js.
147         * MotionMark/tests/simple/simple-canvas-paths.html: Renamed from PerformanceTests/Animometer/tests/simple/simple-canvas-paths.html.
148         * MotionMark/tests/simple/tiled-canvas-image.html: Renamed from PerformanceTests/Animometer/tests/simple/tiled-canvas-image.html.
149         * MotionMark/tests/template/resources/template-canvas.js: Renamed from PerformanceTests/Animometer/tests/template/resources/template-canvas.js.
150         * MotionMark/tests/template/resources/template-css.js: Renamed from PerformanceTests/Animometer/tests/template/resources/template-css.js.
151         * MotionMark/tests/template/resources/template-svg.js: Renamed from PerformanceTests/Animometer/tests/template/resources/template-svg.js.
152         * MotionMark/tests/template/template-canvas.html: Renamed from PerformanceTests/Animometer/tests/template/template-canvas.html.
153         * MotionMark/tests/template/template-css.html: Renamed from PerformanceTests/Animometer/tests/template/template-css.html.
154         * MotionMark/tests/template/template-svg.html: Renamed from PerformanceTests/Animometer/tests/template/template-svg.html.
155
156 2016-12-13  Brady Eidson  <beidson@apple.com>
157
158         Add a PerfTest targeting IDBObjectStore.get().
159         https://bugs.webkit.org/show_bug.cgi?id=165816
160
161         Reviewed by Alex Christensen.
162
163         * IndexedDB/objectstore-get.html: Added.
164
165 2016-12-13  Brady Eidson  <beidson@apple.com>
166
167         Add a PerfTest targeting IDBIndex.get().
168         https://bugs.webkit.org/show_bug.cgi?id=165803
169
170         Reviewed by Alex Christensen.
171
172         * IndexedDB/index-get.html: Added.
173
174 2016-12-12  Brady Eidson  <beidson@apple.com>
175
176         More IndexedDB perf tests.
177         https://bugs.webkit.org/show_bug.cgi?id=165634
178
179         Reviewed by Sam Weinig.
180
181         * IndexedDB/index-multientry.html:
182         * IndexedDB/large-number-of-inserts-responsiveness.html: Added.
183         * IndexedDB/large-number-of-inserts.html: Added.
184         * IndexedDB/objectstore-cursor.html: Added. Runtime cut in 1/4th compared to r209672.
185
186         * resources/runner.js:
187
188 2016-12-12  Commit Queue  <commit-queue@webkit.org>
189
190         Unreviewed, rolling out r209672.
191         https://bugs.webkit.org/show_bug.cgi?id=165766
192
193         IndexedDB/objectstore-cursor.html perf test always times out
194         (Requested by ap on #webkit).
195
196         Reverted changeset:
197
198         "More IndexedDB perf tests."
199         https://bugs.webkit.org/show_bug.cgi?id=165634
200         http://trac.webkit.org/changeset/209672
201
202 2016-12-10  Brady Eidson  <beidson@apple.com>
203
204         More IndexedDB perf tests.
205         https://bugs.webkit.org/show_bug.cgi?id=165634
206
207         Reviewed by Sam Weinig.
208
209         * IndexedDB/index-multientry.html:
210         * IndexedDB/large-number-of-inserts-responsiveness.html: Added.
211         * IndexedDB/large-number-of-inserts.html: Added.
212         * IndexedDB/objectstore-cursor.html: Added.
213         
214         * resources/runner.js: Add "track responsiveness" functionality to PerfTestRunner.
215
216 2016-12-06  Filip Pizlo  <fpizlo@apple.com>
217
218         Concurrent GC should be stable enough to land enabled
219         https://bugs.webkit.org/show_bug.cgi?id=164990
220
221         Reviewed by Geoffrey Garen.
222         
223         Made CDjs more configurable and refined the "large.js" configuration. I was using that one and
224         the new "long.js" configuration to tune concurrent eden GCs.
225         
226         Added a new way of running Splay in browser, which using chartjs to plot the execution times of
227         2000 iterations. This includes the minified chartjs.
228
229         * JetStream/Octane2/splay-detail.html: Added.
230         * JetStream/cdjs/benchmark.js:
231         (benchmarkImpl):
232         (benchmark):
233         * JetStream/cdjs/long.js: Added.
234
235 2016-12-07  Brady Eidson  <beidson@apple.com>
236
237         Add IDB perf tests stressing key size.
238         https://bugs.webkit.org/show_bug.cgi?id=165567
239
240         Reviewed by Alex Christensen.
241
242         * IndexedDB/large-array-keys.html: Added.
243         * IndexedDB/large-binary-keys.html: Added.
244         * IndexedDB/large-string-keys.html: Added.
245
246 2016-12-05  Brady Eidson  <beidson@apple.com>
247
248         Add an IndexedDB perf test to PerformanceTests.
249         https://bugs.webkit.org/show_bug.cgi?id=165430
250
251         Reviewed by Alex Christensen.
252
253         * IndexedDB/index-multientry.html: Added.
254
255 2016-12-02  Filip Pizlo  <fpizlo@apple.com>
256
257         ES6SampleBench should report an average for Steady State so that all of the numbers are comparable
258         https://bugs.webkit.org/show_bug.cgi?id=165325
259
260         Reviewed by Saam Barati.
261
262         This makes all of the numbers that ES6SampleBench reports comparable to each other: they all speak of the time
263         it took to run an iteration of something.
264
265         * ES6SampleBench/results.js:
266         (Results.prototype.reportResult):
267
268 2016-11-18  Jer Noble  <jer.noble@apple.com>
269
270         Add some performance tests for various aspects of media loading and playback.
271         https://bugs.webkit.org/show_bug.cgi?id=164977
272
273         Reviewed by Darin Adler.
274
275         Add new tests for a few aspects of media playback; namely: how quickly  media elements fire
276         the "canplaythrough" event when all data is immediately available, how quickly HLS playback
277         switches resolutions, and at what playback rate MSE-backed video can play without dropping
278         frames. Skip these tests by default, as they're not runnable on all ports, and they require
279         a webserver (run-webkit-httpd) to complete on macOS.
280
281         * Media/HLSCanPlayThrough.html: Added.
282         * Media/HLSGearChange.html: Added.
283         * Media/MSECanPlayThrough.html: Added.
284         * Media/MSEPlaybackRate.html: Added.
285         * Media/StandardCanPlayThrough.html: Added.
286         * Media/hls/1080p/iframe_index.m3u8: Added.
287         * Media/hls/1080p/prog_index.m3u8: Added.
288         * Media/hls/1080p/test.ts: Added.
289         * Media/hls/480p/iframe_index.m3u8: Added.
290         * Media/hls/480p/prog_index.m3u8: Added.
291         * Media/hls/480p/test.ts: Added.
292         * Media/hls/720p/iframe_index.m3u8: Added.
293         * Media/hls/720p/prog_index.m3u8: Added.
294         * Media/hls/720p/test.ts: Added.
295         * Media/hls/720p/test.ts.back: Added.
296         * Media/hls/index.m3u8: Added.
297         * Media/media-source-loader.js: Added.
298         (MediaSourceLoader):
299         (MediaSourceLoader.prototype.loadManifest):
300         (MediaSourceLoader.prototype.loadManifestSucceeded):
301         (MediaSourceLoader.prototype.loadManifestFailed):
302         (MediaSourceLoader.prototype.loadMediaData):
303         (MediaSourceLoader.prototype.loadMediaDataSucceeded):
304         (MediaSourceLoader.prototype.loadMediaDataFailed):
305         (MediaSourceLoader.prototype.get type):
306         (MediaSourceLoader.prototype.get duration):
307         (MediaSourceLoader.prototype.get initSegment):
308         (MediaSourceLoader.prototype.get mediaSegmentsLength):
309         (MediaSourceLoader.prototype.mediaSegments):
310         (MediaSourceLoader.prototype.get everyMediaSegment):
311         * Media/test-fragmented-video.json: Added.
312         * Media/test-fragmented-video.mp4: Added.
313         * Media/test.mp4: Added.
314         * Skipped:
315
316 2016-11-18  Filip Pizlo  <fpizlo@apple.com>
317
318         Concurrent GC should be able to run splay in debug mode and earley/raytrace in release mode with no perf regression
319         https://bugs.webkit.org/show_bug.cgi?id=164282
320
321         Reviewed by Geoffrey Garen and Oliver Hunt.
322         
323         CDjs is a fun benchmark for stressing concurrent GCs, but to really give the GC a good
324         workout you need to increase the amount of work that the test does. This adds a second
325         configuration of the benchmark that has more aircraft. It uses much more memory and causes us
326         to do more GCs and those GCs take longer.
327
328         * JetStream/cdjs/benchmark.js:
329         (benchmarkImpl):
330         (benchmark):
331         * JetStream/cdjs/large.js: Added.
332
333 2016-11-14  Filip Pizlo  <fpizlo@apple.com>
334
335         Unreviewed, revert unintended change.
336
337         * ES6SampleBench/Air/benchmark.js:
338
339 2016-10-03  Saam Barati  <sbarati@apple.com>
340
341         MapHash should speculate on the type of its child node
342         https://bugs.webkit.org/show_bug.cgi?id=161922
343
344         Reviewed by Filip Pizlo.
345
346         I gave the main generator a name so it's easier to see what
347         it is when using the sampling profiler.
348
349         * ES6SampleBench/Basic/ast.js:
350         (Basic.Program):
351
352 2016-09-19  Sergio Villar Senin  <svillar@igalia.com>
353
354         [css-grid] Remove the x2 computation of row sizes with indefinite heights
355         https://bugs.webkit.org/show_bug.cgi?id=162150
356
357         Reviewed by Darin Adler.
358
359         Added a new test case which checks the layout performance of grids inside other grids, i.e,
360         grids acting both as grid container and grid item.
361
362         * Layout/nested-grid.html: Added.
363
364 2016-09-09  Simon Fraser  <simon.fraser@apple.com>
365
366         Perf test Animation/css-accelerated-animation.html failing
367         https://bugs.webkit.org/show_bug.cgi?id=161795
368
369         Skip Animation/css-accelerated-animation.html since this only works as a MobileSafari
370         perf test.
371
372         * Skipped:
373
374 2016-09-08  Simon Fraser  <simon.fraser@apple.com>
375
376         Add a content-animation test with accelerated CSS animations
377         https://bugs.webkit.org/show_bug.cgi?id=161776
378
379         Reviewed by Dean Jackson.
380
381         Add a test that measures the frame rate of accelerated CSS animations.
382
383         This is like css-animation.html (which animates 'left' and 'top') but animates
384         transforms instead. In order to get animation in X and Y we make two nested elements
385         and set transformX() on one, and transformY() on the other.
386
387         * Animation/css-accelerated-animation.html: Added.
388
389 2016-08-23  Saam Barati  <sbarati@apple.com>
390
391         It should be easy to run ES6SampleBench from the jsc shell
392         https://bugs.webkit.org/show_bug.cgi?id=161085
393
394         Reviewed by Yusuke Suzuki.
395
396         This patch makes ES6 sample bench runnable through the `jsc` shell.
397         To do that, you need to be in the PerformanceTests/ES6SampleBench
398         directory and run `jsc cli.js`. To make this work, the benchmark
399         is now aware of being run in two modes: via the browser, and via
400         the shell. Each entry point will set a variable `isInBrowser`,
401         and the benchmark will do different things based on if that
402         variable is true or false.
403
404         * ES6SampleBench/Air/benchmark.js:
405         * ES6SampleBench/Air/stress-test.js:
406         * ES6SampleBench/Basic/benchmark.js:
407         (runBenchmark):
408         * ES6SampleBench/Basic/stress-test.js:
409         * ES6SampleBench/air_benchmark.js:
410         * ES6SampleBench/basic_benchmark.js:
411         * ES6SampleBench/cli.js: Added.
412         (return.doRun):
413         (makeBenchmarkRunner):
414         * ES6SampleBench/driver.js:
415         (Driver):
416         (Driver.prototype._recomputeSummary):
417         (Driver.prototype._iterate.window.setTimeout):
418         (Driver.prototype._iterate):
419         (Driver.prototype._updateIterations):
420         * ES6SampleBench/glue.js:
421         (reportResult):
422         * ES6SampleBench/index.html:
423         * ES6SampleBench/results.js:
424         (Results):
425         (Results.prototype.reportRunning):
426         (Results.prototype.reportDone):
427         (Results.prototype.reportError):
428         * ES6SampleBench/stats.js:
429         (Stats):
430         (Stats.prototype.toString):
431         (Stats.prototype._update):
432
433 2016-08-01  Said Abou-Hallawa  <sabouhallawa@apple.com>
434
435         Add an option to run the MotionMark tests with classic tile size (512x512)
436         https://bugs.webkit.org/show_bug.cgi?id=160371
437
438         Reviewed by Darin Adler.
439
440         Setting the body to large size (3000x3000) and overflow to scroll forces
441         the classic tile size (512x512). This option does not affect the canvas
442         tests because the canvas is displayed on one tile regardless of its size.
443
444         * Animometer/developer.html:
445         * Animometer/resources/debug-runner/animometer.css:
446         (body.showing-test-container.tiles-big):
447         (body.showing-test-container.tiles-classic):
448         * Animometer/resources/debug-runner/animometer.js:
449         * Animometer/resources/runner/animometer.js:
450         (window.benchmarkController.startBenchmark):
451
452 2016-06-29  Filip Pizlo  <fpizlo@apple.com>
453
454         Generators violate bytecode liveness validation
455         https://bugs.webkit.org/show_bug.cgi?id=159279
456
457         Reviewed by Yusuke Suzuki.
458
459         Add Basic to our test harness.
460
461         Also made some cosmetic changes to the benchmark harness.
462
463         * ES6SampleBench/Basic/basic-tests.yaml: Added.
464         * ES6SampleBench/Basic/stress-test.js: Added.
465         (preciseTime):
466         * ES6SampleBench/driver.js:
467         (Driver):
468         (Driver.prototype.start):
469         (Driver.prototype.reportError):
470         * ES6SampleBench/glue.js:
471         * ES6SampleBench/index.html:
472
473 2016-06-28  Filip Pizlo  <fpizlo@apple.com>
474
475         ES6SampleBench should have a harness
476         https://bugs.webkit.org/show_bug.cgi?id=159246
477
478         Reviewed by Saam Barati.
479         
480         This adds a simple web harness for ES6SampleBench. It runs Air and Basic 10 times for 200
481         iterations each and reports three metrics:
482         
483         First iteration: the time it takes for the first iteration to run. This is the first
484         iteration after the benchmark is loaded into the iframe, so it's representative of what
485         would happen if one of these workloads only ran for a short time.
486         
487         Worst 2%: of the last 199 iterations, the average of the worst 2% iterations. If code like
488         any of these workloads was used in an important event handler, you'd want that code to run
489         well in the worst case in addition to having great throughput.
490         
491         Steady state: the total of the last 199 iterations. This is representative of what would
492         happen if you ran code like this for a long time.
493         
494         The total score is the geomean of the firstIteration/worstCase/steadyState numbers of the
495         two benchmarks.
496         
497         The harness does statistics using Student's T-distribution confidence intervals.
498
499         * ES6SampleBench/Basic/benchmark.js:
500         (Benchmark):
501         * ES6SampleBench/air_benchmark.js: Added.
502         * ES6SampleBench/basic_benchmark.js: Added.
503         * ES6SampleBench/driver.js: Added.
504         (Driver):
505         (Driver.prototype.addBenchmark):
506         (Driver.prototype.start):
507         (Driver.prototype.reportResult):
508         (Driver.prototype.reportError):
509         (Driver.prototype._recomputeSummary.Geomean):
510         (Driver.prototype._recomputeSummary.Geomean.prototype.add):
511         (Driver.prototype._recomputeSummary.Geomean.prototype.get result):
512         (Driver.prototype._recomputeSummary):
513         (Driver.prototype._iterate):
514         (Driver.prototype._updateIterations):
515         * ES6SampleBench/glue.js: Added.
516         * ES6SampleBench/index.html: Added.
517         * ES6SampleBench/results.js: Added.
518         (Results):
519         (Results.prototype.get benchmark):
520         (Results.prototype.reset):
521         (Results.prototype.reportRunning):
522         (Results.prototype.reportDone):
523         (Results.prototype.reportResult.averageAbovePercentile):
524         (Results.prototype.reportResult):
525         (Results.prototype.reportError):
526         * ES6SampleBench/stats.js: Added.
527         (Stats):
528         (Stats.prototype.reset):
529         (Stats.prototype.add):
530         (Stats.prototype.get numIterations):
531         (Stats.prototype.valueForIteration):
532         (Stats.get result.tDist):
533         (Stats.prototype.get result):
534         (Stats.prototype.toString):
535         (Stats.prototype._update):
536         * ES6SampleBench/style.css: Added.
537         (body):
538         (body, th, tr):
539         (h1, h2, h3, h4):
540         (h1):
541         (h2):
542         (h3):
543         (hr):
544         (address):
545         (img):
546         (.underline):
547         (ol.loweralpha):
548         (ol.upperalpha):
549         (ol.lowerroman):
550         (ol.upperroman):
551         (ol.arabic):
552         (.banner-link:link, .banner-link:visited):
553         (:link, :visited):
554         (h4 :link, h4 :visited, h5 :link, h5 :visited, h6 :link, h6 :visited):
555         (.anchor:link, .anchor:visited):
556         (* > .anchor:link, * > .anchor:visited):
557         (span:hover .anchor):
558         (a.forbidden, span.forbidden):
559         (a.missing:hover):
560         (a.closed:link, a.closed:visited, span.closed):
561         (pre):
562         (div.code):
563         (div.code pre):
564         (dt):
565         (dd):
566         (dd:last-child):
567         (.site-logo):
568         (.site-logo .tagline):
569         (table):
570         (#contents):
571         (p):
572         (p:last-child):
573         (th):
574         (td):
575
576 2016-06-28  Jon Lee  <jonlee@apple.com>
577
578         Update focus test
579         https://bugs.webkit.org/show_bug.cgi?id=159242
580         rdar://problem/27070007
581
582         Reviewed by Dean Jackson.
583         Provisionally reviewed by Said Abou-Hallawa.
584
585         Move previous test to dom suite, and update the test for better reporting of frame rate, although
586         it uses a different rendering path.
587
588         * Animometer/resources/debug-runner/tests.js: Add to dom suite.
589         * Animometer/tests/dom/focus.html: Copied from PerformanceTests/Animometer/tests/master/focus.html.
590         * Animometer/tests/dom/resources/focus.js: Copied from PerformanceTests/Animometer/tests/master/resources/focus.js.
591         * Animometer/tests/master/focus.html: Remove center element.
592         * Animometer/tests/master/resources/focus.js: Use narrower size range with smaller particles. Remove the
593         container elements. Inline getBlurValue and getOpacityValue since they are only called once.
594
595 2016-06-28  Filip Pizlo  <fpizlo@apple.com>
596
597         Move Air.js and Basic into ES6SampleBench
598
599         Rubber stamped by Geoffrey Garen.
600         
601         We want to group our ES6 benchmarks together, and eventually, we want to give them a
602         common harness.
603
604         * Air.js: Removed.
605         * Air.js/README.md: Removed.
606         * Air.js/airjs-tests.yaml: Removed.
607         * Air.js/all.js: Removed.
608         * Air.js/allocate_stack.js: Removed.
609         * Air.js/arg.js: Removed.
610         * Air.js/basic_block.js: Removed.
611         * Air.js/benchmark.js: Removed.
612         * Air.js/code.js: Removed.
613         * Air.js/custom.js: Removed.
614         * Air.js/frequented_block.js: Removed.
615         * Air.js/insertion_set.js: Removed.
616         * Air.js/inst.js: Removed.
617         * Air.js/liveness.js: Removed.
618         * Air.js/make_dist.sh: Removed.
619         * Air.js/opcode.js: Removed.
620         * Air.js/payload-airjs-ACLj8C.js: Removed.
621         * Air.js/payload-gbemu-executeIteration.js: Removed.
622         * Air.js/payload-imaging-gaussian-blur-gaussianBlur.js: Removed.
623         * Air.js/payload-typescript-scanIdentifier.js: Removed.
624         * Air.js/reg.js: Removed.
625         * Air.js/stack_slot.js: Removed.
626         * Air.js/stress-test.js: Removed.
627         * Air.js/strip-hash.rb: Removed.
628         * Air.js/symbols.js: Removed.
629         * Air.js/test.html: Removed.
630         * Air.js/test.js: Removed.
631         * Air.js/tmp.js: Removed.
632         * Air.js/tmp_base.js: Removed.
633         * Air.js/util.js: Removed.
634         * Basic: Removed.
635         * Basic/ast.js: Removed.
636         * Basic/basic.js: Removed.
637         * Basic/benchmark.js: Removed.
638         * Basic/caseless_map.js: Removed.
639         * Basic/lexer.js: Removed.
640         * Basic/number.js: Removed.
641         * Basic/parser.js: Removed.
642         * Basic/random.js: Removed.
643         * Basic/state.js: Removed.
644         * Basic/test.html: Removed.
645         * Basic/test.js: Removed.
646         * Basic/util.js: Removed.
647         * ES6SampleBench: Added.
648         * ES6SampleBench/Air: Copied from PerformanceTests/Air.js.
649         * ES6SampleBench/Basic: Copied from PerformanceTests/Basic.
