Beef up the ARES-6 explainer text
[WebKit-https.git] / PerformanceTests / ChangeLog
1 2017-02-05  Filip Pizlo  <fpizlo@apple.com>
2
3         Beef up the ARES-6 explainer text
4         https://bugs.webkit.org/show_bug.cgi?id=167867
5
6         Reviewed by Saam Barati.
7         
8         Added appropriate links.
9
10         Added text that explains the subscores.
11
12         * ARES-6/index.html:
13
14 2017-02-06  Jer Noble  <jer.noble@apple.com>
15
16         Playback stalls when a SourceBuffer append causes frame eviction
17         https://bugs.webkit.org/show_bug.cgi?id=167834
18
19         Reviewed by Eric Carlson.
20
21         Add an in-page performance test measuring the amount of time required
22         to append a large amount of media data to a SourceBuffer, and then to
23         completely remove that data 30s at a time.
24
25         Add a microbenchmark for MediaTime which measures the amount of time
26         required to create a 1M entry std::map and traverse the map 1M times.
27
28         * Media/MSERemoveCodedFrames.html: Added.
29         * Media/media-source-loader.js:
30         (MediaSourceLoader.prototype.get duration):
31         * MediaTime/Configurations/Base.xcconfig: Added.
32         * MediaTime/Configurations/DebugRelease.xcconfig: Added.
33         * MediaTime/Makefile: Added.
34         * MediaTime/MediaTime.xcodeproj/project.pbxproj: Added.
35         * MediaTime/main.cpp: Added.
36         (performTest):
37         (test):
38         (main):
39         * Skipped:
40
41 2017-02-06  Saam Barati  <sbarati@apple.com>
42
43         Make ARES-6 work from the CLI again
44         https://bugs.webkit.org/show_bug.cgi?id=167895
45
46         Reviewed by Michael Saboff.
47
48         * ARES-6/driver.js:
49         (Driver.prototype.readyTrigger):
50         (Driver.prototype.disableTrigger):
51         * ARES-6/stats.js:
52         (Stats.prototype.toString.span.span):
53         (Stats.prototype.toString.return.span):
54         (Stats.prototype.toString):
55
56 2017-02-06  Ryan Haddad  <ryanhaddad@apple.com>
57
58         Skip ARES-6 on performance bots
59         https://bugs.webkit.org/show_bug.cgi?id=167863
60
61         Unreviewed test gardening.
62
63         ES6SampleBench was renamed to ARES-6, so this should be updated in PerformanceTests/Skipped as well.
64
65         * Skipped:
66
67 2017-02-05  Filip Pizlo  <fpizlo@apple.com>
68
69         Change ES6SampleBench into ARES-6
70         https://bugs.webkit.org/show_bug.cgi?id=167863
71
72         Rubber stamped by Saam Barati.
73         
74         This imports changes that Jonathan Davis (jond@apple.com) made to ES6SampleBench to make it look
75         super awesome. Its now called ARES-6.
76
77         * ARES-6: Copied from PerformanceTests/ES6SampleBench.
78         * ARES-6/ARES-6.svg: Added.
79         * ARES-6/driver.js:
80         (Driver):
81         (Driver.prototype.readyTrigger):
82         (Driver.prototype.disableTrigger):
83         (Driver.prototype.start):
84         (Driver.prototype._updateIterations):
85         * ARES-6/glue.js:
86         * ARES-6/index.html:
87         * ARES-6/results.js:
88         (Results.prototype.reportRunning):
89         (Results.prototype.reportDone):
90         * ARES-6/stats.js:
91         (Stats.prototype.toString.return.span):
92         (Stats.prototype.toString.span.span):
93         (Stats.prototype.toString): Deleted.
94         * ARES-6/style.css: Removed.
95         * ARES-6/styles.css: Added.
96         (html):
97         (body):
98         (header,):
99         (p a):
100         (p a:hover):
101         (.flip):
102         (header):
103         (h2,):
104         (p):
105         (.about h2):
106         (.logo):
107         (.start):
108         (.start.ready):
109         (.start.ready:hover):
110         (.start:after):
111         (.start.ready .testrun):
112         (.tests):
113         (.test):
114         (.overall h2,):
115         (.test .indicator):
116         (.test .indicator.running):
117         (@keyframes test-running):
118         (100%):
119         (.score):
120         (.score label):
121         (.score .value):
122         (.score .units):
123         (.score .margin):
124         (.overall):