650         * Skipped:
651
652 2016-06-24  Filip Pizlo  <fpizlo@apple.com>
653
654         Add a ES6 generator benchmark
655         https://bugs.webkit.org/show_bug.cgi?id=159101
656
657         Rubber stamped by Keith Miller.
658         
659         This adds a Basic interpreter loosely based on ECMA-55:
660         
661         http://www.ecma-international.org/publications/files/ECMA-ST-WITHDRAWN/ECMA-55,%201st%20Edition,%20January%201978.pdf
662         
663         It includes a lexer that is a generator, a parser that uses regular expressions, and an AST
664         walk interpreter where the walking functions for statements are generators that call each
665         other with yield*. This enables the interpreter to look like an AST walk even though it can
666         yield at INPUT and PRINT statements.
667         
668         This also uses for-of, classes, Map, and WeakMap. It also uses deprecated-but-awesome RegExp
669         features like RegExp.lastMatch and RegExp.rightContext. I did it that way because this is
670         how I've always written lexers in dynamic languages; see offlineasm's lex() method in
671         parser.rb for example.
672         
673         The benchmark runs a handful of simple Basic programs. The longest-running one computes
674         prime numbers.
675         
676         Includes a command-line and web harness. On my machine it runs in 2-3 seconds.
677
678         * Basic: Added.
679         * Basic/ast.js: Added.
680         (Basic.NumberApply):
681         (Basic.Variable):
682         (Basic.Const):
683         (Basic.NumberPow):
684         (Basic.NumberMul):
685         (Basic.NumberDiv):
686         (Basic.NumberNeg):
687         (Basic.NumberAdd):
688         (Basic.NumberSub):
689         (Basic.StringVar):
690         (Basic.Equals):
691         (Basic.NotEquals):
692         (Basic.LessThan):
693         (Basic.GreaterThan):
694         (Basic.LessEqual):
695         (Basic.GreaterEqual):
696         (Basic.GoTo):
697         (Basic.GoSub):
698         (Basic.Def):
699         (Basic.Let):
700         (Basic.If):
701         (Basic.Return):
702         (Basic.Stop):
703         (Basic.On):
704         (sideState.shouldStop):
705         (Basic.For):
706         (Basic.Next):
707         (Basic.Print):
708         (Basic.Input):
709         (Basic.Read):
710         (Basic.Restore):
711         (Basic.Dim):
712         (Basic.Randomize):
713         (Basic.End):
714         (Basic.Program):
715         * Basic/basic.js: Added.
716         (prepare):
717         (simulate):
718         * Basic/benchmark.js: Added.
719         (Benchmark):
720         (Benchmark.prototype.runIteration.expect):
721         (Benchmark.prototype.runIteration):
722         (runBenchmark):
723         * Basic/caseless_map.js: Added.
724         (CaselessMap):
725         * Basic/lexer.js: Added.
726         (lex.consumeWhitespace):
727         (lex.consume):
728         (lex):
729         * Basic/lexer_test.js: Added.
730         * Basic/number.js: Added.
731         (NumberValue):
732         (NumberValue.prototype.get value):
733         (NumberValue.prototype.apply):
734         (NumberValue.prototype.leftApply):
735         (NumberValue.prototype.assign):
736         (NumberArray.):
737         (NumberArray):
738         (NumberArray.prototype.apply):
739         (NumberArray.prototype.leftApply):
740         (NumberFunction):
741         (NumberFunction.prototype.apply):
742         (NumberFunction.prototype.leftApply):
743         (NativeFunction):
744         (NativeFunction.prototype.apply):
745         (NativeFunction.prototype.leftApply):
746         * Basic/parser.js: Added.
747         (parse):
748         (parse.pushToken):
749         (parse.peekToken):
750         (parse.consumeKind):
751         (parse.consumeToken):
752         (parse.parseVariable):
753         (parse.parseNumericExpression.parsePrimary):
754         (parse.parseNumericExpression.parseFactor):
755         (parse.parseNumericExpression.parseTerm):
756         (parse.parseNumericExpression):
757         (parse.parseConstant):
758         (parse.parseStringExpression):
759         (parse.isStringExpression):
760         (parse.parseRelationalExpression):
761         (parse.parseNonNegativeInteger):
762         (parse.parseGoToStatement):
763         (parse.parseGoSubStatement):
764         (parse.parseStatement):
765         (parse.parseStatements):
766         * Basic/random.js: Added.
767         (createRNG):
768         (createRNGWithFixedSeed):
769         (createRNGWithRandomSeed):
770         * Basic/state.js: Added.
771         (State):
772         (State.prototype.getValue):
773         (State.prototype.getSideState):
774         (State.prototype.abort):
775         (State.prototype.validate):
776         * Basic/test.html: Added.
777         * Basic/test.js: Added.
778         * Basic/util.js: Added.
779         (this.performance.performance.now.currentTime):
780         (else.this.preciseTime.currentTime):
781         (else.currentTime):
782         * Skipped: Make sure that we don't run Basic yet.
783
784 2016-06-21  Jon Lee  <jonlee@apple.com>
785
786         Update canvas size when benchmark begins
787         https://bugs.webkit.org/show_bug.cgi?id=159010
788
789         Reviewed by Dean Jackson.
790
791         Previously we would evaluate the media queries and assign the canvas size to the test
792         on body.onload. Instead, do it once the user starts the benchmark.
793
794         * Animometer/resources/debug-runner/animometer.js: Move benchmarkController.determineCanvasSize()
795         to benchmarkController._startBenchmark.
796         * Animometer/resources/runner/animometer.js: Ditto.
797
798 2016-06-21  Jon Lee  <jonlee@apple.com>
799
800         Add new timestamp option
801         https://bugs.webkit.org/show_bug.cgi?id=159006
802
803         Reviewed by Dean Jackson.
804
805         Add a new option to take timestamps from the rAF callback.
806
807         * Animometer/developer.html: Added option.
808         * Animometer/tests/resources/main.js:
809         (Benchmark): Fall back to using the rAF timestamp if performance.now()
810         is not available.
811         (Benchmark._animateLoop): Update how timestamp is set. Prefer to use
812         local var instead of accessing private var.
813
814 2016-06-21  Jon Lee  <jonlee@apple.com>
815
816         Improvements to Animometer benchmark
817         https://bugs.webkit.org/show_bug.cgi?id=157738
818
819         Reviewed by Dean Jackson.
820         Provisionally reviewed by Said Abou-Hallawa.
821
822         Update tests.
823
824         * Animometer/tests/master/text.html: Ensure only three text sizes for
825         the three canvases.
826         * Animometer/tests/master/focus.html: Reduce the text size to fit with smaller
827         particle sizes.
828         * Animometer/tests/master/resources/focus.js: Remove the quadratic distribution for
829         particle sizes, and make it linear. Reduce the size variance. Shuffle the math to
830         reduce some calculations per frame. Fix the placement of the particles which might
831         otherwise be culled.
832         * Animometer/tests/master/resources/image-data.js: Reduce the particle size
833         to encourage larger scores.
834
835 2016-06-21  Jon Lee  <jonlee@apple.com>
836
837         Improvements to Animometer benchmark
838         https://bugs.webkit.org/show_bug.cgi?id=157738
839
840         Reviewed by Dean Jackson.
841         Provisionally reviewed by Said Abou-Hallawa.
842
843         Include confidence interval for the final score, and store the canvas
844         size in the serialization so that it is accurately shown in results.
845
846         * Animometer/developer.html: Add a "confidence" div.
847         * Animometer/index.html: Ditto. Convert "mean" to "confidence".
848         * Animometer/resources/debug-runner/animometer.js: Look at options, and
849         if the configuration is included, update the body class based on it
850         (similar to what we do when we first load the page). That way, if you
851         drag-and-drop previous results in, that configuration is reflected in
852         the dashboard. Show the full confidence interval.
853
854         * Animometer/resources/debug-runner/animometer.css:
855         * Animometer/resources/debug-runner/animometer.js: Style update.
856         * Animometer/resources/runner/animometer.css:
857
858         * Animometer/resources/runner/animometer.js:
859         (_processData): Propagate the confidence interval values out and calculate
860         the lower and upper bounds. For now, shortcut the aggregate calculation,
861         since we only go through one iteration.
862         (this._processData.calculateScore): Propagate the confidence interval out
863         to be next to the score. Depending on the controller these values are
864         calculated differently.
865         (this._processData._getResultsProperty): Convenience function.
866         (this._processData.get score): Refactor.
867         (this._processData.get scoreLowerBound): Get the aggregate lower bound.
868         (this._processData.get scoreUpperBound): Get the aggregate upper bound.
869         (window.sectionsManager.setSectionScore):
870         (window.benchmarkController._startBenchmark): When the benchmark starts, note
871         the canvas size and add it to options. That way it will be automatically be
872         serialized.
873         (window.benchmarkController.showResults): Include the maximum deviation
874         percentage.
875         * Animometer/resources/runner/lines.svg: Make the background line up with the
876         skew.
877         * Animometer/resources/runner/tests.js:
878         (Headers.details.text): Refactor.
879         * Animometer/resources/statistics.js:
880         (Statistics.largestDeviationPercentage): Convenience function to calculate
881         the deviation percentage on either end and return the largest deviation.
882         * Animometer/resources/strings.js:
883
884         Allow specifying a regression profile to use instead of taking the one
885         with the lowest error.
886
887         Address an issue in the focus test when the regression calculated ends
888         up overestimating the change point, causing a cascade of tougher
889         ramps. The reason behind this is that at max complexity of an initial
890         ramp, the frame length is very high, and it influences the second
891         segment of the piecewise regression strongly, causing it to be very
892         steep. As a result, the first segment, expected to be flat, ends up
893         covering a higher range of complexity. That makes the change point
894         much higher than it should be. To avoid this, we will add a sanity
895         check on the maximum value of the ramp. If the regression's projected
896         value at the maximum complexity of the current ramp is very slow (less
897         than 20 fps), 1) reduce the maximum complexity by 20%, and 2) do not
898         include the regression's change point in the change point estimator.
899         That estimator is used as the midpoint of the next ramp, and including
900         the change point from a poor regression can bake in the error. The
901         controller already knows how to adjust for ramps that are too easy for
902         the system.
903
904         * Animometer/resources/runner/animometer.js:
905         (this._processData.findRegression): Takes a preferred profile and gives that to
906         Regression.
907         (this._processData.calculateScore): With the ramp controller, take the profile
908         of the ramp that was used the most when calculating the ramp's regression. That
909         profile is what is used for the test's score.
910         * Animometer/resources/statistics.js:
911         (Regression.Utilities.createClass): Update to take an options object which can
912         specify a profile to calculate with. Otherwise it will continue to use both and
913         select the one with the lower error.
914         (_calculateRegression): Fix an issue where we claim 0 error if the regression
915         calculation fails due to divide-by-zero. Instead reject that regression calculation
916         by giving it Number.MAX_VALUE.
917         * Animometer/resources/strings.js: New strings for marking a regression as flat
918         or slope.
919         * Animometer/tests/resources/main.js:
920         (RampController): Rename the thresholds for clarity. Add a threshold that, if
921         exceeded, will lower the maximum complexity of the next ramp.
922         (tune): Relax the cdf check to consider whether the interval definitely falls in
923         the desired frame length threshold.
924         (processSamples): Include the profile in the ramp.
925
926         Update ramp controller test. Increase the length of the test to 30 seconds, and extend
927         the interval to 120 ms during sampling. Improve the estimation of the ramp parameters.
928
929         * Animometer/developer.html: Change default to 30 seconds, and don't show the progress bar
930         by default.
931         * Animometer/resources/runner/animometer.js: Change default to 30 seconds.
932         * Animometer/tests/resources/main.js: A number of improvements to the ramp controller, in
933         the order in which they appear in the patch:
934
935         - With a longer test length use longer ramps with longer intervals to get more data at each
936         complexity. Keep the 100 ms interval length during the ramp up phase since we don't need to
937         spend more time there to find the right order of magnitude, but increase it during the
938         ramps to 120 ms.
939         - The ramp linearly interpolates the complexity to render based on its timestamp, but it would
940         never sample the minimum complexity. Instead of lerping max to min complexity from time
941         0 to t where t is the ramp length, instead lerp from 0 to (t - intervalSampleLength) so that
942         we can have at least one interval sample at the min complexity for that ramp.
943         - Some regression calculations only come out with one line segment rather than the two
944         we expect. This could be due to a noisy ramp or the ramp's range is too narrow. If that's the
945         case, influence the minimum complexity of the next ramp towards the lowest bound of 1, so that
946         we ensure that at least part of the ramp is covering a complexity range that the system can
947         handle at full 60.
948         - Remove an assignment to interpolatedFrameLength since that is never subsequently used.
949
950         Update the format used to serialize the results for analysis.
951
952         Each data point used to have named properties for fields like complexity and frame rate.
953         In addition the serialized numbers had rounding errors that took up many characters.
954         Update the format by introducing a new data container called SampleData, which contains a
955         field map. The map maps a string to an array index. Each data point is an array, so, to
956         get a stat, use the field map to get the array index into the data point. This allows future
957         versions to track other data, and reduces the size of the output string by two-thirds.
958
959         * Animometer/resources/extensions.js:
960         (Utilities.toFixedNumber): Add convenience function that truncates the number to a fixed
961         precision, and converts it back to a number.
962         (SampleData): New class that contains sample data and a field map that maps properties to
963         an array index.
964         (get length): Number of data points.
965         (addField): Add a field to the field map.
966         (push): Add a data point.
967         (sort): Sort the data.
968         (slice): Return new SampleData object with sliced data.
969         (forEach): Iterate over the data with provided function.
970         (createDatum):
971         (getFieldInDatum): Returns the data point associated with the field name by looking it up
972         in the field map in the datum provided, which can be the datum object itself (an array) or
973         an index into the data member variable.
974         (setFieldInDatum): Sets the data point associated with the field name.
975         (at): Returns the data point at the provided index.
976         (toArray): Serializes the data where the field map serves as property names for each point.
977
978         * Animometer/resources/debug-runner/graph.js:
979         (updateGraphData): Remove unused _testData. Convert the data to the old array format for the
980         graph to use, since the old format was much easier to work with when displaying the graphs.
981         (onGraphTypeChanged): For some controllers, no alternative score or mean is provided.
982         * Animometer/resources/runner/animometer.css:
983         * Animometer/resources/runner/animometer.js: Refactor to use SampleData. Update JSON output
984         to only go to 3 digits of precision for purposes of reducing the data size.
985         * Animometer/resources/strings.js: Add new strings to put into the field maps.
986         * Animometer/tests/resources/main.js: Refactor to use SampleData.
987
988         * Animometer/developer.html:
989         * Animometer/index.html: Restructure results table for both pages. Add charset attribute to
990         tests.js include.
991         * Animometer/resources/debug-runner/animometer.css: Clear out styles from release runner.
992         * Animometer/resources/debug-runner/graph.js:
993         (onGraphTypeChanged): Update score and mean if bootstrap results are available from the
994         controller, since not all controllers do bootstrapping.
995         * Animometer/resources/debug-runner/tests.js: Update header text.
996         * Animometer/resources/runner/animometer.css: Include confidence interval in results.
997         * Animometer/resources/runner/animometer.js:
998         (ResultsTable._addHeader): Header contents can be HTML, so use innerHTML instead.
999         (ResultsTable._addBody): Add tbody element.
1000         (ResultsTable._addTest): Allow a data cell to invoke a JS function to get its contents.
1001         (window.benchmarkController.showResults): Add table that includes tests' confidence intervals.
1002         * Animometer/resources/runner/tests.js:
1003         (Headers.details.text): Add new details table that includes bootstrap confidence interval.
1004         The interval can be asymmetric, but for simplicity, report the maximum deviation percentage
1005         on either side of the bootstrap median.
1006         * Animometer/resources/statistics.js:
1007         (bootstrap): Include the confidence percentage in the return object.
1008
1009         Report canvas size in results.
1010
1011         * Animometer/developer.html: Add markup to indicate whether a small, medium, or large
1012         canvas was used.
1013         * Animometer/index.html: Ditto.
1014         * Animometer/resources/debug-runner/animometer.js: Call determineCanvasSize().
1015         * Animometer/resources/runner/animometer.css: Update styles to set the canvas based on the
1016         body class size.
1017         * Animometer/resources/runner/animometer.js:
1018         (window.benchmarkController.initialize): Update styles to set the canvas based on the
1019         body class size.
1020         (window.benchmarkController.determineCanvasSize): Run various media queries and set the body
1021         class based on the size of the device.
1022
1023         * Animometer/developer.html: Refactor to include the main CSS file, and redo
1024         the layout so that it doesn't rely on flexbox.
1025         * Animometer/resources/debug-runner/animometer.css:
1026         * Animometer/resources/debug-runner/animometer.js:
1027         (updateDisplay): Since various parts of the script alter the body class, we can't
1028         replace the className directly. Instead, remove all display-based values and then add
1029         the one that was selected.
1030         * Animometer/resources/debug-runner/graph.js:
1031         (updateGraphData): To set the size of the graph, use window.innerHeight.
1032         * Animometer/resources/runner/animometer.js:
1033         (window.sectionsManager.showSection): Since various parts of the script alter the body
1034         class, we can't replace the className directly. Remove all of the section classes
1035         individually and then add the one desired.
1036         * Animometer/tests/resources/stage.css: Remove -apple-system as a font to use in the
1037         stage.
1038
1039 2016-06-12  Filip Pizlo  <fpizlo@apple.com>
1040
1041         Fix round-down goof in Air.js's ShuffleCustom.forEachArg
1042         https://bugs.webkit.org/show_bug.cgi?id=158674
1043
1044         Reviewed by Michael Saboff.
1045         
1046         x / 3 * 3 is not how you round down to multiples of 3 in JavaScript. You need to do
1047         Math.floor(x / 3) * 3 instead.
1048         
1049         This is a benign change, because having extra arguments to a Shuffle is not something we
1050         actually take advantage of yet. But I think it's best to match the original C++ code's
1051         intent.
1052
1053         * Air.js/custom.js:
1054         (const.ShuffleCustom.forEachArg):
1055
1056 2016-06-11  Filip Pizlo  <fpizlo@apple.com>
1057
1058         Unreviewed, skip Air.js. It got unskipped when I renamed it from JSAir.
1059
1060         * Skipped:
1061
1062 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1063
1064         Make it easy to package Air.js
1065         https://bugs.webkit.org/show_bug.cgi?id=158652
1066
1067         Reviewed by Benjamin Poulain.
1068         
1069         If you want to give Air.js to someone, you can now do:
1070         
1071         cd PerformanceTests/Air.js
1072         ./make_dist.sh
1073         
1074         This will create Air.js.tar.gz and a directory called Air.js. If you want to send someone
1075         the benchmark, just send either of those things and tell them to look at index.html.
1076         
1077         You may have to edit make_dist.sh for wherever you have Markdown.pl. You can get that from
1078         https://daringfireball.net/projects/markdown/
1079
1080         * Air.js/make_dist.sh: Added.
1081         * Air.js/test.html:
1082
1083 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1084
1085         Unreviewed, make the link to test.html into an actual link.
1086
1087         * Air.js/README.md:
1088
1089 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1090
1091         Fix some text in Air.js/README.md
1092         https://bugs.webkit.org/show_bug.cgi?id=158650
1093
1094         Reviewed by Benjamin Poulain.
1095         
1096         I read the text again and found bugs:
1097         
1098         - We never actually say how to run the benchmark. This change adds a blurb about how to run
1099           it.
1100
1101         - We both say that allocateStack is responsible for the bulk of the running time and that
1102           we haven't measured where the bulk of the time is spent. This changes the text to say that
1103           it was a goal to make allocateStack be the hottest part of the benchmark, but that we did
1104           not measure this.
1105
1106         * Air.js/README.md:
1107
1108 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1109
1110         Air.js should have some documentation
1111         https://bugs.webkit.org/show_bug.cgi?id=158648
1112
1113         Reviewed by Keith Miller.
1114         
1115         I want to be able to point people at a document if they want to know more about this
1116         benchmark.
1117
1118         * Air.js/README.md: Added.
1119
1120 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1121
1122         Rename JSAir to Air.js.
1123
1124         Rubber stamped by Mark Lam.
1125
1126         * Air.js: Copied from PerformanceTests/JSAir.
1127         * Air.js/airjs-tests.yaml: Copied from PerformanceTests/JSAir/jsair-tests.yaml.
1128         * Air.js/benchmark.js: Copied from PerformanceTests/JSAir/benchmark.js.
1129         (Benchmark):
1130         * Air.js/jsair-tests.yaml: Removed.
1131         * Air.js/payload-airjs-ACLj8C.js: Copied from PerformanceTests/JSAir/payload-jsair-ACLj8C.js.
1132         * Air.js/payload-jsair-ACLj8C.js: Removed.
1133         * Air.js/stress-test.js: Copied from PerformanceTests/JSAir/stress-test.js.
1134         * Air.js/test.html: Copied from PerformanceTests/JSAir/test.html.
1135         * Air.js/test.js: Copied from PerformanceTests/JSAir/test.js.
1136         * JSAir: Removed.
1137         * JSAir/all.js: Removed.
1138         * JSAir/allocate_stack.js: Removed.
1139         * JSAir/arg.js: Removed.
1140         * JSAir/basic_block.js: Removed.
1141         * JSAir/benchmark.js: Removed.
1142         * JSAir/code.js: Removed.
1143         * JSAir/custom.js: Removed.
1144         * JSAir/frequented_block.js: Removed.
1145         * JSAir/insertion_set.js: Removed.
1146         * JSAir/inst.js: Removed.
1147         * JSAir/jsair-tests.yaml: Removed.
1148         * JSAir/liveness.js: Removed.
1149         * JSAir/opcode.js: Removed.
1150         * JSAir/payload-gbemu-executeIteration.js: Removed.
1151         * JSAir/payload-imaging-gaussian-blur-gaussianBlur.js: Removed.
1152         * JSAir/payload-jsair-ACLj8C.js: Removed.
1153         * JSAir/payload-typescript-scanIdentifier.js: Removed.
1154         * JSAir/reg.js: Removed.
1155         * JSAir/stack_slot.js: Removed.
1156         * JSAir/stress-test.js: Removed.
1157         * JSAir/strip-hash.rb: Removed.
1158         * JSAir/symbols.js: Removed.
1159         * JSAir/test.html: Removed.
1160         * JSAir/test.js: Removed.
1161         * JSAir/tmp.js: Removed.
1162         * JSAir/tmp_base.js: Removed.
1163         * JSAir/util.js: Removed.
1164
1165 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1166
1167         JSC Stress Test failing: jsair-tests.yaml/test.js.ftl-eager-no-cjit
1168         https://bugs.webkit.org/show_bug.cgi?id=158571
1169
1170         Reviewed by Keith Miller.
1171         
1172         Introduce a different harness for run-jsc-stress-tests, which runs for a minimum of 10
1173         iterations and then tries to do any number of "bonus" iterations until it's been running for
1174         two seconds. Since this is the sort of test that isn't really meant to stress anything in
1175         particular, I think it's OK if it is time-limited in this way. The worst case is that some
1176         of its failures will be flaky, but I think that they would have been flaky anyway given the
1177         complexity of the test.
1178
1179         * JSAir/benchmark.js:
1180         (Benchmark):
1181         (Benchmark.prototype.runIteration):
1182         (benchmark): Deleted.
1183         * JSAir/jsair-tests.yaml:
1184         * JSAir/stress-test.js: Added.
1185         (preciseTime):
1186         * JSAir/test.html:
1187         * JSAir/test.js:
1188
1189 2016-06-09  Filip Pizlo  <fpizlo@apple.com>
1190
1191         Unreviewed, teach the perf bots not to run JSAir.
1192
1193         * Skipped:
1194
1195 2016-06-08  Filip Pizlo  <fpizlo@apple.com>
1196
1197         Use more ES6 features in JSAir
1198         https://bugs.webkit.org/show_bug.cgi?id=158497
1199
1200         Reviewed by Keith Miller.
1201         
1202         This improves JSAir with the following ES6 features suggested by JoePeck:
1203         
1204         - String interpolation.
1205         - Destructuring inside PatchCustom.
1206         - Default arguments.
1207         
1208         All of these things are on hot paths.
1209         
1210         Note that I didn't use string interpolation everywhere that I could, only in those places
1211         where it made the code more readable. In Ruby, I used the style that if the interpolation
1212         expression has any non-trivial stuff (like a ternary operator, a chain of calls, or embedded
1213         strings) then it's better to use regular strcat. I think that's what I carried over to here.
1214         
1215         Note that the previous change (Add result validation to JSAir) also made the Proxy code not
1216         dead, though it's not necessarily on the hot path. The Proxy isn't called into frequently
1217         but it's used from a function that is otherwise hot, so if calling into the Proxy prevents
1218         that function from being optimized then it will hurt so good.
1219         
1220         I also reenabled tail calls in a few places.
1221         
1222         This change doesn't seem to change the performance of the benchmark for us. That's expected
1223         since these ES6 features are cheap. Note that this claim doesn't include Proxy, which was
1224         added in a separate change and that change did make the benchmark overall more expensive.
1225
1226         * JSAir/allocate_stack.js:
1227         (allocateStack):
1228         * JSAir/arg.js:
1229         (Arg.createBitImm64):
1230         (Arg.createAddr):
1231         (Arg.createStack):
1232         (Arg.logScale):
1233         (Arg.createIndex):
1234         * JSAir/basic_block.js:
1235         (BasicBlock.get headerString):
1236         (BasicBlock.prototype.get if):
1237         (BasicBlock):
1238         * JSAir/benchmark.js:
1239         (benchmark):
1240         * JSAir/code.js:
1241         (Code):
1242         (Code.prototype.addBlock):
1243         (Code.prototype.addStackSlot):
1244         (Code.prototype.newTmp):
1245         (Code.prototype.get size):
1246         (Code.prototype.get blocks):
1247         (Code.prototype.get stackSlots):
1248         (Code.prototype.tmps):
1249         (Code.prototype.get callArgAreaSize):
1250         (Code.prototype.toString):
1251         * JSAir/custom.js:
1252         (const.PatchCustom.forEachArg):
1253         * JSAir/inst.js:
1254         (Inst):
1255         * JSAir/reg.js:
1256         (Reg.prototype.toString):
1257         * JSAir/util.js:
1258         (symbolName):
1259         (lowerSymbolName):
1260
1261 2016-06-07  Filip Pizlo  <fpizlo@apple.com>
1262
1263         Add result validation to JSAir
1264         https://bugs.webkit.org/show_bug.cgi?id=158493
1265
1266         Reviewed by Saam Barati.
1267         
1268         This adds the ability to hash a Code in a way that matches the C++ code's hashing of Code.
1269         This allows us to check if the Code that JSAir sees is the code that C++ saw. We use this to
1270         check the Code before and after allocateStack, and compare against hashes we got from C++.
1271         
1272         Doing this uncovered bugs. roundUpToMultipleOf wasn't doing anything. allocateStack was not
1273         allocating things correctly because I was concatting a Set to an Array, which doesn't really
1274         work. Now these bugs are fixed.
1275         
1276         The checking step adds to the running time so I reduced the number of iterations. The
1277         benchmark spends a decent amount of its time computing Code hashes; I think it's around 1/3
1278         total. This is probably OK. It's better to verify the results even if the running time is
1279         not all in the "core" of the algorithm.
1280         
1281         Also add a run-jsc-stress-tests yaml file to allow this to run as a test.
1282
1283         * JSAir/allocate_stack.js:
1284         * JSAir/arg.js:
1285         (Arg.createImm):
1286         (Arg.createBigImm):
1287         (Arg.createBitImm):
1288         (Arg.createBitImm64):
1289         (Arg.createWidth):
1290         (Arg.createSpecial):
1291         (Arg.prototype.get kind):
1292         (Arg.prototype.get isTmp):
1293         (Arg.prototype.get isImm):
1294         (Arg.prototype.get isSomeImm):
1295         (Arg.prototype.get isSomeBigImm):
1296         (Arg.prototype.get isCondition):
1297         (Arg.prototype.get isWidth):
1298         (Arg.prototype.get isSpecial):
1299         (Arg.prototype.get isAlive):
1300         (Arg.prototype.get tmp):
1301         (Arg.prototype.get value):
1302         (Arg.prototype.get lowValue):
1303         (Arg.prototype.get highValue):
1304         (Arg.prototype.get base):
1305         (Arg.prototype.get isGP):
1306         (Arg.prototype.get isFP):
1307         (Arg.prototype.isValidForm):
1308         (Arg.prototype.get isInvertible):
1309         (Arg.kindCode):
1310         (Arg.prototype.hash):
1311         (Arg.prototype.toString):
1312         (Arg):
1313         * JSAir/basic_block.js:
1314         (BasicBlock.get successorBlocks):
1315         * JSAir/benchmark.js:
1316         (benchmark):
1317         * JSAir/code.js:
1318         (Code.prototype.setFrameSize):
1319         (Code.prototype.hash):
1320         (Code.prototype.toString):
1321         (Code):
1322         * JSAir/inst.js:
1323         (Inst.prototype.get hasNonArgEffects):
1324         (Inst.prototype.hash):
1325         (Inst.prototype.toString):
1326         (Inst):
1327         * JSAir/jsair-tests.yaml: Added.
1328         * JSAir/opcode.js:
1329         (Inst_forEachArg):
1330         (Inst_hasNonArgEffects):
1331         (opcodeCode):
1332         * JSAir/payload-gbemu-executeIteration.js:
1333         (createPayloadGbemuExecuteIteration):
1334         * JSAir/payload-imaging-gaussian-blur-gaussianBlur.js:
1335         (createPayloadImagingGaussianBlurGaussianBlur):
1336         * JSAir/payload-jsair-ACLj8C.js:
1337         (createPayloadJSAirACLj8C):
1338         * JSAir/payload-typescript-scanIdentifier.js:
1339         (createPayloadTypescriptScanIdentifier):
1340         * JSAir/reg.js:
1341         (Reg.prototype.get isReg):
1342         (Reg.prototype.hash):
1343         (Reg.prototype.toString):
1344         * JSAir/stack_slot.js:
1345         (StackSlot.prototype.setOffsetFromFP):
1346         (StackSlot.prototype.hash):
1347         (StackSlot.prototype.toString):
1348         * JSAir/symbols.js:
1349         (relCondCode):
1350         (resCondCode):
1351         (doubleCondCode):
1352         * JSAir/test.html:
1353         * JSAir/tmp.js:
1354         (Tmp.prototype.get isReg):
1355         (Tmp.prototype.hash):
1356         (Tmp.prototype.toString):
1357         * JSAir/util.js:
1358         (roundUpToMultipleOf):
1359         (symbolName):
1360
1361 2016-06-07  Filip Pizlo  <fpizlo@apple.com>
1362
1363         Implement Air::allocateStack() in ES6 to see how much of a bad idea that is
1364         https://bugs.webkit.org/show_bug.cgi?id=158318
1365
1366         Reviewed by Saam Barati.
1367         
1368         This adds a new benchmark for us to play with called JSAir. It's a complete ES6
1369         implementation of Air's allocateStack() phase along with all of Air needed to run that
1370         phase. This includes things like stack slots, registers, temporaries, basic blocks,
1371         instructions, and all of the code for iterating over, inspecting, and modifying those
1372         things.
1373         
1374         To make this work, JSC can now dump Air just before allocateStack() in the form of JS code
1375         that creates a Code object that matches exactly what C++ Air saw. This benchmark comprises
1376         four Air IRs:
1377         
1378         - Octane/gbemu's largest function, executeIteration.
1379         - Kraken/imaging-gaussian-blur's largest function in OSR entry mode, gaussuanBlur.
1380         - Octane/typescript's largest function that is actually hot, scanIdentifier.
1381         - JSAir's largest hot function, which is anonymous, so we call it by its hash (ACLj8C).
1382         
1383         This runs in about 2 seconds on my machine in JSC trunk. It includes both a commandline
1384         harness and a web harness. JSAir is almost exactly 100x slower in ES6 in WebKit than the
1385         C++ Air::allocateStack() phase on which it is based.
1386         
1387         JSAir uses the following ES6 features:
1388         
1389         - Symbol.
1390         - for-of.
1391         - arrow functions.
1392         - Map/Set.
1393         - let/const.
1394         - classes.
1395         
1396         All of these things are used in anger and should end up on the hot path. There is also code
1397         that uses Proxies, but it ends up being dead.
1398         
1399         We can improve this even more:
1400         I still need to add result validation: https://bugs.webkit.org/show_bug.cgi?id=158493
1401         I want to make it use more ES6 features: https://bugs.webkit.org/show_bug.cgi?id=158497
1402         
1403         * JSAir: Added.
1404         * JSAir/all.js: Added.
1405         * JSAir/allocate_stack.js: Added.
1406         (allocateStack.attemptAssignment):
1407         (allocateStack.assign):
1408         (allocateStack.interfere):
1409         (allocateStack.):
1410         (allocateStack):
1411         * JSAir/arg.js: Added.
1412         (Arg):
1413         (Arg.isAnyUse):
1414         (Arg.isColdUse):
1415         (Arg.isWarmUse):
1416         (Arg.cooled):
1417         (Arg.isEarlyUse):
1418         (Arg.isLateUse):
1419         (Arg.isAnyDef):
1420         (Arg.isEarlyDef):
1421         (Arg.isLateDef):
1422         (Arg.isZDef):
1423         (Arg.typeForB3Type):
1424         (Arg.widthForB3Type):
1425         (Arg.conservativeWidth):
1426         (Arg.minimumWidth):
1427         (Arg.bytes):
1428         (Arg.widthForBytes):
1429         (Arg.createTmp):
1430         (Arg.fromReg):
1431         (Arg.createImm):
1432         (Arg.createBigImm):
1433         (Arg.createBitImm):
1434         (Arg.createBitImm64):
1435         (Arg.createAddr):
1436         (Arg.createStack):
1437         (Arg.createCallArg):
1438         (Arg.createStackAddr):
1439         (Arg.isValidScale):
1440         (Arg.logScale):
1441         (Arg.createIndex):
1442         (Arg.createRelCond):
1443         (Arg.createResCond):
1444         (Arg.createDoubleCond):
1445         (Arg.createWidth):
1446         (Arg.prototype.get kind):
1447         (Arg.prototype.get isTmp):
1448         (Arg.prototype.get isImm):
1449         (Arg.prototype.get isBigImm):
1450         (Arg.prototype.get isBitImm):
1451         (Arg.prototype.get isBitImm64):
1452         (Arg.prototype.get isSomeImm):
1453         (Arg.prototype.get isAddr):
1454         (Arg.prototype.get isStack):
1455         (Arg.prototype.get isCallArg):
1456         (Arg.prototype.get isIndex):
1457         (Arg.prototype.get isMemory):
1458         (Arg.prototype.get isStackMemory):
1459         (Arg.prototype.get isRelCond):
1460         (Arg.prototype.get isResCond):
1461         (Arg.prototype.get isDoubleCond):
1462         (Arg.prototype.get isCondition):
1463         (Arg.prototype.get isWidth):
1464         (Arg.prototype.get isAlive):
1465         (Arg.prototype.get tmp):
1466         (Arg.prototype.get value):
1467         (Arg.prototype.get base):
1468         (Arg.prototype.get hasOffset):
1469         (Arg.prototype.get offset):
1470         (Arg.prototype.get stackSlot):
1471         (Arg.prototype.get index):
1472         (Arg.prototype.get scale):
1473         (Arg.prototype.get logScale):
1474         (Arg.prototype.get width):
1475         (Arg.prototype.get isGPTmp):
1476         (Arg.prototype.get isFPTmp):
1477         (Arg.prototype.get isGP):
1478         (Arg.prototype.get isFP):
1479         (Arg.prototype.get hasType):
1480         (Arg.prototype.get type):
1481         (Arg.prototype.isType):
1482         (Arg.prototype.isCompatibleType):
1483         (Arg.prototype.get isGPR):
1484         (Arg.prototype.get gpr):
1485         (Arg.prototype.get isFPR):
1486         (Arg.prototype.get fpr):
1487         (Arg.prototype.get isReg):
1488         (Arg.prototype.get reg):
1489         (Arg.isValidImmForm):
1490         (Arg.isValidBitImmForm):
1491         (Arg.isValidBitImm64Form):
1492         (Arg.isValidAddrForm):
1493         (Arg.isValidIndexForm):
1494         (Arg.prototype.isValidForm):
1495         (Arg.prototype.forEachTmpFast):
1496         (Arg.prototype.usesTmp):
1497         (Arg.prototype.forEachTmp):
1498         (Arg.prototype.is):
1499         (Arg.prototype.as):
1500         (Arg.prototype.forEachFast):
1501         (Arg.prototype.forEach):
1502         (Arg.extract):
1503         (Arg.forEachFast):
1504         (Arg.forEach):
1505         (Arg.prototype.get condition):
1506         (Arg.prototype.get isInvertible):
1507         (Arg.prototype.toString):
1508         * JSAir/basic_block.js: Added.
1509         (BasicBlock):
1510         (BasicBlock.prototype.get index):
1511         (BasicBlock.prototype.get size):
1512         (BasicBlock.prototype.Symbol.iterator):
1513         (BasicBlock.prototype.at):
1514         (BasicBlock.get last):
1515         (BasicBlock.get insts):
1516         (BasicBlock.get numSuccessors):
1517         (BasicBlock.get successors):
1518         (BasicBlock.get successorBlocks.):
1519         (BasicBlock.get successorBlocks):
1520         (BasicBlock.set get numPredecessors):
1521         (BasicBlock.get predecessors):
1522         (BasicBlock.get frequency):
1523         (BasicBlock.get headerString):
1524         (BasicBlock.get footerString):
1525         * JSAir/benchmark.js: Added.
1526         (benchmark):
1527         * JSAir/code.js: Added.
1528         (Code):
1529         (Code.prototype.addBlock):
1530         (Code.prototype.addStackSlot):
1531         (Code.prototype.newTmp):
1532         (Code.prototype.get size):
1533         (Code.prototype.at):
1534         (Code.prototype.Symbol.iterator):
1535         (Code.prototype.get blocks):
1536         (Code.prototype.get stackSlots):
1537         (Code.prototype.tmps):
1538         (Code.prototype.get callArgAreaSize):
1539         (Code.prototype.requestCallArgAreaSize):
1540         (Code.prototype.get frameSize):
1541         (Code.prototype.setFrameSize):
1542         (Code.prototype.toString):
1543         * JSAir/custom.js: Added.
1544         (const.ShuffleCustom.forEachArg):
1545         (const.ShuffleCustom.hasNonArgNonControlEffects):
1546         (const.PatchCustom.forEachArg):
1547         (const.PatchCustom.hasNonArgNonControlEffects):
1548         (const.CCallCustom.forEachArg):
1549         (const.CCallCustom.hasNonArgNonControlEffects):
1550         (const.ColdCCallCustom.forEachArg):
1551         (const.ColdCCallCustom.hasNonArgNonControlEffects):
1552         * JSAir/frequented_block.js: Added.
1553         (FrequentedBlock):
1554         (FrequentedBlock.prototype.toString):
1555         * JSAir/insertion_set.js: Added.
1556         (Insertion):
1557         (Insertion.prototype.get index):
1558         (Insertion.prototype.get element):
1559         (Insertion.prototype.lessThan):
1560         (InsertionSet):
1561         (InsertionSet.prototype.appendInsertion):
1562         (InsertionSet.prototype.append):
1563         (InsertionSet.prototype.execute):
1564         * JSAir/inst.js: Added.
1565         (Inst):
1566         (Inst.prototype.append):
1567         (Inst.prototype.clear):
1568         (Inst.prototype.get opcode):
1569         (Inst.prototype.get args):
1570         (Inst.prototype.visitArg):
1571         (Inst.prototype.forEachTmpFast):
1572         (Inst.prototype.forEachArg):
1573         (Inst.prototype.forEachTmp):
1574         (Inst.prototype.forEach):
1575         (Inst.forEachDef):
1576         (Inst.forEachDefWithExtraClobberedRegs):
1577         (Inst.prototype.get hasNonArgEffects):
1578         (Inst.prototype.toString):
1579         * JSAir/liveness.js: Added.
1580         (Liveness):
1581         (Liveness.prototype.get thing):
1582         (Liveness.prototype.get code):
1583         (Liveness.prototype.get liveAtHead):
1584         (Liveness.prototype.get liveAtTail):
1585         (Liveness.prototype.localCalc.LocalCalc):
1586         (Liveness.prototype.localCalc.LocalCalc.prototype.get liveSet):
1587         (Liveness.prototype.localCalc.LocalCalc.prototype.execute):
1588         (Liveness.prototype.localCalc):
1589         * JSAir/opcode.js: Added.
1590         (Inst_forEachArg):
1591         (Inst_hasNonArgEffects):
1592         * JSAir/payload-gbemu-executeIteration.js: Added.
1593         (createPayloadGbemuExecuteIteration):
1594         * JSAir/payload-imaging-gaussian-blur-gaussianBlur.js: Added.
1595         (createPayloadImagingGaussianBlurGaussianBlur):
1596         * JSAir/payload-jsair-ACLj8C.js: Added.
1597         (createPayloadJSAirACLj8C):
1598         * JSAir/payload-typescript-scanIdentifier.js: Added.
1599         (createPayloadTypescriptScanIdentifier):
1600         * JSAir/reg.js: Added.
1601         (Reg):
1602         (Reg.fromReg):
1603         (Reg.prototype.get index):
1604         (Reg.prototype.get type):
1605         (Reg.prototype.get name):
1606         (Reg.prototype.get isCalleeSave):
1607         (Reg.prototype.get isReg):
1608         (Reg.prototype.toString):
1609         (Reg.extract):
1610         (Reg.forEachFast):
1611         (Reg.forEach):
1612         (newGPR):
1613         (Reg.gprs.Reg.fprs.Reg.calleeSaveGPRs.Reg.calleeSaveFPRs.Reg.calleeSaves):
1614         * JSAir/stack_slot.js: Added.
1615         (StackSlot):
1616         (StackSlot.prototype.get byteSize):
1617         (StackSlot.prototype.get kind):
1618         (StackSlot.prototype.get isLocked):
1619         (StackSlot.prototype.get isSpill):
1620         (StackSlot.prototype.get index):
1621         (StackSlot.prototype.ensureSize):
1622         (StackSlot.prototype.get alignment):
1623         (StackSlot.prototype.get offsetFromFP):
1624         (StackSlot.prototype.setOffsetFromFP):
1625         (StackSlot.prototype.toString):
1626         (StackSlot.extract):
1627         (StackSlot.forEachFast):
1628         (StackSlot.forEach):
1629         * JSAir/symbols.js: Added.
1630         * JSAir/test.html: Added.
1631         * JSAir/test.js: Added.
1632         * JSAir/tmp.js: Added.
1633         (Tmp):
1634         (Tmp.fromReg):
1635         (Tmp.prototype.get index):
1636         (Tmp.prototype.get type):
1637         (Tmp.prototype.get isReg):
1638         (Tmp.prototype.toString):
1639         (Tmp.extract):
1640         (Tmp.forEachFast):
1641         (Tmp.forEach):
1642         * JSAir/tmp_base.js: Added.
1643         (TmpBase.prototype.get isGP):
1644         (TmpBase.prototype.get isFP):
1645         (TmpBase.prototype.get isGPR):
1646         (TmpBase.prototype.get isFPR):
1647         (TmpBase.prototype.get reg):
1648         (TmpBase.prototype.get gpr):
1649         (TmpBase.prototype.get fpr):
1650         (TmpBase):
1651         * JSAir/util.js: Added.
1652         (isRepresentableAsInt32):
1653         (addIndexed):
1654         (roundUpToMultipleOf):
1655         (symbolName):
1656         (mergeIntoSet):
1657         (nonEmptyRangesOverlap):
1658         (rangesOverlap):
1659         (removeAllMatching):
1660         (swap):
1661         (bubble):
1662         (bubbleSort):
1663         (currentTime):
1664         (else.currentTime):
1665
1666 2016-05-26  Yusuke Suzuki  <utatane.tea@gmail.com>
1667
1668         [JSC] Allow JSBench to use precise time
1669         https://bugs.webkit.org/show_bug.cgi?id=158050
1670
1671         Reviewed by Geoffrey Garen.
1672
1673         * JSBench/amazon-chrome-win/urem.js:
1674         (else.window.performance.window.performance.now.currentTimeInMS):
1675         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1676         (else.else.currentTimeInMS):
1677         (onload.cb):
1678         (onload):
1679         * JSBench/amazon-chrome/urem.js:
1680         (else.window.performance.window.performance.now.currentTimeInMS):
1681         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1682         (else.else.currentTimeInMS):
1683         (onload.cb):
1684         (onload):
1685         * JSBench/amazon-firefox-win/urm.js:
1686         (else.window.performance.window.performance.now.currentTimeInMS):
1687         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1688         (else.else.currentTimeInMS):
1689         (onload.cb):
1690         (onload):
1691         * JSBench/amazon-firefox/urm.js:
1692         (else.window.performance.window.performance.now.currentTimeInMS):
1693         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1694         (else.else.currentTimeInMS):
1695         (onload.cb):
1696         (onload):
1697         * JSBench/amazon-safari/urem.js:
1698         (else.window.performance.window.performance.now.currentTimeInMS):
1699         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1700         (else.else.currentTimeInMS):
1701         (onload.cb):
1702         (onload):
1703         * JSBench/facebook-chrome-win/urem.js:
1704         (else.window.performance.window.performance.now.currentTimeInMS):
1705         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1706         (else.else.currentTimeInMS):
1707         (onload.cb):
1708         (onload):
1709         * JSBench/facebook-chrome/urem.js:
1710         (else.window.performance.window.performance.now.currentTimeInMS):
1711         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1712         (else.else.currentTimeInMS):
1713         (onload.cb):
1714         (onload):
1715         * JSBench/facebook-firefox-win/urem.js:
1716         (else.window.performance.window.performance.now.currentTimeInMS):
1717         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1718         (else.else.currentTimeInMS):
1719         (onload.cb):
1720         (onload):
1721         * JSBench/facebook-firefox/urem.js:
1722         (else.window.performance.window.performance.now.currentTimeInMS):
1723         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1724         (else.else.currentTimeInMS):
1725         (onload.cb):
1726         (onload):
1727         * JSBench/facebook-safari/urem.js:
1728         (else.window.performance.window.performance.now.currentTimeInMS):
1729         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1730         (else.else.currentTimeInMS):
1731         (onload.cb):
1732         (onload):
1733         * JSBench/google-chrome-win/urem.js:
1734         (else.window.performance.window.performance.now.currentTimeInMS):
1735         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1736         (else.else.currentTimeInMS):
1737         (onload.cb):
1738         (onload):
1739         * JSBench/google-chrome/urem.js:
1740         (else.window.performance.window.performance.now.currentTimeInMS):
1741         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1742         (else.else.currentTimeInMS):
1743         (onload.cb):
1744         (onload):
1745         * JSBench/google-firefox-win/urem.js:
1746         (else.window.performance.window.performance.now.currentTimeInMS):
1747         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1748         (else.else.currentTimeInMS):
1749         (onload.cb):
1750         (onload):
1751         * JSBench/google-firefox/uem.js:
1752         (else.window.performance.window.performance.now.currentTimeInMS):
1753         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1754         (else.else.currentTimeInMS):
1755         (onload.cb):
1756         (onload):
1757         * JSBench/google-safari/urem.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/harness.js:
1764         (runBenchmark.window.currentTimeInMS):
1765         (runBenchmark.else.window.currentTimeInMS):
1766         * JSBench/twitter-chrome-win/rem.js:
1767         (else.window.performance.window.performance.now.currentTimeInMS):
1768         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1769         (else.else.currentTimeInMS):
1770         (onload.cb):
1771         (onload):
1772         * JSBench/twitter-chrome/urem.js:
1773         (else.window.performance.window.performance.now.currentTimeInMS):
1774         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1775         (else.else.currentTimeInMS):
1776         (onload.cb):
1777         (onload):
1778         * JSBench/twitter-firefox-win/urem.js:
1779         (else.window.performance.window.performance.now.currentTimeInMS):
1780         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1781         (else.else.currentTimeInMS):
1782         (onload.cb):
1783         (onload):
1784         * JSBench/twitter-firefox/urem.js:
1785         (else.window.performance.window.performance.now.currentTimeInMS):
1786         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1787         (else.else.currentTimeInMS):
1788         (onload.cb):
1789         (onload):
1790         * JSBench/twitter-safari/urem.js:
1791         (else.window.performance.window.performance.now.currentTimeInMS):
1792         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1793         (else.else.currentTimeInMS):
1794         (onload.cb):
1795         (onload):
1796         * JSBench/yahoo-chrome-win/urem.js:
1797         (else.window.performance.window.performance.now.currentTimeInMS):
1798         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1799         (else.else.currentTimeInMS):
1800         (onload.cb):
1801         (onload):
1802         * JSBench/yahoo-chrome/urem.js:
1803         (else.window.performance.window.performance.now.currentTimeInMS):
1804         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1805         (else.else.currentTimeInMS):
1806         (onload.cb):
1807         (onload):
1808         * JSBench/yahoo-firefox-win/urem.js:
1809         (else.window.performance.window.performance.now.currentTimeInMS):
1810         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1811         (else.else.currentTimeInMS):
1812         (onload.cb):
1813         (onload):
1814         * JSBench/yahoo-firefox/urem.js:
1815         (else.window.performance.window.performance.now.currentTimeInMS):
1816         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1817         (else.else.currentTimeInMS):
1818         (onload.cb):
1819         (onload):
1820         * JSBench/yahoo-safari/urem.js:
1821         (else.window.performance.window.performance.now.currentTimeInMS):
1822         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1823         (else.else.currentTimeInMS):
1824         (onload.cb):
1825         (onload):
1826
1827 2016-05-25  Geoffrey Garen  <ggaren@apple.com>
1828
1829         replaceable own properties seem to ignore replacement after property caching
1830         https://bugs.webkit.org/show_bug.cgi?id=158091
1831
1832         Reviewed by Darin Adler.
1833
1834         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
1835         * MallocBench/MallocBench/Benchmark.cpp:
1836         * MallocBench/MallocBench/Interpreter.cpp:
1837         (Interpreter::doMallocOp):
1838         * MallocBench/MallocBench/Interpreter.h:
1839         * MallocBench/MallocBench/fastMallocLog.63316.ops: Added.
1840         * MallocBench/MallocBench/jetstream.cpp: Added.
1841         (benchmark_jetstream):
1842         * MallocBench/MallocBench/jetstream.h: Added.
1843
1844 2016-05-25  Keith Miller  <keith_miller@apple.com>
1845
1846         Unreviewed, add JSBench to the skipped list for now since it doesn't
1847         work currently.
1848
1849         * Skipped:
1850
1851 2016-05-17  Keith Miller  <keith_miller@apple.com>
1852         We should have JSBench in PerformanceTests
1853         https://bugs.webkit.org/show_bug.cgi?id=157952
1854
1855         Rubber-stamped by Saam Barati.
1856
1857         There are some slight changes to the layout of the test directory
1858         to make it work nicely with run-jsc-benchmarks. Before JSBench
1859         had each of the browser specific sub-tests in a sub-directory.
1860         These have been flattened e.g. amazon/safari/ has become
1861         amazon-safari/.
1862
1863         * JSBench/amazon-chrome-win/urem.html: Added.
1864         * JSBench/amazon-chrome-win/urem.js: Added.
1865         * JSBench/amazon-chrome/urem.html: Added.
1866         * JSBench/amazon-chrome/urem.js: Added.
1867         * JSBench/amazon-firefox-win/urm.html: Added.
1868         * JSBench/amazon-firefox-win/urm.js: Added.
1869         * JSBench/amazon-firefox/urm.html: Added.
1870         * JSBench/amazon-firefox/urm.js: Added.
1871         * JSBench/amazon-safari/urem.html: Added.
1872         * JSBench/amazon-safari/urem.js: Added.
1873         * JSBench/browsercheck.js: Added.
1874         * JSBench/facebook-chrome-win/urem.html: Added.
1875         * JSBench/facebook-chrome-win/urem.js: Added.
1876         * JSBench/facebook-chrome/urem.html: Added.
1877         * JSBench/facebook-chrome/urem.js: Added.
1878         * JSBench/facebook-firefox-win/urem.html: Added.
1879         * JSBench/facebook-firefox-win/urem.js: Added.
1880         * JSBench/facebook-firefox/urem.html: Added.
1881         * JSBench/facebook-firefox/urem.js: Added.
1882         * JSBench/facebook-safari/urem.html: Added.
1883         * JSBench/facebook-safari/urem.js: Added.
1884         * JSBench/google-chrome-win/urem.html: Added.
1885         * JSBench/google-chrome-win/urem.js: Added.
1886         * JSBench/google-chrome/urem.html: Added.
1887         * JSBench/google-chrome/urem.js: Added.
1888         * JSBench/google-firefox-win/urem.html: Added.
1889         * JSBench/google-firefox-win/urem.js: Added.
1890         * JSBench/google-firefox/uem.html: Added.
1891         * JSBench/google-firefox/uem.js: Added.
1892         * JSBench/google-safari/urem.html: Added.
1893         * JSBench/google-safari/urem.js: Added.
1894         * JSBench/harness.html: Added.
1895         * JSBench/harness.js: Added.
1896         * JSBench/harness.py: Added.
1897         * JSBench/index.html: Added.
1898         * JSBench/reload.html: Added.
1899         * JSBench/twitter-chrome-win/rem.html: Added.
1900         * JSBench/twitter-chrome-win/rem.js: Added.
1901         * JSBench/twitter-chrome/urem.html: Added.
1902         * JSBench/twitter-chrome/urem.js: Added.
1903         * JSBench/twitter-firefox-win/urem.html: Added.
1904         * JSBench/twitter-firefox-win/urem.js: Added.
1905         * JSBench/twitter-firefox/urem.html: Added.
1906         * JSBench/twitter-firefox/urem.js: Added.
1907         * JSBench/twitter-safari/urem.html: Added.
1908         * JSBench/twitter-safari/urem.js: Added.
1909         * JSBench/yahoo-chrome-win/urem.html: Added.
1910         * JSBench/yahoo-chrome-win/urem.js: Added.
1911         * JSBench/yahoo-chrome/urem.html: Added.
1912         * JSBench/yahoo-chrome/urem.js: Added.
1913         * JSBench/yahoo-firefox-win/urem.html: Added.
1914         * JSBench/yahoo-firefox-win/urem.js: Added.
1915         * JSBench/yahoo-firefox/urem.html: Added.
1916         * JSBench/yahoo-firefox/urem.js: Added.
1917         * JSBench/yahoo-safari/urem.html: Added.
1918         * JSBench/yahoo-safari/urem.js: Added.
1919
1920 2016-05-20  Ryosuke Niwa  <rniwa@webkit.org>
1921
1922         run-benchmark's results should contain Animometer's debug output
1923         https://bugs.webkit.org/show_bug.cgi?id=157941
1924
1925         Reviewed by Stephanie Lewis.
1926
1927         Made developer.html support the JSON generated by run-benchmark which stores everything under debugOutput.
1928
1929         * Animometer/resources/debug-runner/animometer.js:
1930         (Utilities.initialize): Unwrap debugOutput in the case run-benchmark's result JSON is used.
1931
1932 2016-05-18  Timothy Hatcher  <timothy@apple.com>
1933
1934         Make Animometer work in all browsers
1935         https://bugs.webkit.org/show_bug.cgi?id=157855
1936         rdar://problem/26338521
1937
1938         Reviewed by Jon Lee.
1939
1940         * Animometer/resources/runner/logo.svg: Use a symbol instead
1941         of putting the id on the root svg. This works in all browsers.
1942
1943 2016-05-07  Timothy Hatcher  <timothy@apple.com>
1944
1945         Update Animometer design
1946         https://bugs.webkit.org/show_bug.cgi?id=157449
1947
1948         Reviewed by Darin Adler.
1949
1950         * Animometer/developer.html: Stop including runner/animometer.css.
1951         * Animometer/index.html: Udpated elements to fit new styles.
1952
1953         * Animometer/resources/debug-runner/animometer.css:
1954         Copy the old runner/animometer.css to here so the debug runner is unaffected.
1955
1956         * Animometer/resources/runner/animometer.css:
1957         New styles for the design. Drop flex box for broad browser support.
1958
1959         * Animometer/resources/runner/animometer.js:
1960         (ResultsTable.clear): Use textContent.
1961         (window.sectionsManager.showSection): Add class to the body.
1962         (window.sectionsManager.setSectionScore): Use textContent, not innerHTML.
1963
1964         * Animometer/resources/runner/crystal.svg: Added.
1965         * Animometer/resources/runner/lines.svg: Added.
1966         * Animometer/resources/runner/logo.svg: Added.
1967
1968         * Animometer/resources/runner/tests.js: Shorten test name and title case them.
1969
1970 2016-05-06  Manuel Rego Casasnovas  <rego@igalia.com>
1971
1972         [css-grid] Unprefix CSS Grid Layout properties
1973         https://bugs.webkit.org/show_bug.cgi?id=157137
1974
1975         Reviewed by Simon Fraser.
1976
1977         Remove "-webkit" prefix from all the grid layout properties,
1978         including the display value.
1979         Update the source code to remove the prefix where it was used too.
1980
1981         * Layout/auto-grid-lots-of-data.html:
1982         * Layout/fixed-grid-lots-of-data.html:
1983         * Layout/fixed-grid-lots-of-stretched-data.html:
1984
1985 2016-04-25  Geoffrey Garen  <ggaren@apple.com>
1986
1987         bmalloc: Misc improvements to MallocBench
1988         https://bugs.webkit.org/show_bug.cgi?id=157004
1989
1990         Reviewed by Darin Adler.
1991
1992         * MallocBench/run-malloc-benchmarks: Added --memory and --memory_warning
1993         modes for focused memory testing.
1994
1995         * MallocBench/MallocBench/Benchmark.cpp:
1996         (Benchmark::printReport): Clarified output.
1997
1998         (Benchmark::currentMemoryBytes): Added compressed memory because top
1999         does the same. (It always happens to zero in the benchmarks we run. But
2000         this is good for sanity.)
2001
2002         * MallocBench/MallocBench/CommandLine.cpp: Moved up to 8 runs to reduce
2003         variance.
2004
2005         * MallocBench/MallocBench/alloc_free.cpp:
2006         (benchmark_alloc_free): Cycle a single allocation in order to stress
2007         the effect of merging on calls to madvise.
2008
2009         * MallocBench/MallocBench/big.cpp:
2010         (benchmark_big): Graduated to 8kB-128kB because medium tests up to 8 and
2011         our large allocator doesn't kick in until 64kB.
2012
2013         * MallocBench/MallocBench/medium.cpp:
2014         (benchmark_medium): Test all the way down to 1kB because our large
2015         allocator used to service 1kB allocations and 1kB is an interesting
2016         middle size where memory is unusually large but allocation throughput
2017         still matters.
2018
2019         * MallocBench/MallocBench/stress.cpp:
2020         (benchmark_stress): Reduced the churn count to match stress_aligned
2021         because this test was taking too long to complete.
2022
2023         * MallocBench/MallocBench/stress_aligned.cpp:
2024         (benchmark_stress_aligned): Our new large allocator can handle even
2025         more absurdly large values.
2026
2027 2016-04-25  Simon Fraser  <simon.fraser@apple.com>
2028
2029         Add a content animation test that uses SVG animation.
2030         https://bugs.webkit.org/show_bug.cgi?id=156827
2031
2032         * Animation/svg-animation.html: Added.
2033         * Skipped:
2034
2035 2016-04-22  Simon Fraser  <simon.fraser@apple.com>
2036
2037         Skip two content animation tests which are only meant for iOS testing.
2038
2039         * Animation/css-animation.html: Added.
2040         * Animation/raf-animation.html: Added.
2041
2042         * Skipped:
2043
2044 2016-04-20  Simon Fraser  <simon.fraser@apple.com>
2045
2046         Add content animation tests to benchmark_runner, and allow the runner to collect device data as part of the results
2047         https://bugs.webkit.org/show_bug.cgi?id=156827
2048
2049         Add two files missing from the previous commit.
2050
2051         * Animation/css-animation.html: Added.
2052         * Animation/raf-animation.html: Added.
2053
2054 2016-04-20  Simon Fraser  <simon.fraser@apple.com>
2055
2056         Add content animation tests to benchmark_runner, and allow the runner to collect device data as part of the results
2057         https://bugs.webkit.org/show_bug.cgi?id=156827
2058
2059         Reviewed by Stephanie Lewis.
2060
2061         Two new files for the content-animation suite.
2062
2063         After patching for use in the benchmark, these tests are a little different to other benchmarks.
2064         The test content itself does not collect performance data; the tests spit out a boilerplate JSON
2065         result with a placeholder for framerate, which is collected in native code.
2066
2067         When run as a benchmark, test completion involves a couple of bounces between the test and MobileSafari.
2068         Test completion is initiated by a setTimeout() in the test, at which point location.hash is set to "#done".
2069         MobileSafari detects that, and asynchronously dumps performance data. Once this is complete, MobileSafari
2070         changes the URL has to "#submit", which triggers the XHR to the test relay that indicates the test is
2071         complete.
2072
2073         * Animation/css-animation.html: Added.
2074         * Animation/raf-animation.html: Added.
2075
2076 2016-04-15  Jon Lee  <jonlee@apple.com>
2077
2078         Animometer test could report a NaN
2079         https://bugs.webkit.org/show_bug.cgi?id=156646
2080
2081         Reviewed by Darin Adler.
2082         Provisionally reviewed by Said Abou-Hallawa.
2083
2084         * Animometer/tests/resources/main.js:
2085         (didFinishInterval): The ramp controller has a first phase where it ramps up the complexity and
2086         reacts based on how well the system handles the load. The assumption was that it would handle at least
2087         1 particle easily. That is not always the case. As a result, an interpolation calculation could
2088         end up setting an upper bound of NaN.
2089
2090         This occurs because we never get out of the first tier, so this._lastTierComplexity is undefined.
2091         Now that we guarantee a minimum complexity of 1, modify the conditional to check for this._lastTierComplexity
2092         before interpolating the upper bound of the first ramp. In the case where the system struggles with
2093         1 particle, set it to currentComplexity.
2094
2095 2016-04-09  Jon Lee  <jonlee@apple.com>
2096
2097         Update Animometer to accommodate different screens
2098         https://bugs.webkit.org/show_bug.cgi?id=156449
2099
2100         Reviewed by Darin Adler.
2101         Provisionally reviewed by Said Abou-Hallawa.
2102
2103         * Animometer/index.html: Wrap button in a container to add padding at the bottom.
2104         * Animometer/resources/debug-runner/animometer.css:
2105         (@media screen and (min-device-width: 1800px)): Deleted.
2106         (@media screen and (min-width: 1800px)): Cannot use min-device-width since it may match incorrectly.
2107         (screen and (max-device-height: 414px) and (orientation: landscape)): Some devices swap device width
2108         and height with orientation change.
2109         * Animometer/resources/runner/animometer.css: Similar.
2110         (screen and (min-device-width: 1024px) and (orientation: landscape)):
2111         (screen and (max-device-height: 414px) and (orientation: landscape)):
2112         (.frame-container): On smaller iPhones, adding 1px prevents the navigation bars from appearing.
2113         (@media screen and (min-device-width: 768px) and (max-device-width: 1024px)): Deleted.
2114         (@media (min-device-height: 768px) and (max-device-height: 1024px)): Target iPad Airs and similar.
2115         (@media screen and (min-device-width: 1024px) and (max-device-width: 1366px)): Deleted.
2116         (@media screen and (max-device-width: 1024px) and (min-device-height: 1366px)): Target iPad Pro.
2117         (#results footer): Add padding below the button for testing again.
2118         * Animometer/tests/master/multiply.html: Remove the center text.
2119         * Animometer/tests/master/resources/text.js: Update the test so that in every frame the text moves.
2120         * Animometer/tests/master/text.html: Update the text sizing depending on the size of the device.
2121
2122 2016-04-08  Jon Lee  <jonlee@apple.com>
2123
2124         Have Animometer benchmark always start with complexity of 1
2125         https://bugs.webkit.org/show_bug.cgi?id=156432
2126
2127         Reviewed by Ryosuke Niwa.
2128
2129         * Animometer/tests/resources/main.js: Update the default Controller and RampController to
2130         set its minimum complexities to 1 instead of 0.
2131
2132 2016-04-08  Jon Lee  <jonlee@apple.com>
2133
2134         Fix SVG benchmark test
2135         https://bugs.webkit.org/show_bug.cgi?id=156410
2136
2137         Reviewed by Dean Jackson.
2138
2139         * Animometer/resources/extensions.js: Update Point.zero to be a static Point.
2140         * Animometer/tests/simple/resources/tiled-canvas-image.js:
2141         (Stage.call._setupTiles): Refactor.
2142         * Animometer/tests/master/resources/particles.js:
2143         (Particle.prototype.reset): Use Point.center.
2144         (complexity): We are not using a gradient background anymore, so remove the +1.
2145         * Animometer/tests/master/resources/svg-particles.js: Update to use SVG transform
2146         instead of CSS transform.
2147
2148 2016-04-06  Jon Lee  <jonlee@apple.com>
2149
2150         Update master benchmark with SVG test
2151         https://bugs.webkit.org/show_bug.cgi?id=156273
2152
2153         Reviewed by Dean Jackson.
2154         Provisionally reviewed by Said Abou-Hallawa.
2155
2156         Switch masks tests for SVG path test.
2157
2158         * Animometer/resources/debug-runner/tests.js: Move mask test here.
2159         * Animometer/resources/runner/tests.js: Add SVG test here.
2160         * Animometer/tests/dom/particles.html: Renamed from PerformanceTests/Animometer/tests/master/particles.html.
2161         * Animometer/tests/dom/resources/dom-particles.js: Renamed from PerformanceTests/Animometer/tests/master/resources/dom-particles.js.
2162
2163         * Animometer/tests/master/resources/particles.js: Add minPosition for bounds checking. Prevents particle from being
2164         partially obscured.
2165         * Animometer/tests/master/resources/svg-particles.js: Added.
2166         (Particle): The particle is either a path object or a rect using a path as a clip. The
2167         same path is used either way. For each particle create a linear gradient with a random
2168         rotation.
2169         (SVGParticlesStage): Look in #shapes to see how many different kinds of paths are available.
2170         This makes the test more generic in case other shapes need to be tested.
2171         * Animometer/tests/master/svg-particles.html: Added. Have two defs, one that houses each
2172         particle's gradient, and one that holds the shape templates.
2173
2174 2016-03-27  Hunseop Jeong  <hs85.jeong@samsung.com>
2175
2176         [EFL] REGRESSION(r188793): It made 200 layout tests and Bindings/event-target-wrapper.html performance test fail
2177         https://bugs.webkit.org/show_bug.cgi?id=148470 
2178
2179         Reviewed by Darin Adler.
2180
2181         * Skipped: Unskip the Bindings/event-target-wrapper test.
2182
2183 2016-03-23  Jon Lee  <jonlee@apple.com>
2184
2185         Update focus and leaves tests
2186         https://bugs.webkit.org/show_bug.cgi?id=155825
2187
2188         Reviewed by Simon Fraser.
2189
2190         Particles would improperly disappear, before they were fully off stage. Update the calculation.
2191         * Animometer/tests/dom/resources/leaves.js:
2192         * Animometer/tests/master/resources/leaves.js:
2193
2194         * Animometer/tests/master/focus.html: Update the center element styling.
2195         * Animometer/tests/master/resources/focus.js:
2196         (FocusElement.Utilities.createClass): Give enough space in the container for each particle for the blur
2197         to spread. It should be about 3x more space than the blur radius on each side.
2198         (Stage.call.initialize): Ditto for the center element.
2199
2200 2016-03-23  Jon Lee  <jonlee@apple.com>
2201
2202         Update support for other platforms
2203         https://bugs.webkit.org/show_bug.cgi?id=155824
2204
2205         Reviewed by Simon Fraser.
2206
2207         * Animometer/resources/debug-runner/animometer.css: Remove extraneous media query. All of it should
2208         be handled in the release stylesheet.
2209         * Animometer/resources/debug-runner/animometer.js: Add support for browsers that return an HTMLCollection
2210         when retrieving radio elements with a name, instead of the selected value.
2211         * Animometer/resources/runner/animometer.css: Adjust the queries for devices of varying widths.
2212         (.frame-container): Move overflow:hidden to #stage.
2213         * Animometer/tests/master/resources/stage.css: Consolidate into tests/resources/stage.css.
2214         * Animometer/tests/resources/stage.css:
2215         (body): Remove overflow:hidden.
2216         (#stage): Set overflow:hidden here.
2217         (#center-text): Moved from master/resources/stage.css.
2218
2219         Update stylesheet hrefs.
2220         * Animometer/tests/master/canvas-stage.html:
2221         * Animometer/tests/master/focus.html:
2222         * Animometer/tests/master/leaves.html:
2223         * Animometer/tests/master/multiply.html:
2224         * Animometer/tests/master/particles.html:
2225         * Animometer/tests/master/text.html: Adjust text size to fit in the device screen.
2226
2227 2016-03-23  Jon Lee  <jonlee@apple.com>
2228
2229         Make the benchmark require device to be in landscape orientation
2230         https://bugs.webkit.org/show_bug.cgi?id=155822
2231         rdar://problem/25258650
2232
2233         Reviewed by Ryosuke Niwa.
2234
2235         Check that the device is in landscape orientation. If not, disable the button to start it,
2236         and show a warning message.
2237
2238         * Animometer/developer.html: Add a message about how the browser should be set up. Include a
2239         warning message that appears if the orientation is incorrect on mobile devices.
2240         * Animometer/index.html: Ditto.
2241
2242         * Animometer/resources/debug-runner/animometer.css: Migrate the .hidden rule into the release
2243         stylesheet. Update the style to accommodate the new UI.
2244         * Animometer/resources/runner/animometer.css:
2245
2246         * Animometer/resources/runner/animometer.js:
2247         (window.benchmarkController.initialize): Add the orientation listener is needed.
2248         (window.benchmarkController.addOrientationListenerIfNecessary): Only mobile devices need this. Check
2249         to see for support of window.orientation.
2250         (window.benchmarkController._orientationChanged): Toggle the warning based on whether we match the
2251         landscape query. Set a state variable, which is needed for the debug harness. Call updateStartButtonState.
2252         (window.benchmarkController.updateStartButtonState):
2253         * Animometer/resources/debug-runner/animometer.js: Change _updateStartButtonState to return
2254         a boolean about whether at least one test is selected. That will be used in
2255         benchmarkController.updateStartButtonState(). Move the load event listener to the release version.
2256         (window.benchmarkController.updateStartButtonState): Override the release version, and also check
2257         that at least one test is selected.
2258
2259 2016-03-22  Geoffrey Garen  <ggaren@apple.com>
2260
2261         MallocBench: consolidate regression testing for aligned allocation
2262         https://bugs.webkit.org/show_bug.cgi?id=155762
2263
2264         Reviewed by Michael Saboff.
2265
2266         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
2267         * MallocBench/MallocBench/Benchmark.cpp:
2268         * MallocBench/MallocBench/memalign.cpp: Removed.
2269         * MallocBench/MallocBench/memalign.h: Removed. The stress_aligned test
2270         covers this and much more.
2271
2272         * MallocBench/MallocBench/stress_aligned.cpp:
2273         (benchmark_stress_aligned): Include specific tests for extreme sizes
2274         and alignments.
2275
2276 2016-03-21  Jon Lee  <jonlee@apple.com>
2277
2278         Update benchmark tests
2279         https://bugs.webkit.org/show_bug.cgi?id=155723
2280
2281         Reviewed by Darin Adler.
2282         Provisionally reviewed by Said Abou-Hallawa.
2283
2284         Add quadratic and bezier segments to the canvas path test.
2285
2286         * Animometer/resources/runner/tests.js: Some of the query strings are unnecessary.
2287         Rename the test.
2288         * Animometer/tests/master/resources/canvas-stage.js:
2289         (tune): Update to be able to take an array of possible constructors. Choose one randomly.
2290         * Animometer/tests/master/resources/canvas-tests.js:
2291         (CanvasLinePoint.Utilities.createClass): Move the point selection out to a separate
2292         function called randomPoint() for reuse. Move X_LOOPS, Y_LOOPS, and offsets into the
2293         class definition.
2294         (randomPoint): Scale the grid down a little bit so that the lines along the edge of the
2295         canvas are not cut off when the stroke size is thick.
2296         (CanvasQuadraticSegment): Added.
2297         (CanvasBezierSegment): Added.
2298         (SimpleCanvasStage): Pass in an array of the different segment types. Since line segments
2299         are short compared to the curved ones, make it twice as likely to render a line segment.
2300         (SimpleCanvasStage.animate): Update the drawing code so that we render all line segments.
2301
2302         Add a helper method that selects a random element from an array.
2303
2304         * Animometer/tests/resources/main.js:
2305         (Stage.randomElementInArray): Select a random element from the provided array.
2306
2307         * Animometer/tests/bouncing-particles/resources/bouncing-tagged-images.js: Refactor.
2308         * Animometer/tests/dom/resources/leaves.js: Ditto.
2309         * Animometer/tests/master/resources/dom-particles.js: Ditto.
2310         * Animometer/tests/master/resources/image-data.js: Ditto.
2311         * Animometer/tests/master/resources/leaves.js: Ditto.
2312         * Animometer/tests/simple/resources/simple-canvas-paths.js: Ditto.
2313
2314         Add canvas tests that includes all stroke and fill paths. This makes it possible to avoid having to
2315         include the full simple canvas suite for perf testing.
2316
2317         * Animometer/resources/debug-runner/tests.js: Add new tests. Move the canvas test into the Canvas
2318         suite.
2319         * Animometer/tests/simple/resources/simple-canvas-paths.js: Add a CanvasStroke and CanvasFill particle
2320         that random selects an object to render.
2321
2322         * Animometer/resources/debug-runner/tests.js: Move 3D suite before basic canvas suite.
2323         When updating the perf bot script, we will include the suites up to this one, but exclude
2324         the basic canvas suite.
2325
2326         Move compositing transforms test to HTML suite and remove the empty Miscellaneous suite.
2327
2328         * Animometer/resources/debug-runner/tests.js:
2329         * Animometer/tests/dom/compositing-transforms.html: Renamed from PerformanceTests/Animometer/tests/misc/compositing-transforms.html.
2330         * Animometer/tests/dom/resources/compositing-transforms.js: Renamed from PerformanceTests/Animometer/tests/misc/resources/compositing-transforms.js.
2331
2332         Clean up miscellaneous test suite. Add a canvas ellipse test, and remove the other
2333         canvas tests.
2334
2335         * Animometer/resources/debug-runner/tests.js: Add ellipse tests to the simple suite.
2336         * Animometer/tests/misc/canvas-electrons.html: Removed.
2337         * Animometer/tests/misc/canvas-stars.html: Removed.
2338         * Animometer/tests/misc/resources/canvas-electrons.js: Removed.
2339         * Animometer/tests/misc/resources/canvas-stars.js: Removed.
2340         * Animometer/tests/simple/resources/simple-canvas-paths.js: Add ellipse primitives.
2341
2342         Merge text tests together into one.
2343
2344         * Animometer/resources/runner/tests.js: Remove international.html.
2345         * Animometer/tests/master/international.html: Removed.
2346         * Animometer/tests/master/resources/text.js:
2347         (animate): Update styling. Manually calculate gradients for the shadow particles.
2348         Reduce the step size for y direction to avoid cutting text off at the margins. Increase
2349         step for x since there will be a little more room.
2350         * Animometer/tests/master/text.html: Add more translations and lay it out in a table.
2351
2352 2016-03-20  Jon Lee  <jonlee@apple.com>
2353
2354         Add a link to show debug data
2355         https://bugs.webkit.org/show_bug.cgi?id=155724
2356
2357         Reviewed by Simon Fraser.
2358
2359         * Animometer/developer.html: Attach the onclick handler to the score.
2360         * Animometer/index.html: Ditto.
2361
2362 2016-03-18  Dean Jackson  <dino@apple.com>
2363
2364         Add a basic WebGL test to Animometer
2365         https://bugs.webkit.org/show_bug.cgi?id=155475
2366         <rdar://problem/25156860>
2367
2368         Reviewed by Simon Fraser and Jon Lee.
2369
2370         Add a pretty simple WebGL test to Animometer
2371         which tries to draw a number of animated triangles.
2372         The result still needs a bit of tuning.
2373
2374         * Animometer/resources/debug-runner/tests.js: Add a new
2375         Suite for "3D" tests.
2376         * Animometer/tests/3d/resources/webgl.js: Added.
2377         * Animometer/tests/3d/webgl.html: Added.
2378
2379 2016-03-18  Jon Lee  <jonlee@apple.com>
2380
2381         Add support for statically linking to a specific test
2382         https://bugs.webkit.org/show_bug.cgi?id=155631
2383
2384         Rubber-stamped by Darin Adler.
2385
2386         * Animometer/developer.html: Update to call restartBenchmark instead so that it works for
2387         both the preset and the manually set test sessions.
2388         * Animometer/resources/debug-runner/animometer.css:
2389         (.tree .link): Style the "link" UI.
2390         * Animometer/resources/debug-runner/animometer.js: Move updateDisplay from suitesManager
2391         to optionsManager since this is a visual option and has nothing to do with the suites
2392         settings. Update the construction of the test UI by including a "link" after each test.
2393         Clicking on that link brings up a JS prompt with a URL and a query string with the current
2394         parameters of the controller, and selected test. Pasting this URL into the location bar will
2395         automatically start running the selected test.
2396         (suitesManager.suitesFromQueryString): Iterate through the Suites and tests and find the one
2397         that matches the provided parameters. Returns an object similar to the form in
2398         suitesManager.updateLocalStorageFromUI.
2399         (benchmarkController.initialize): After settings up the events and options, try parsing the
2400         query string and running the benchmark immediately. Otherwise, fall back to the form.
2401         (benchmarkController.startBenchmark): Store the options and suites into member variables for
2402         reuse in restartBenchmark.
2403         (benchmarkController.startBenchmarkImmediatelyIfEncoded): Convert the query string to an object.
2404         If that's successful, find the suite and test referenced in the query string. Start the benchmark if
2405         the search for the test succeeded.
2406         * Animometer/resources/extensions.js:
2407         (Utilities.stripNonASCIICharacters): Helper method to convert the name of the suite and test
2408         into a query-string-friendly version.
2409         (Utilities.convertObjectToQueryString): Helper method to convert an object to query string
2410         format.
2411         (Utilities.convertQueryStringToObject): Helper method to convert query string into an object
2412         with properties and values.
2413
2414 2016-03-16  Jon Lee  <jonlee@apple.com>
2415
2416         Add a new benchmark test
2417         https://bugs.webkit.org/show_bug.cgi?id=155570
2418
2419         Reviewed by Simon Fraser.
2420
2421         New Leaves test includes various image sizes and opacity.
2422
2423         * Animometer/resources/debug-runner/tests.js: Add it to the HTML test suite.
2424         * Animometer/tests/dom/leaves.html: Added.
2425         * Animometer/tests/dom/resources/leaves.js: Added. Override the
2426         (Particle.call.reset): Uses a range of sizes, and opacity.
2427         (Particle.call.animate): Opacity goes up then down. When it hits 0, reset the particle.
2428         (Particle.call.move): Set transform and opacity.
2429         * Animometer/tests/master/resources/leaves.js: Get rid of the closure so that it
2430         can be used in this test. Update the relative path so that it works in both the master
2431         and dom test suite.
2432
2433 2016-03-15  Simon Fraser  <simon.fraser@apple.com>
2434
2435         Add developer Animometer test that bounces P3-tagged images
2436         https://bugs.webkit.org/show_bug.cgi?id=155511
2437
2438         Reviewed by Tim Horton.
2439         
2440         Add a test for rendering performance of tagged images. The 5 images are tagged
2441         with the Display P3 colorspace.
2442
2443         * Animometer/resources/debug-runner/tests.js:
2444         * Animometer/tests/bouncing-particles/bouncing-tagged-images.html: Added.
2445         * Animometer/tests/bouncing-particles/resources/bouncing-tagged-images.js: Added.
2446         * Animometer/tests/bouncing-particles/resources/image1.jpg: Added.
2447         * Animometer/tests/bouncing-particles/resources/image2.jpg: Added.
2448         * Animometer/tests/bouncing-particles/resources/image3.jpg: Added.
2449         * Animometer/tests/bouncing-particles/resources/image4.jpg: Added.
2450         * Animometer/tests/bouncing-particles/resources/image5.jpg: Added.
2451
2452 2016-03-09  Jon Lee  <jonlee@apple.com>
2453
2454         Enhance existing Animometer tests
2455         https://bugs.webkit.org/show_bug.cgi?id=155261
2456
2457         Reviewed by Simon Fraser.
2458
2459         * Animometer/tests/master/resources/canvas-tests.js: Add a gradient to
2460         the filled circles. To expose more of the gradient, add another circle.
2461         (SimpleCanvasStage.call.animate): For each frame, create a gradient with
2462         undulating stop points and colors. Fill the circles twice; once with
2463         the solid color, and once with the gradient.
2464
2465         * Animometer/tests/master/resources/dom-particles.js: Refactor the
2466         emission variables into a separate stage for this test. Add a colorOffset
2467         variable to make the colors of each particle slightly different, since
2468         the ramp controller can add large numbers of particles all at once, which
2469         would otherwise get all the same color.
2470         * Animometer/tests/master/resources/particles.js:
2471         (initialize): Remove the code specific to the SVG mask test.
2472
2473         * Animometer/tests/master/resources/image-data.js:
2474         (initialize): Remove unused local variable.
2475         * Animometer/tests/master/resources/multiply.js:
2476         (initialize): Make the test harder by adding more total particles.
2477
2478 2016-03-09  Jon Lee  <jonlee@apple.com>
2479
2480         Add text tests
2481         https://bugs.webkit.org/show_bug.cgi?id=155257
2482
2483         Reviewed by Simon Fraser.
2484
2485         * Animometer/resources/extensions.js:
2486         (UnitBezier.Utilities.createClass): Add a class that computes Bezier points
2487         assuming that two of the control points are at (0,0) and (1,1). Taken from
2488         WebCore/platform/graphics/UnitBezier.h
2489         * Animometer/resources/runner/tests.js: Add a test for English text and one
2490         for international text.
2491         * Animometer/tests/master/international.html: Added.
2492         * Animometer/tests/master/resources/text.js: Added. The test assumes there is
2493         #template div which it will copy. The copies are placed behind the template, and
2494         are set with different colors each frame. They are moved around with CSS transform.
2495         * Animometer/tests/master/text.html: Added.
2496
2497         Remove the other text tests, since these ones cover the same techniques.
2498         * Animometer/tests/text/layering-text.html: Removed.
2499         * Animometer/tests/text/resources/layering-text.js: Removed.
2500         * Animometer/tests/text/resources/text-boxes.js: Removed.
2501         * Animometer/tests/text/text-boxes.html: Removed.
2502         * Animometer/resources/debug-runner/tests.js:
2503
2504 2016-03-09  Jon Lee  <jonlee@apple.com>
2505
2506         Add a new image test
2507         https://bugs.webkit.org/show_bug.cgi?id=155232
2508
2509         Reviewed by Dean Jackson.
2510         Provisionally reviewed by Said Abou-Hallawa.
2511
2512         The image test renders PNGs and moves them with translate and rotate
2513         transforms. Each particle has a lifetime, and when the lifetime ends or
2514         the particle goes offscreen, it resets itself somewhere on the stage.
2515
2516         * Animometer/resources/debug-runner/tests.js: Remove the CSS bouncing PNG
2517         images test, because this one tests the same technique.
2518         * Animometer/resources/runner/tests.js: Added here as "Leaves".
2519         * Animometer/tests/master/leaves.html: Added.
2520         * Animometer/tests/master/resources/particles.js: Refactor out the parts
2521         specific to the DOM particles test. Consequently make velocity a public
2522         member.
2523         (initialize): The options parameter is never used, so remove it.
2524         (animate): Remove unused local variable.
2525         * Animometer/tests/master/resources/leaves.js: Maintains a focal point that
2526         moves back and forth across the canvas. That point affects the velocity
2527         of the particles. When the focal point is on the leftmost side, it is
2528         between [-6, -2], and set in reset(). When the focal point is on the rightmost
2529         side, it will be from [2, 6].
2530         * Animometer/tests/master/resources/dom-particles.js: Move JS specific to
2531         this test here from particles.js.
2532
2533         New images.
2534         * Animometer/tests/master/resources/compass100.png: Added.
2535         * Animometer/tests/master/resources/console100.png: Added.
2536         * Animometer/tests/master/resources/contribute100.png: Added.
2537         * Animometer/tests/master/resources/debugger100.png: Added.
2538         * Animometer/tests/master/resources/inspector100.png: Added.
2539         * Animometer/tests/master/resources/layout100.png: Added.
2540         * Animometer/tests/master/resources/performance100.png: Added.
2541         * Animometer/tests/master/resources/script100.png: Added.
2542         * Animometer/tests/master/resources/shortcuts100.png: Added.
2543         * Animometer/tests/master/resources/standards100.png: Added.
2544         * Animometer/tests/master/resources/storage100.png: Added.
2545         * Animometer/tests/master/resources/styles100.png: Added.
2546         * Animometer/tests/master/resources/timeline100.png: Added.
2547
2548 2016-03-03  Jon Lee  <jonlee@apple.com>
2549
2550         Add ability to retrieve raw data from release harness
2551         https://bugs.webkit.org/show_bug.cgi?id=155026
2552
2553         Reviewed by Simon Fraser.
2554
2555         * Animometer/developer.html: Remove the special UI in the debug
2556         harness.
2557         * Animometer/resources/runner/animometer.css: Add styles for the
2558         overlay.
2559         * Animometer/resources/runner/animometer.js: Let 'j' show the JSON
2560         results, but only if the overlay doesn't exist. Add 'esc' key to dismiss
2561         the overlay.
2562         (window.benchmarkController.selectResults): Cycle the cases around
2563         so that the first 's' press selects both the benchmark score and
2564         the individual test scores.
2565         * Animometer/resources/debug-runner/animometer.css: Remove
2566         unneeded rules.
2567         * Animometer/resources/debug-runner/animometer.js: Make the same
2568         call to handleKeyPress.
2569
2570 2016-03-03  Jon Lee  <jonlee@apple.com>
2571
2572         Make sure multiply test particles have at least some opacity
2573         https://bugs.webkit.org/show_bug.cgi?id=155027
2574
2575         Reviewed by Simon Fraser.
2576
2577         * Animometer/tests/master/resources/multiply.js:
2578         (initialize): Have the elements spin a little faster to make it
2579         more obvious when the system is being stressed.
2580         (_addTile):
2581         (animate): Make sure each element being animated has at least 1%
2582         opacity.
2583
2584 2016-03-03  Jon Lee  <jonlee@apple.com>
2585
2586         Update image test
2587         https://bugs.webkit.org/show_bug.cgi?id=154962
2588
2589         Rubber-stamped by Darin Adler.
2590
2591         * Animometer/tests/master/image-data.html: Make each canvas have a compositing layer.
2592         * Animometer/tests/master/resources/image-data.js:
2593         (initialize): Update to load multiple images serially.
2594         (_loadImage):
2595         (tune): Use display instead of visibility. When showing an element anew, move it somewhere else.
2596         (_createTestElement): Set up the element, refactor out placement of element to _refreshElement.
2597         (_refreshElement): Place element in a tile grid.
2598         (animate):
2599         (_getRandomNeighboringPixelIndex): Use a more conservative distribution to make the effect last a little longer.
2600         * Animometer/tests/master/resources/compass.svg: Added.
2601         * Animometer/tests/master/resources/console.svg: Added.
2602         * Animometer/tests/master/resources/contribute.svg: Added.
2603         * Animometer/tests/master/resources/debugger.svg: Added.
2604         * Animometer/tests/master/resources/inspector.svg: Added.
2605         * Animometer/tests/master/resources/layout.svg: Added.
2606         * Animometer/tests/master/resources/performance.svg: Added.
2607         * Animometer/tests/master/resources/script.svg: Added.
2608         * Animometer/tests/master/resources/shortcuts.svg: Added.
2609         * Animometer/tests/master/resources/standards.svg: Added.
2610         * Animometer/tests/master/resources/storage.svg: Added.
2611         * Animometer/tests/master/resources/styles.svg: Added.
2612         * Animometer/tests/master/resources/timeline.svg: Added.
2613
2614 2016-03-02  Jon Lee  <jonlee@apple.com>
2615
2616         Add some new controllers, and refine tests
2617         https://bugs.webkit.org/show_bug.cgi?id=154914
2618
2619         Reviewed by Simon Fraser.
2620
2621         Improve tests.
2622
2623         * Animometer/tests/master/focus.html: Move each particle into a layer. Put the particle
2624         in a container element. Expand the container element by the max blur radius, and clip
2625         overflow. This way, when the blur is applied, it avoids causing layer resizes.
2626         * Animometer/tests/master/resources/focus.js:
2627         (FocusElement.Utilities.createClass): Create a container element, and put the particle
2628         inside.
2629         (hide): Set display:none.
2630         (show): Set display:block.
2631         (animate): Apply filters to the container element.
2632         (FocusStage.call.initialize): Instead of inserting and removing elements from the DOM,
2633         keep them in the stage, but set the display style instead. Use this._offsetIndex to
2634         keep track of which elements are displayed.
2635         (FocusStage.call.animate): Move some calculations around to avoid doing unneeded math.
2636         (FocusStage.call.getBlurValue): Make sure elements are always blurred.
2637         (FocusStage.call.getOpacityValue): Make sure elements have some opacity.
2638         * Animometer/tests/master/resources/multiply.js: Have the particles on the edge of the
2639         stage appear less black when the complexity gets large enough.
2640         (tune): Have this._distanceFactor calculate the factor. Avoid calculating square root each
2641         frame.
2642
2643 2016-03-02  Jon Lee  <jonlee@apple.com>
2644
2645         Add some new controllers, and refine tests
2646         https://bugs.webkit.org/show_bug.cgi?id=154914
2647
2648         Reviewed by Simon Fraser.
2649
2650         Add a controller that centers around 30 fps instead of 60 fps.
2651
2652         * Animometer/developer.html: Add a new option.
2653         * Animometer/resources/debug-runner/animometer.js:
2654         * Animometer/resources/runner/animometer.js:
2655         (this._processData.findRegression): When calculating the complexity-frameLength regression,
2656         check the controller, and use a 30 fps baseline if needed.
2657         * Animometer/resources/statistics.js:
2658         (Regression.Utilities.createClass): Update to allow clients to specify the baseline
2659         frame length.
2660         * Animometer/tests/resources/main.js:
2661         (tune): Override some of the constants in RampController. Move those constants out for
2662         easier reading.
2663
2664         Add a fixed controller, with no step.
2665
2666         * Animometer/developer.html: Add a controller that takes no step.
2667         * Animometer/resources/debug-runner/animometer.js:
2668         * Animometer/tests/resources/main.js:
2669         (Rotater.Utilities.createClass):
2670
2671         Switch to ramp controller as default.
2672
2673         * Animometer/developer.html: Increase the test length to 20 seconds.
2674         * Animometer/resources/debug-runner/animometer.js: Rename the "adjustment" field to "controller"
2675         since that is a more accurate description.
2676         * Animometer/resources/debug-runner/graph.js:
2677         * Animometer/resources/runner/animometer.js: Update preferences for release suite.
2678
2679 2016-02-26  Jon Lee  <jonlee@apple.com>
2680
2681         Address Dean's comments in 154673.
2682
2683         * Animometer/developer.html:
2684         * Animometer/resources/debug-runner/animometer.js:
2685         * Animometer/resources/debug-runner/graph.js:
2686
2687 2016-02-25  Jon Lee  <jonlee@apple.com>
2688
2689         Update animation benchmark and tests
2690         https://bugs.webkit.org/show_bug.cgi?id=154673
2691
2692         Reviewed by Dean Jackson.
2693
2694         Update test visuals.
2695
2696         * Animometer/tests/master/resources/canvas-tests.js:
2697         (CanvasLineSegment.Utilities.createClass): Line segments near the edge of the stage get clipped.
2698         Update circle position and radius to minimize impact.
2699         * Animometer/tests/master/resources/particles.js: Get rid of rotating gradient background, and
2700         have 3 locations for emitting particles.
2701         * Animometer/tests/master/resources/multiply.js: Update distance metric so that fringe tiles get
2702         more color.
2703         (initialize): Fix some of the math for laying out the tiles.
2704         (animate): When a tile isn't used, set visibility: hidden.
2705
2706 2016-02-24  Jon Lee  <jonlee@apple.com>
2707
2708         Update animation benchmark and tests
2709         https://bugs.webkit.org/show_bug.cgi?id=154673
2710
2711         Reviewed by Dean Jackson.
2712
2713         Update the ramp controller.
2714
2715         The controller refines the complexity interval to test across.
2716
2717         * Animometer/resources/statistics.js: Add functions that estimate cumulative distribution function.
2718         (Regression): For the flat regression, force the first segment to be at 60 fps.
2719         (valueAt): Add convenience function to return interpolated value based on the regression used.
2720         (_calculateRegression): Include the number of points included for both segments, and the piecewise
2721         errors.
2722         * Animometer/tests/resources/math.js: Make the Kalman estimator subclass Experiment, and allow it
2723         to be reset.
2724
2725         * Animometer/tests/resources/main.js: Initialize the tier such that it starts at 10^0 = 1.
2726         Increase the number of ramps. Maintain three FPS thresholds-- the frame rate of interest, a limit
2727         on the lowest FPS we care to go for later interpolation, and a minimum FPS threshold we want to
2728         aim for each ramp. Also keep three estimators: a running average of the change point, a minimum
2729         boundary for each ramp, and an estimator for all the frames within an interval. The first two
2730         are used to determine the parameters of the next ramp, and the latter allows us to refine the
2731         parameters.
2732         (update): During the tier phase, it is possible that the highest complexity possible for a test
2733         won't stress the system enough to trigger stopping the tier phase and transitioning to the ramps.
2734         If the complexity doesn't change when going to the next tier, we've maxed the test out, and move
2735         on. When the tier phase completed, turn off Controller.frameLengthEstimator, which estimates the
2736         FPS at each tier.
2737         (tune): At each interval, look at the confidence distribution of being on the 60 FPS side or the
2738         slow side. If the slowest FPS we achieve at the ramp's maximum complexity is not at least
2739         _fpsRampSlowThreshold, then increase the maximum complexity. If we ever achieve 60 FPS, increase
2740         the ramp's minimum complexity to that level. If, at an even lower complexity, a glitch causes the
2741         FPS to drop, we reset the minimum complexity.
2742
2743         Have the bootstrap calculation occur between tests. Clean up harness.
2744
2745         * Animometer/resources/debug-runner/animometer.js: Run bootstrap after a test has
2746         completed to avoid doing all of it at the end before showing the results. Clean up
2747         parameters being passed around.
2748         * Animometer/resources/debug-runner/tests.js:
2749         (text):
2750         * Animometer/resources/runner/animometer.js:
2751         (this._processData.calculateScore): Save the results to the same object holding the data.
2752         (this._processData._processData): In the case where a file is dragged, calculate the score
2753         serially. Grab the results object and move it to the results variable and remove it from
2754         the data object. This avoids serializing the results into the JSON.
2755         (this._processData.findRegression): Include the samples used for bootstrapping. Reduce the
2756         resample size to shorten the wait.
2757         * Animometer/resources/runner/benchmark-runner.js:
2758         * Animometer/resources/statistics.js:
2759         (bootstrap): Update how bootstrapData is sorted. In some regression results the mix of
2760         floats and integers causes an alphabetical sort to occur.
2761         * Animometer/resources/strings.js:
2762
2763         Add meta charset so that encodings between harness and test match.
2764
2765         * Animometer/tests/bouncing-particles/bouncing-canvas-images.html:
2766         * Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html:
2767         * Animometer/tests/bouncing-particles/bouncing-css-images.html:
2768         * Animometer/tests/bouncing-particles/bouncing-css-shapes.html:
2769         * Animometer/tests/bouncing-particles/bouncing-svg-images.html:
2770         * Animometer/tests/bouncing-particles/bouncing-svg-shapes.html:
2771         * Animometer/tests/master/canvas-stage.html:
2772         * Animometer/tests/master/focus.html:
2773         * Animometer/tests/master/image-data.html:
2774         * Animometer/tests/master/multiply.html:
2775         * Animometer/tests/master/particles.html:
2776         * Animometer/tests/misc/canvas-electrons.html:
2777         * Animometer/tests/misc/canvas-stars.html:
2778         * Animometer/tests/misc/compositing-transforms.html:
2779         * Animometer/tests/simple/simple-canvas-paths.html:
2780         * Animometer/tests/simple/tiled-canvas-image.html:
2781         * Animometer/tests/template/template-canvas.html:
2782         * Animometer/tests/template/template-css.html:
2783         * Animometer/tests/template/template-svg.html:
2784         * Animometer/tests/text/layering-text.html:
2785         * Animometer/tests/text/text-boxes.html:
2786
2787         Update test harness reporting.
2788
2789         * Animometer/developer.html: Add missing meta charset.
2790         * Animometer/index.html: Remove unnecessary utf-8 declaration.
2791         * Animometer/resources/debug-runner/animometer.css: Add convenience classes for
2792         formatting the results table.
2793         * Animometer/resources/debug-runner/animometer.js: Adjust which stats are shown.
2794         * Animometer/resources/debug-runner/tests.js: Display bootstrapping statistics.
2795         * Animometer/resources/strings.js: Move strings not used by the release harness.
2796
2797         Switch to a pseudo-random number generator.
2798
2799         * Animometer/resources/statistics.js: Add a Pseudo class, with a simple
2800         pseudo-random number generator.
2801         (_calculateRegression): Reset the generator before running bootstrap.
2802         (bootstrap): Deleted.
2803
2804         Replace Math.random with Pseudo.random.
2805         * Animometer/tests/master/resources/canvas-tests.js:
2806         * Animometer/tests/master/resources/focus.js:
2807         * Animometer/tests/master/resources/particles.js:
2808         * Animometer/tests/resources/main.js:
2809
2810         Use bootstrapping to get confidence interval in the breakpoint.
2811
2812         For the ramp controller, calculate the piecewise regression, and then use
2813         bootstrapping in order to find the 95% confidence interval. Use the raw data.
2814
2815         * Animometer/developer.html: Default to the complexity graph. Add a legend
2816         checkbox to toggle visibility of the bootstrap score and histogram.
2817         * Animometer/resources/debug-runner/animometer.css: Make some more space to show
2818         the old raw and average scores in the legend. Add new styles for the data.
2819         * Animometer/resources/debug-runner/graph.js:
2820         (_addRegressionLine): Allow passing an array for the variance bar tied to the
2821         regression line. Now |stdev| is |range|.
2822         (createComplexityGraph): Add bootstrap median, and overlay a histogram of
2823         the bootstrap samples. Switch raw samples from circles to X's.
2824         (onComplexityGraphOptionsChanged): Allow toggling of the bootstrap data.
2825         (onGraphTypeChanged): Move the regressions for the raw and average samples to the
2826         legend. In the subtitle use the bootstrap median, and include the 95% confidence
2827         interval.
2828         * Animometer/resources/runner/animometer.js:
2829         (this._processData.findRegression): Factor out the code that determines which
2830         samples to include when calculating the piecewise regression. For series that have
2831         many samples, or a wider range of recorded complexities, throw away the 2.5%
2832         lowest and highest samples before calculating the regression. Keep all samples
2833         if the number of samples to regress is small or the range of complexities is
2834         narrow.
2835         (this._processData._calculateScore): Factor out regression calculation to
2836         findRegression(). Bootstrap the change point of the regression. The score is the
2837         median.
2838         * Animometer/resources/statistics.js:
2839         (_calculateRegression): Correct an issue in the calculation of the regression, where
2840         the denominator can be 0.
2841         (bootstrap): Template for bootstrapping. Create a bootstrap sample array, Create
2842         re-samples by random selection with replacement. Return the 95% confidence samples,
2843         the bootstrap median, mean, and the data itself.
2844         * Animometer/resources/strings.js: Add bootstrap.
2845         * Animometer/tests/resources/main.js:
2846         (processSamples): Don't prematurely cut the sample data.
2847
2848         Fix graph drawing.
2849
2850         * Animometer/resources/debug-runner/animometer.js: Add spacing in the JSON output.
2851         Multiple tests output a lot of JSON and can hang when selecting JSON with no whitespace.
2852         * Animometer/resources/debug-runner/animometer.css:
2853         (#complexity-graph .series.raw circle): Update the color.
2854         * Animometer/resources/debug-runner/graph.js: Use the FPS axis instead of the
2855         complexity axis, which can vary in domain. For determining the complexity domain,
2856         only use samples after samplingTimeOffset.
2857
2858         Allow dropping results JSON.
2859
2860         * Animometer/developer.html: Add a button.
2861         * Animometer/resources/debug-runner/animometer.css:
2862         * Animometer/resources/debug-runner/animometer.js: Read the data and go straight
2863         to the dashboard. With JSON output, write out only the options and the raw data.
2864
2865         Teach the harness to evaluate the samples and determine the test score.
2866
2867         This will allow us to update how the score is calculated separately from the samples recorded.
2868         This also prepares the harness to be able to accept JSON of prior runs.
2869
2870         * Animometer/resources/strings.js: Clean up and remove unneeded strings and reduce some of the
2871         hierarchy.
2872         * Animometer/resources/debug-runner/tests.js: Update to use the new strings.
2873
2874         * Animometer/tests/resources/main.js: Allow all controllers to show a complexity-FPS graph.
2875         (_processComplexitySamples): Factor out some of the sample processing done in the ramp
2876         controller for the benefit of the other controllers. |complexitySamples| contains a list of
2877         samples. Sort the samples by complexity. Optionally remove the top x% of samples.
2878         Group them, and calculate distribution of samples within the same complexity, and add those as
2879         new entries into |complexityAverageSamples|.
2880         (Controller.processSamples): Move the code responsible for determining the complexity and FPS
2881         scores out to ResultsDashboard. The structure of the data returned by the controller is:
2882
2883         {
2884             controller: [time-regression, time-regression, ...], // optional, data specific to controller
2885             marks: [...],
2886             samples: {                    // all of the sample data
2887                 controller: [...],
2888                 complexity: [...],        // processed from controller samples
2889                 complexityAverage: [...], // processed from complexity samples
2890             }
2891         }
2892
2893         (AdaptiveController.processSamples): Adding the target frame length is no longer necessary; we
2894         now pass the test options to the graph.
2895         (Regression): Move to statistics.js.
2896         * Animometer/resources/statistics.js: Move Regression to here. Add a check if the sampling range
2897         only contains one sample, since we cannot calculate a regression from one sample point.
2898
2899         Teach the test harness to evaluate the data.
2900         * Animometer/resources/runner/animometer.js:
2901         (ResultsDashboard): Store the options used to run the test and the computed results/score separately
2902         from the data. The results are stored as:
2903
2904         {
2905             score: /* geomean of iteration score */,
2906             iterationsResults: [
2907                 {
2908                     score: /* geomean of tests */,
2909                     testsResults: {
2910                         suiteName: {
2911                             testName: {
2912                                 controller: {
2913                                     average:
2914                                     concern:
2915                                     stdev:
2916                                     percent:
2917                                 },
2918                                 frameLength: { ... },
2919                                 complexity: {
2920                                     complexity:
2921                                     stdev:
2922                                     segment1:
2923                                     segment2:
2924                                 },
2925                                 complexityAverage: { ... }
2926                             },
2927                             testName: { ... },
2928                         },
2929                         ... next suite ...
2930                     }
2931                 },
2932                 { ...next iteration... }
2933             ]
2934         }
2935
2936         * Animometer/resources/debug-runner/animometer.js: Pass options around instead of relying
2937         on what was selected in the form. This will later allow for dropping previous results, and
2938         using those runs' options when calculating scores.
2939         (ResultsTable._addGraphButton): Simplify button action by using attached test data.
2940         * Animometer/resources/debug-runner/graph.js: Refactor to use the data.
2941
2942         Consolidate JS files, and move statistics out to a separate JS.
2943
2944         Preparation for having the Controller only handle recording and storage of the samples,
2945         and leave the evaluation of the test score out to the harness. Move Experiment to
2946         a new statistics.js, where Regression will also eventually go. Get rid of algorithm.js
2947         and move it to utilities.js since the Heap is used only for Experiments.
2948
2949         * Animometer/tests/resources/algorithm.js: Removed. Heap is in utilities.js.
2950         * Animometer/tests/resources/sampler.js: Removed. Experiment is in statistics.js,
2951         Sampler in main.js.
2952         * Animometer/tests/resources/main.js: Move Sampler here.
2953         * Animometer/resources/statistics.js: Added. Move Statistics and Experiment here.
2954         * Animometer/resources/extensions.js: Move Heap here. Attach static method to create
2955         a max or min heap to Heap, instead of a new Algorithm object.
2956
2957         Update JS files.
2958         * Animometer/developer.html:
2959         * Animometer/index.html:
2960         * Animometer/tests/bouncing-particles/bouncing-canvas-images.html:
2961         * Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html:
2962         * Animometer/tests/bouncing-particles/bouncing-css-images.html:
2963         * Animometer/tests/bouncing-particles/bouncing-css-shapes.html:
2964         * Animometer/tests/bouncing-particles/bouncing-svg-images.html:
2965         * Animometer/tests/bouncing-particles/bouncing-svg-shapes.html:
2966         * Animometer/tests/master/canvas-stage.html:
2967         * Animometer/tests/master/focus.html:
2968         * Animometer/tests/master/image-data.html:
2969         * Animometer/tests/master/multiply.html:
2970         * Animometer/tests/master/particles.html:
2971         * Animometer/tests/misc/canvas-electrons.html:
2972         * Animometer/tests/misc/canvas-stars.html:
2973         * Animometer/tests/misc/compositing-transforms.html:
2974         * Animometer/tests/simple/simple-canvas-paths.html:
2975         * Animometer/tests/simple/tiled-canvas-image.html:
2976         * Animometer/tests/template/template-canvas.html:
2977         * Animometer/tests/template/template-css.html:
2978         * Animometer/tests/template/template-svg.html:
2979         * Animometer/tests/text/layering-text.html:
2980         * Animometer/tests/text/text-boxes.html:
2981
2982         Fix the cursor in the graph analysis when the min
2983         complexity is not 0.
2984
2985         * Animometer/resources/debug-runner/graph.js:
2986         (_addRegression):
2987         (createComplexityGraph):
2988
2989 2016-02-23  Geoffrey Garen  <ggaren@apple.com>
2990
2991         Fix some issues in MallocBench
2992         https://bugs.webkit.org/show_bug.cgi?id=154600
2993
2994         Reviewed by Oliver Hunt.
2995
2996         * MallocBench/MallocBench.xcodeproj/project.pbxproj: Make nimlang.ops
2997         a part of the build so that it copies to the right place for execution.
2998
2999         * MallocBench/MallocBench/Interpreter.cpp:
3000         (Interpreter::Interpreter): Don't try to open .ops files for writing
3001         because we're only going to read and we might need extra permissions to
3002         write.
3003
3004         (Interpreter::~Interpreter): Give more context when opening a file fails
3005         to help with debugging.
3006
3007         * MallocBench/MallocBench/stress.cpp:
3008         (benchmark_stress): Reduce iterations to complete in less than 10 seconds.
3009
3010 2016-02-23  Michael Saboff  <msaboff@apple.com>
3011
3012         Unreviewed change to revert extraneous changes made part of change set 196955.
3013
3014         * MallocBench/MallocBench/Interpreter.cpp:
3015         (Interpreter::doMallocOp):
3016         (Interpreter::Thread::switchTo):
3017         (writeData): Deleted.
3018
3019 2016-02-19  Michael Saboff  <msaboff@apple.com>
3020
3021         MallocBench: Added recording for nimlang website, new recording details and added new options
3022         https://bugs.webkit.org/show_bug.cgi?id=154485
3023
3024         Reviewed by Geoff Garen.
3025
3026         Added new capabilities to MallocBench.  These include:
3027             Added a recording of http://nim-lang.org/docs/lib.html.
3028             Added thread id to the recording and the ability to playback switching threads in MallocBench
3029             Added aligned allocations to recordings and the ability to playback
3030             Added --use-thread-id option to honor recorded thread ids
3031             Added --detailed-report to output remaining allocations by size after playback
3032             Added --no-warmup to not run the warm up iteration
3033
3034         Changed the way that options are passed down to the benchmarks.  Instead of passing individual
3035         boolean or numeric option values, just pass a reference the CommandLine itself.  Each benchmark
3036         can access the options that are appropriate.  The Benchmark class also uses the options for
3037         is parallel, run counts and warm up.
3038
3039         Added thread id and aligned malloc to the Op by noticing that structure padding and Opcode allowed
3040         for another 32 bits of data.  Breaking that unused 32 bits into a 16 bit thread id value and a
3041         16 bit log base 2 of the alignment for aligned malloc allowed for existing recordings to playback
3042         without any incompatibilities.
3043
3044         Threaded operation is simulated by creating threads as needed.  As long as the next Op's thread id
3045         is the same as the last, operation continues as normal.  When the next Op has a different thread id,
3046         we switch to that thread using the shared Op stream to continue playing back.  There is a mutex to
3047         assure that only one thread is really running at a time and a condition variable used to wait
3048         that the current thread id matches each block thread's thread id.  This doesn't simulate true
3049         concurrent threading, but is instead plays back Ops recorded for multiple thread faithfully.
3050
3051         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
3052         * MallocBench/MallocBench/Benchmark.cpp:
3053         (deallocateHeap):
3054         (Benchmark::Benchmark):
3055         (Benchmark::runOnce):
3056         (Benchmark::run):
3057         * MallocBench/MallocBench/Benchmark.h:
3058         (Benchmark::isValid):
3059         * MallocBench/MallocBench/CommandLine.cpp:
3060         (CommandLine::printUsage):
3061         * MallocBench/MallocBench/CommandLine.h:
3062         (CommandLine::isValid):
3063         (CommandLine::benchmarkName):
3064         (CommandLine::isParallel):
3065         (CommandLine::useThreadID):
3066         (CommandLine::detailedReport):
3067         (CommandLine::warmUp):
3068         (CommandLine::heapSize):
3069         (CommandLine::runs):
3070         * MallocBench/MallocBench/Interpreter.cpp:
3071         (Interpreter::Interpreter):
3072         (Interpreter::run):
3073         (Interpreter::readOps):
3074         (Interpreter::doOnSameThread):
3075         (Interpreter::switchToThread):
3076         (Interpreter::detailedReport):
3077         (compute2toPower):
3078         (writeData):
3079         (Interpreter::doMallocOp):
3080         (Interpreter::Thread::Thread):
3081         (Interpreter::Thread::stop):
3082         (Interpreter::Thread::~Thread):
3083         (Interpreter::Thread::runThread):
3084         (Interpreter::Thread::waitToRun):
3085         (Interpreter::Thread::switchTo):
3086         * MallocBench/MallocBench/Interpreter.h:
3087         (Interpreter::Thread::isMainThread):
3088         * MallocBench/MallocBench/alloc_free.cpp: Added.
3089         (benchmark_alloc_free):
3090         * MallocBench/MallocBench/alloc_free.h: Added.
3091         * MallocBench/MallocBench/balloon.cpp:
3092         (benchmark_balloon):
3093         * MallocBench/MallocBench/balloon.h:
3094         * MallocBench/MallocBench/big.cpp:
3095         (benchmark_big):
3096         * MallocBench/MallocBench/big.h:
3097         * MallocBench/MallocBench/churn.cpp:
3098         (benchmark_churn):
3099         * MallocBench/MallocBench/churn.h:
3100         * MallocBench/MallocBench/facebook.cpp:
3101         (benchmark_facebook):
3102         * MallocBench/MallocBench/facebook.h:
3103         * MallocBench/MallocBench/flickr.cpp:
3104         (benchmark_flickr):
3105         (benchmark_flickr_memory_warning):
3106         * MallocBench/MallocBench/flickr.h:
3107         * MallocBench/MallocBench/fragment.cpp:
3108         (validate):
3109         (benchmark_fragment):
3110         (benchmark_fragment_iterate):
3111         * MallocBench/MallocBench/fragment.h:
3112         * MallocBench/MallocBench/list.cpp:
3113         (benchmark_list_allocate):
3114         (benchmark_list_traverse):
3115         * MallocBench/MallocBench/list.h:
3116         * MallocBench/MallocBench/main.cpp:
3117         (main):
3118         * MallocBench/MallocBench/medium.cpp:
3119         (benchmark_medium):
3120         * MallocBench/MallocBench/medium.h:
3121         * MallocBench/MallocBench/memalign.cpp:
3122         (test):
3123         (benchmark_memalign):
3124         * MallocBench/MallocBench/memalign.h:
3125         * MallocBench/MallocBench/message.cpp:
3126         (benchmark_message_one):
3127         (benchmark_message_many):
3128         * MallocBench/MallocBench/message.h:
3129         * MallocBench/MallocBench/nimlang.cpp: Added.
3130         (benchmark_nimlang):
3131         * MallocBench/MallocBench/nimlang.h: Added.
3132         * MallocBench/MallocBench/nimlang.ops: Added.
3133         * MallocBench/MallocBench/realloc.cpp:
3134         (benchmark_realloc):
3135         * MallocBench/MallocBench/realloc.h:
3136         * MallocBench/MallocBench/reddit.cpp:
3137         (benchmark_reddit):
3138         (benchmark_reddit_memory_warning):
3139         * MallocBench/MallocBench/reddit.h:
3140         * MallocBench/MallocBench/stress.cpp:
3141         (deallocate):
3142         (benchmark_stress):
3143         * MallocBench/MallocBench/stress.h:
3144         * MallocBench/MallocBench/stress_aligned.cpp:
3145         (benchmark_stress_aligned):
3146         * MallocBench/MallocBench/stress_aligned.h:
3147         * MallocBench/MallocBench/theverge.cpp:
3148         (benchmark_theverge):
3149         (benchmark_theverge_memory_warning):
3150         * MallocBench/MallocBench/theverge.h:
3151         * MallocBench/MallocBench/tree.cpp:
3152         (benchmark_tree_allocate):
3153         (benchmark_tree_traverse):
3154         (benchmark_tree_churn):
3155         * MallocBench/MallocBench/tree.h:
3156         * MallocBench/run-malloc-benchmarks:
3157
3158 2016-02-11  Jon Lee  <jonlee@apple.com>
3159
3160         Fix a missing refactoring.
3161
3162         * Animometer/tests/master/resources/multiply.js:
3163         (animate): Move to Utilities.lerp.
3164         (_lerp): Deleted.
3165
3166 2016-02-10  Jon Lee  <jonlee@apple.com>
3167
3168         Add new benchmark tests.
3169         https://bugs.webkit.org/show_bug.cgi?id=154063
3170
3171         Provisionally reviewed by Said Abou-Hallawa.
3172
3173         Add tests for get/put image data, filters, opacity, and css transforms.
3174
3175         * Animometer/resources/runner/benchmark-runner.js:
3176         (_runBenchmarkAndRecordResults): Update the body background color to match that of
3177         the stage.
3178         (this._runNextIteration): Clear the background color style for the results page.
3179         * Animometer/resources/runner/tests.js:
3180         * Animometer/tests/master/focus.html: Added.
3181         * Animometer/tests/master/image-data.html: Added.
3182         * Animometer/tests/master/multiply.html: Added.
3183         * Animometer/tests/master/resources/focus.js: Added.
3184         * Animometer/tests/master/resources/image-data.js: Added.
3185         * Animometer/tests/master/resources/multiply.js: Added.
3186         * Animometer/tests/master/resources/stage.css: Move common styles out.
3187         * Animometer/tests/resources/main.js: Update Stage.randomBool to use Math.random.
3188         Add Stage.randomSign for randomly setting a direction. Add the notion of the
3189         current timestamp of the test to Benchmark, since some animations cycle through
3190         colors and rely on an incremental counter like the time.
3191
3192 2016-02-09  Said Abou-Hallawa  <sabouhallawa@apple.com>
3193
3194         Add internal benchmark tests for CSS mix-blend-modes and filters
3195         https://bugs.webkit.org/show_bug.cgi?id=154058
3196
3197         Provisionally reviewed by Jon Lee.
3198
3199         * Animometer/resources/debug-runner/tests.js: Include the new tests in the
3200         "HTML suite" of the debug runner.
3201
3202         * Animometer/resources/extensions.js:
3203         (Utilities.browserPrefix):
3204         (Utilities.setElementPrefixedProperty): Utility functions to allow setting
3205         prefixed style properties.
3206
3207         * Animometer/tests/bouncing-particles/resources/bouncing-css-shapes.js:
3208         Set the mix-blend-mode and the filter to some random values if the options
3209         of the test requested that.
3210
3211         * Animometer/tests/bouncing-particles/resources/bouncing-particles.js:
3212         (parseShapeParameters): Parse the url options "blend" and "filter" and set
3213         the corresponding flags.
3214
3215         * Animometer/tests/resources/main.js:
3216         (randomStyleMixBlendMode):
3217         (randomStyleFilter): Return random mix-blend-mode and filter.
3218
3219 2016-02-08  Jon Lee  <jonlee@apple.com>
3220
3221         Add a ramp controller
3222         https://bugs.webkit.org/show_bug.cgi?id=154028
3223
3224         Provisionally reviewed by Said Abou-Hallawa.
3225
3226         Enhance the graph to include a complexity-fps graph, in addition
3227         to the time graph.
3228
3229         * Animometer/developer.html: Add a ramp option.
3230         * Animometer/resources/debug-runner/animometer.css: Update the style.
3231         * Animometer/resources/strings.js: Flatten the Strings.text constants.
3232         * Animometer/resources/debug-runner/animometer.js:
3233         (ResultsTable.call._addGraphButton): Refactor.
3234         (ResultsTable.call._addTest): Add regression data.
3235         (benchmarkController): Add a form that allows the user to switch between the two forms,
3236         Add a form that allows the user to toggle different data. Hide certain header columns
3237         depending on the selected controller.
3238         * Animometer/resources/debug-runner/graph.js: Add the complexity regressions.
3239         * Animometer/resources/debug-runner/tests.js: Add headers for the ramp results.
3240         * Animometer/resources/runner/animometer.js:
3241         (ResultsTable): If a header is disabled don't include them in _flattenedHeaders.
3242         * Animometer/tests/resources/main.js:
3243         (Controller): Allow options to specify the capacity for sample arrays.
3244         (Regression): A piecewise regression that tries to fit a slope and a flat profile.
3245         (_calculateRegression): Options can fix the slope and bias when calculating the minimal
3246         error. Sweep across the samples in time (which could be backward depending on the controller)
3247         and calculate the intersection point.
3248         (RampController): This controller assumes that the target frame rate is below
3249         58 FPS. It runs in two stages. The first stage quickly determines the order of
3250         magnitude of objects needed to stress the system by the setting the complexity
3251         to increasingly difficult tiers. Perform a series of ramps descending from a
3252         high-water mark of complexity. The complexity needed to reach the target frame
3253         length is done by performing a piecewise regression on each ramp, and track a
3254         running average of these values. For the next ramp, make that running average
3255         the center of the ramp. With a minimum complexity of 0, the high-water mark is
3256         twice that average. The score is based on the highest complexity that can
3257         reach 60 fps.
3258
3259 2016-02-08  Jon Lee  <jonlee@apple.com>
3260
3261         Address Said's comments on the benchmark, and do some clean up.
3262
3263         * Animometer/developer.html:
3264         * Animometer/resources/debug-runner/animometer.css: Add styles for averages.
3265         * Animometer/resources/debug-runner/animometer.js: Use the right
3266         Strings constants.
3267         * Animometer/resources/debug-runner/graph.js:
3268         (_addRegressionLine): Add missing code to draw the line and standard
3269         deviation highlight.
3270         (onGraphTypeChanged): Remove unneeded variables
3271         (onTimeGraphOptionsChanged):
3272         * Animometer/resources/runner/benchmark-runner.js:
3273         (_runBenchmarkAndRecordResults): Rename samplers to suiteResults and
3274         _suitesSamplers to _suitesResults.
3275         * Animometer/tests/resources/main.js:
3276         (results): Call processSamples().
3277         (update): Change sampling timestamp comparison.
3278         (_animateLoop): Move shouldStop call to before the update.
3279         * Animometer/tests/resources/sampler.js:
3280         (process): Rename to processSamples().
3281
3282 2016-02-07  Jon Lee  <jonlee@apple.com>
3283
3284         Teach Controller to measure intervals, and turn off the frame length estimator.
3285
3286         * Animometer/tests/resources/main.js: Default interval length is 100 ms.
3287         (start): Set the first interval.
3288         (_measureAndResetInterval): Reports the average frame length of the interval that just
3289         completed, and sets up the next interval.
3290         (update): If there is no length, then just use the estimator per frame, otherwise the
3291         estimator measures per interval. Add a didFinishInterval for subclasses to process
3292         prior to recording the sample. Update tune() to include whether an interval had
3293         finished.
3294         (StepController): Step controllers don't measure on an interval basis.
3295
3296 2016-02-07  Jon Lee  <jonlee@apple.com>
3297
3298         Minor refactoring. Rename Controller._estimator to Controller._frameLengthEstimator
3299         and switch the parameters for start(), update(), and tune(), so that the timestamp
3300         is first and stage is second.
3301
3302         * Animometer/tests/resources/main.js:
3303
3304 2016-02-07  Jon Lee  <jonlee@apple.com>
3305
3306         Move ResultsTable functionality not needed for release tests out.
3307         Move reporting of score and mean to selection of the time-based graph.
3308
3309         * Animometer/developer.html: Rename graph-options to time-graph-options.
3310         * Animometer/resources/debug-runner/animometer.js:
3311         (DeveloperResultsTable): Moved from runner/animometer.js. Switch from mean
3312         values to "average" objects which can hold stdev. Move graph button and
3313         calculation of noisy measurements here. Sophisticated header processing
3314         is not needed in release suite.
3315         (populateTable): Use DeveloperResultsTable.
3316         * Animometer/resources/debug-runner/graph.js: Pull time graph creation to
3317         its own function, and add a new onGraphTypeChanged handler in preparation
3318         of a complexity graph to be added later.
3319         * Animometer/resources/runner/animometer.js:
3320         (ResultsTable): Simplify to just handle test names and scores.
3321
3322 2016-02-07  Jon Lee  <jonlee@apple.com>
3323
3324         Tests: reuse objects already made.
3325
3326         Avoid thrash of object creation and removal by maintaining an index that
3327         moves along the array as the adjust values change. If the tune value
3328         requires more objects than the maximum size of the object array, then create
3329         new objects. This means that the object array size never decreases.
3330
3331         * Animometer/tests/master/resources/canvas-stage.js: Maintain a separate
3332         offsetIndex. For these tests, we want to avoid drawing the oldest objects,
3333         so the scene will draw the object at offsetIndex to the end of the array.
3334         (tune): Reverse the logic since "removal" of objects is much simpler and
3335         involves simply changing the offsetIndex.
3336         (animate): Update the for loop to draw from offsetIndex to the end.
3337         (complexity): Update the definition.
3338         * Animometer/tests/master/resources/canvas-tests.js: Maintain a separate
3339         offsetIndex. For these tests, we want to avoid drawing the newest objects,
3340         so the scene will draw the object at index 0 to the object at offsetIndex.
3341         (SimpleCanvasStage.animate): Fly-by removal of local stage variable,
3342         which is unneeded. Update the for loop to draw from offsetIndex to the end.
3343         * Animometer/tests/simple/resources/simple-canvas-paths.js:
3344         (SimpleCanvasStage.animate): Update the for loop to draw from 0 to
3345         offsetIndex.
3346         * Animometer/tests/simple/resources/simple-canvas.js:
3347         (tune): Update logic. Here, offsetIndex represents the boundary of the last
3348         index to render.
3349         (animate): Update the for loop to draw from 0 to offsetIndex.
3350         (complexity): Update the definition.
3351
3352 2016-02-07  Jon Lee  <jonlee@apple.com>
3353
3354         Tests: refactor and update styles.
3355
3356         * Animometer/tests/resources/main.js: Add helper methods that return
3357         a color that hue rotates based on the date, and a counter value that
3358         increases based on the date. Fix randomInt() to not bias against the min and
3359         max values.
3360
3361         * Animometer/tests/master/resources/canvas-tests.js: Use new helper methods.
3362         * Animometer/tests/master/resources/dom-particles.js: Ditto.
3363         * Animometer/tests/master/resources/particles.js: Ditto.
3364         * Animometer/tests/simple/resources/simple-canvas-paths.js: Refactor to
3365         use a rotating color instead of a random color. The fast switching of color
3366         is too vivid to watch.
3367
3368         * Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js:
3369         (BouncingSvgParticlesStage.call.createGradient): Fix the gradient so
3370         that the last stop is located at the end.
3371
3372 2016-02-07  Jon Lee  <jonlee@apple.com>
3373
3374         Refactor tune() to not return the complexity of the scene.
3375
3376         We have stage.complexity() now, so returning the complexity through tune
3377         is unnecessary.
3378
3379         * Animometer/tests/bouncing-particles/resources/bouncing-particles.js:
3380         * Animometer/tests/master/resources/canvas-stage.js:
3381         * Animometer/tests/master/resources/particles.js:
3382         * Animometer/tests/misc/resources/canvas-electrons.js:
3383         * Animometer/tests/misc/resources/canvas-stars.js:
3384         * Animometer/tests/resources/main.js:
3385         * Animometer/tests/simple/resources/simple-canvas.js:
3386         * Animometer/tests/simple/resources/tiled-canvas-image.js:
3387         * Animometer/tests/template/resources/template-canvas.js:
3388         * Animometer/tests/template/resources/template-css.js:
3389         * Animometer/tests/template/resources/template-svg.js:
3390         * Animometer/tests/text/resources/layering-text.js:
3391
3392 2016-02-07  Jon Lee  <jonlee@apple.com>
3393
3394         Make the fixed controller a step controller instead. Halfway through the test
3395         it will bump up the complexity 4-fold. Calculate the step timestamp using options
3396         instead of a separate parameter to the Controller constructor.
3397
3398         * Animometer/developer.html: Change value to "step"
3399         * Animometer/resources/debug-runner/animometer.js:
3400         (window.suitesManager.updateEditsElementsState): Show number inputs when set to "step".
3401         * Animometer/tests/resources/main.js:
3402         (update): Provide a hook for subclasses to tune.
3403         (StepController): Maintain a flag determining whether we've stepped, and the time
3404         we should step.
3405         (Benchmark): Use the new StepController.
3406
3407 2016-02-07  Jon Lee  <jonlee@apple.com>
3408
3409         Adjust the FPS graph scale.
3410
3411         Instead of making the FPS graph linearly scale, scale it based on the frame length,
3412         but show the data in terms of FPS. Because it is inversely proportional, and most
3413         of the data never gets below 20, concentrate the axis from 20-60 FPS, since otherwise
3414         over half of the available graph space ends up blank.
3415
3416         This means we should convert all of the FPS data to frame length data.
3417
3418         * Animometer/resources/debug-runner/graph.js: Update the domain to be based on
3419         frame length in milliseconds instead of FPS. Update the cursor to consider all of the
3420         values being shown, and then pick the min and max values to represent the length of the
3421         cursor.
3422         * Animometer/resources/runner/animometer.js:
3423         * Animometer/resources/strings.js:
3424         * Animometer/tests/resources/main.js:
3425         (processSamples): Add the ability to only sample a range of the data instead of everything
3426         after an offset index. Update sampler to record the frame lengths instead of the frame
3427         rate.
3428
3429 2016-02-07  Jon Lee  <jonlee@apple.com>
3430
3431         Add option to use different methods for retrieving a timestamp.
3432
3433         * Animometer/developer.html: Add performance.now and Date.now options.
3434         * Animometer/resources/runner/animometer.js: Default to performance.now.
3435         (window.benchmarkController.startBenchmark):
3436         * Animometer/tests/resources/main.js: Tie the desired method to _getTimestamp.
3437         (run): Use _getTimestamp.
3438         (_animateLoop): Ditto.
3439
3440 2016-02-07  Jon Lee  <jonlee@apple.com>
3441
3442         Allow adding any number of markers to the graph. The markers can be labeled
3443         and contain timestamp and sample index data. Make it a part of the controller
3444         rather than keeping it in the sampler.
3445
3446         * Animometer/resources/debug-runner/animometer.css: Add styles for markers
3447         * Animometer/resources/debug-runner/graph.js: Create the markers and add
3448         text labels.
3449         * Animometer/resources/runner/animometer.js: Assume the samplingTimeOffset
3450         is just one of the marks provided.
3451         * Animometer/resources/strings.js: Add Strings.json.marks.
3452         * Animometer/tests/resources/main.js:
3453         (Controller): Keep marks here. They are keyed by the marker name, so no two
3454         markers should have the same name.
3455         (recordFirstSample): Refactor to use mark.
3456         (mark): Allows for arbitrary data if needed later. The timestamp maintained
3457         is relative to the absolute start timestamp.
3458         (containsMark): Checks whether a mark with a specific comment exists.
3459         (processSamples): Removes the _startTimestamp offset from the marks before
3460         setting it in results.
3461         * Animometer/tests/resources/sampler.js: Remove marks.
3462
3463 2016-02-07  Jon Lee  <jonlee@apple.com>
3464
3465         Get rid of options member variable in Benchmark.
3466
3467         Options are only needed when initializing the stage or benchmark, so there's no
3468         need to also keep a reference to it.
3469
3470         * Animometer/tests/resources/main.js: Get rid of options variable in Benchmark.
3471         Pass options to Controllers and Stages.
3472         (Controller.Utilities.createClass):
3473         (Benchmark.Utilities.createClass):
3474         (get options): Deleted.
3475
3476         * Animometer/tests/bouncing-particles/resources/bouncing-canvas-images.js:
3477         * Animometer/tests/bouncing-particles/resources/bouncing-canvas-particles.js:
3478         * Animometer/tests/bouncing-particles/resources/bouncing-canvas-shapes.js:
3479         * Animometer/tests/bouncing-particles/resources/bouncing-css-images.js:
3480         * Animometer/tests/bouncing-particles/resources/bouncing-css-shapes.js:
3481         * Animometer/tests/bouncing-particles/resources/bouncing-particles.js:
3482         * Animometer/tests/bouncing-particles/resources/bouncing-svg-images.js:
3483         * Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js:
3484         * Animometer/tests/master/resources/canvas-stage.js:
3485         * Animometer/tests/master/resources/canvas-tests.js:
3486         * Animometer/tests/master/resources/particles.js:
3487         * Animometer/tests/misc/resources/canvas-electrons.js:
3488         * Animometer/tests/misc/resources/canvas-stars.js:
3489         * Animometer/tests/misc/resources/compositing-transforms.js:
3490         * Animometer/tests/simple/resources/simple-canvas-paths.js:
3491         * Animometer/tests/simple/resources/tiled-canvas-image.js:
3492         * Animometer/tests/template/resources/template-canvas.js:
3493         * Animometer/tests/template/resources/template-css.js:
3494         * Animometer/tests/template/resources/template-svg.js:
3495         * Animometer/tests/text/resources/layering-text.js:
3496
3497 2016-02-07  Jon Lee  <jonlee@apple.com>
3498
3499         Update how the benchmark is run
3500         https://bugs.webkit.org/show_bug.cgi?id=153960
3501
3502         Provisionally reviewed by Said Abou-Hallawa.
3503
3504         Introduce the notion of a Controller. It is responsible for recording, updating,
3505         and processing the statistics and complexity of the benchmark. This allows
3506         plugging in different Controllers.
3507
3508         This strips most of the functionality from Animator and BenchmarkState, so fold
3509         what's left into Benchmark. Now, Benchmarks only own a stage and a controller, but
3510         are responsible for driving the animation loop.
3511
3512         Rewrite Animator._shouldRequestAnotherFrame into two different Controllers. One
3513         maintains a fixed complexity, and the other adapts the complexity to meet a
3514         fixed FPS.
3515
3516         Fix the Kalman estimator to be modeled on a scalar variable with no model.
3517
3518         * Animometer/tests/resources/main.js: Remove BenchmarkState and Animator, and
3519         replace it with a Controller. Add a FixedController and refactor the previous controller
3520         to an AdaptiveController.
3521
3522         (Controller): Controllers own the estimator and the sampler. When a new frame is
3523         displayed, the animation loop calls update(). The estimator and sampler record
3524         stats, then tune. Samplers can track multiple series of data. The basic controller
3525         tracks timestamp, complexity, and estimated frame rate.
3526                 The Kalman estimation is based on the frame length rather than the frame
3527         rate. Because FPS is inverse proportional to frame length, in the case where the measured
3528         frame length is very small, the FPS ends up being a wildly large number (in the order of
3529         600-1000 "FPS"), and it pulls the estimator up drastically enough that it takes a while
3530         for it to settle back down. Using frame length reduces the impact of these spikes.
3531                 Converging the estimation takes enough time to avoid initializing it immediately
3532         when the benchmark starts. Instead, the benchmark runs for a brief period of time (100ms)
3533         before running it in earnest. Allow controllers an opportunity to set the complexity
3534         before starting recording.
3535                 When the benchmark is complete, the controller has an opportunity to process
3536         the samples. The default implementation calculates the raw FPS based on the time
3537         difference of the samples, and calculates the complexity score. This is moved from
3538         Benchmark.processSamples.
3539
3540         (Controller): Initialize timestamps. These are at first relative to the start of the
3541         benchmark, but are offset by the absolute start time during start(). By default maintain
3542         3 data series, but subclasses can override.
3543         (start): Calls recordFirstSample() for subclasses to override if needed.
3544         (recordFirstSample): For basic controller, start sampling at the beginning.
3545         (update): Update the frame length estimator and sample.
3546         (shouldStop): Checks that the time is before _endTimestamp.
3547         (results): Returns the processed samples.
3548         (processSamples): Iterate through the sample data and collate them. Include scores.
3549
3550         (FixedComplexityController): Controller that tunes the stage to the desired complexity
3551         prior to starting, and keeps it at that complexity.
3552
3553         (AdaptiveController): Have the estimator estimate the interval frame rate instead of the
3554         raw frame rate.
3555                 The previous version of this controller ignored the frame that came after the
3556         adjustment. The raw FPS show that whatever noise the scene change adds is negligible
3557         compared to the noise of the system overall. Stop ignoring that frame and include all
3558         frames in the measurements.
3559
3560         (Benchmark): Remove dependency on animator, and instantiate a runner based on what is
3561         selected. Most of the loop's functionality is in Controller, so remove here.
3562         (Benchmark.run): Remove start() since it is only called from run(), and fold it in here.
3563         (Benchmark._animateLoop): Fold in from Animator.animateLoop. Let the benchmark run for
3564         a brief period before calling Controller.start().
3565
3566         * Animometer/tests/resources/math.js: Fix the Kalman estimator. The filter estimates
3567         a scalar variable, and makes basic assumptions regarding the model. As a result
3568         none of the linear algebra classes are needed, so remove Matrix, Vector3, and Matrix3.
3569         (SimpleKalmanEstimator): Calculate the gain based on the provided process and
3570         measurement errors.
3571         (KalmanEstimator): Deleted.
3572         (IdentityEstimator): Deleted.
3573         (PIDController): Refactor to use the Utilities.createClass() helper.
3574
3575         The Kalman filter algorithm is explained here http://greg.czerniak.info/guides/kalman1/.
3576         The state, represented by a scalar, is the estimated frame length. There is no user
3577         transition of the state, and the state is the same as the measurement. With this model,
3578         the estimation error converges, so calculate the gain ahead of time.
3579
3580         * Animometer/developer.html: Remove fixed-after-warmup since it is not useful.
3581         Replace the option to toggle the estimator, and make it possible to customize the
3582         estimator's error parameters. Show raw FPS by default, and remove interval FPS,
3583         which will be shown instead of the filtered raw FPS.
3584         * Animometer/resources/debug-runner/animometer.css: Put the header behind the graph.
3585         Remove #intervalFPS rules; move the color to #filteredFPS.
3586         * Animometer/resources/debug-runner/graph.js:
3587         (updateGraphData): Update the hr style to force the layout to be calculated
3588         correctly. Change the tick format to be in terms of seconds, since the timestamps
3589         are in milliseconds. Remove interval data.
3590         * Animometer/resources/runner/animometer.js:
3591         (window.benchmarkController.startBenchmark): Set Kalman parameters.
3592         * Animometer/resources/runner/benchmark-runner.js:
3593         (_runBenchmarkAndRecordResults): When a benchmark completes, expect it to return
3594         the final data, rather than passing a sampler from the controller. This avoids
3595         needing to expose the sampler variable in the benchmark.
3596         * Animometer/tests/resources/sampler.js:
3597         (process): Move the setting of the target frame rate to AdaptiveController.
3598
3599 2016-02-06  Jon Lee  <jonlee@apple.com>
3600
3601         Code clean up: Move Rotater function closer to Stage static methods.
3602         The Rotater is used together with those methods; keep them close.
3603
3604         * Animometer/tests/resources/main.js:
3605
3606 2016-02-06  Jon Lee  <jonlee@apple.com>
3607
3608         Update the JS includes due to ResultsTable move.
3609
3610         * Animometer/developer.html:
3611         * Animometer/index.html:
3612
3613 2016-02-06  Jon Lee  <jonlee@apple.com>
3614
3615         Move createElement and createSVGElement to Utilities.
3616
3617         * Animometer/resources/extensions.js:
3618         (Utilities.createElement): Added.
3619         (Utilities.createSVGElement): Added.
3620         (DocumentExtension.createElement): Deleted.
3621         (DocumentExtension.createSvgElement): Deleted.
3622
3623         * Animometer/resources/debug-runner/animometer.js:
3624         * Animometer/resources/runner/animometer.js:
3625         * Animometer/tests/bouncing-particles/resources/bouncing-svg-images.js:
3626         * Animometer/tests/bouncing-particles/resources/bouncing-svg-particles.js:
3627         * Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js:
3628
3629 2016-02-06  Jon Lee  <jonlee@apple.com>
3630
3631         Add a convenience function for creating a class.
3632
3633         The pattern for creating a class is common enough to add as a Utilities
3634         helper function. It also makes it easy to collapse class definitions when
3635         editing.
3636
3637         * Animometer/resources/debug-runner/animometer.js: Move ProgressBar definition,
3638         since it is only used here.
3639         * Animometer/resources/runner/animometer.js: Move ResultsDashboard and
3640         ResultsTable definition, since it is only used here.
3641         * Animometer/resources/extensions.js: Move Utilities definition to the top. Convert
3642         Point, Insets, SimplePromise.
3643         (ProgressBar): Moved to animometer.js.
3644         (ResultsDashboard): Moved to animometer.js.
3645         (ResultsTable): Moved to animometer.js.
3646         * Animometer/resources/runner/benchmark-runner.js: Convert BenchmarkRunnerState,
3647         BenchmarkRunner.
3648         * Animometer/tests/resources/main.js: Convert Rotater, Stage, Animator, Benchmark.
3649         * Animometer/tests/resources/sampler.js: Convert Experiment, Sampler.
3650
3651         Convert test primitives.
3652         * Animometer/tests/master/resources/canvas-tests.js: Convert CanvasLineSegment,
3653         CanvasArc, CanvasLinePoint.
3654         * Animometer/tests/simple/resources/simple-canvas-paths.js: Convert CanvasLineSegment,
3655         CanvasLinePoint, CanvasQuadraticSegment, CanvasQuadraticPoint, CanvasBezierSegment,
3656         CanvasBezierPoint, CanvasArcToSegment, CanvasArcToSegmentFill, CanvasArcSegment,
3657         CanvasArcSegmentFill, CanvasRect, CanvasRectFill.
3658         * Animometer/tests/simple/resources/tiled-canvas-image.js: Convert CanvasImageTile.
3659
3660 2016-02-06  Jon Lee  <jonlee@apple.com>
3661
3662         Minor improvements to debug harness.
3663
3664         * Animometer/developer.html:
3665         * Animometer/resources/debug-runner/animometer.css:
3666         (#suites): Put the complexity text boxes closer to the test names.
3667         (#options):
3668         (#rawFPS circle): Make the interval FPS appear as a separate data series, with a line.
3669         (#intervalFPS path):
3670         (#intervalFPS circle):
3671         * Animometer/resources/debug-runner/animometer.js:
3672         (window.optionsManager.updateLocalStorageFromUI): Convert number inputs from text.
3673         (window.suitesManager._onChangeTestCheckbox): Refactor to take a checkbox.
3674         (window.suitesManager._createTestElement): Enhance such that typing into the complexity
3675         input will automatically select that test for running.
3676         (window.suitesManager.updateLocalStorageFromJSON): Make the harness work for private
3677         browsing.
3678         * Animometer/resources/debug-runner/graph.js: Separate the intervalFPS data, and show
3679         more accuracy in timestamps.
3680
3681 2016-02-06  Jon Lee  <jonlee@apple.com>
3682
3683         Refactor helper methods for getting random values for a stage.
3684
3685         Instead of requiring a Stage instance, just attach it to the Stage object.
3686
3687         * Animometer/tests/bouncing-particles/resources/bouncing-canvas-shapes.js:
3688         * Animometer/tests/bouncing-particles/resources/bouncing-css-shapes.js:
3689         * Animometer/tests/bouncing-particles/resources/bouncing-particles.js:
3690         * Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js:
3691         * Animometer/tests/master/resources/canvas-tests.js:
3692         * Animometer/tests/master/resources/particles.js:
3693         * Animometer/tests/misc/resources/canvas-electrons.js:
3694         * Animometer/tests/misc/resources/canvas-stars.js:
3695         * Animometer/tests/misc/resources/compositing-transforms.js:
3696         * Animometer/tests/resources/main.js:
3697         * Animometer/tests/simple/resources/simple-canvas-paths.js:
3698
3699 2016-02-05  Said Abou-Hallawa  <sabouhallawa@apple.com>
3700
3701         Add a new graphics test for CanvasRenderingContext2D functions: getImageData and putImageData
3702         https://bugs.webkit.org/show_bug.cgi?id=151716
3703
3704         Reviewed by Darin Adler.
3705
3706         The purpose of this test is to measure the performance of getImageData
3707         and putImageData functions. This test draws a background on the canvas
3708         and then gets some random tiles from this background and draw them in
3709         destinations different from their original sources.
3710
3711         * Animometer/resources/debug-runner/tests.js: Adding the new test to the canvas simple tests suite.
3712
3713         * Animometer/resources/extensions.js:
3714         (Array.prototype.shuffle): Shuffles the elements of an array.
3715
3716         (Point.zero): Returns a new Point object whose x and y are equal zero.
3717         (Point.prototype.str): Used for debugging the Point object.
3718
3719         * Animometer/tests/simple/resources/tiled-canvas-image.js: Added.
3720         (CanvasImageTile):
3721         (CanvasImageTile.prototype.getImageData):
3722         (CanvasImageTile.prototype.putImageData):
3723         (Stage.call.initialize):
3724         (Stage.call._createTiles):
3725         (Stage.call._nextTilePosition):
3726         (Stage.call.tune):
3727         (Stage.call._drawBackground):
3728         (Stage.call.animate):
3729         (Stage.call.complexity):
3730         (Stage.call):
3731         * Animometer/tests/simple/tiled-canvas-image.html: Added.
3732
3733 2016-01-07  Jon Lee  <jonlee@apple.com>
3734
3735         Fix new test.
3736
3737         * Animometer/resources/runner/tests.js: Wrong URL from an
3738         old patch.
3739         * Animometer/tests/master/particles.html:
3740         * Animometer/tests/master/resources/dom-particles.js:
3741         (Particle.call.reset): Figured out a simpler way to set up
3742         the particles.
3743         (this.move.reset): Deleted.
3744         (this.move._applyAttributes): Deleted.
3745         * Animometer/tests/master/resources/particles.js:
3746         (Particle): Call move() after reset().
3747
3748 2016-01-07  Jon Lee  <jonlee@apple.com>
3749
3750         Update benchmark test suite
3751         https://bugs.webkit.org/show_bug.cgi?id=152679
3752
3753         Reviewed by Simon Fraser.
3754
3755         Add a new test. The test has a rotating background
3756         gradient, and does a better job physically simulating
3757         particles.
3758
3759         * Animometer/resources/extensions.js: Teach Point to take constants as well as other Points.
3760         (Point.prototype.length): Added.
3761         (Point.prototype.normalize): Added.
3762         * Animometer/resources/runner/tests.js: Add the test to the master suite.
3763         * Animometer/tests/master/particles.html: Added.
3764         * Animometer/tests/master/resources/particles.js: Added. Parent class for different kinds of particles.
3765         (Particle):
3766         (Particle.prototype.reset): If the particle starts slowing down in terms of its animation,
3767         reset it.
3768         (Particle.prototype.animate): Bounce off the walls elastically, and include gravity.
3769         (Particle.prototype.move): Subclasses should override.
3770         (ParticlesStage): Stage includes a rotating gradient background.
3771         * Animometer/tests/master/resources/dom-particles.js: Added. Creates a <div> and adds it to
3772         the stage.
3773         * Animometer/tests/resources/star.svg: Added.
3774
3775 2016-01-03  Jon Lee  <jonlee@apple.com>
3776
3777         Update benchmark test suite
3778         https://bugs.webkit.org/show_bug.cgi?id=152679
3779
3780         Reviewed by Simon Fraser.
3781
3782         Move algorithm.js and sampler.js to tests/ and benchmark-runner.js to runner/.
3783
3784         Needed by both harnesses.
3785         * Animometer/resources/runner/benchmark-runner.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/benchmark-runner.js.
3786         * Animometer/developer.html:
3787         * Animometer/index.html:
3788
3789         Needed only by the tests. Move to tests/. Statistics, in sampler.js, is used by ResultsDashboard, so move that
3790         into extensions.js.
3791         * Animometer/resources/extensions.js:
3792         * Animometer/tests/resources/algorithm.js: Renamed from PerformanceTests/Animometer/resources/algorithm.js.
3793         * Animometer/tests/resources/sampler.js: Renamed from PerformanceTests/Animometer/resources/sampler.js.
3794         * Animometer/tests/bouncing-particles/bouncing-canvas-images.html:
3795         * Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html:
3796         * Animometer/tests/bouncing-particles/bouncing-css-images.html:
3797         * Animometer/tests/bouncing-particles/bouncing-css-shapes.html:
3798         * Animometer/tests/bouncing-particles/bouncing-svg-images.html:
3799         * Animometer/tests/bouncing-particles/bouncing-svg-shapes.html:
3800         * Animometer/tests/master/canvas-stage.html:
3801         * Animometer/tests/misc/canvas-electrons.html:
3802         * Animometer/tests/misc/canvas-stars.html:
3803         * Animometer/tests/misc/compositing-transforms.html:
3804         * Animometer/tests/simple/simple-canvas-paths.html:
3805         * Animometer/tests/template/template-canvas.html:
3806         * Animometer/tests/template/template-css.html:
3807         * Animometer/tests/template/template-svg.html:
3808         * Animometer/tests/text/layering-text.html:
3809         * Animometer/tests/text/text-boxes.html:
3810
3811 2016-01-07  Jon Lee  <jonlee@apple.com>
3812
3813         Update benchmark test suite
3814         https://bugs.webkit.org/show_bug.cgi?id=152679
3815
3816         Reviewed by Simon Fraser.
3817
3818         Fix tests for other browsers.
3819
3820         * Animometer/resources/extensions.js:
3821         (Point.elementClientSize): Some browsers return 0 for SVG clientWidth and clientHeight.
3822         Use getBoundingClientRect() instead.
3823         * Animometer/tests/misc/resources/canvas-electrons.js:
3824         (CanvasElectron.prototype._draw): Some browsers don't support ellipse.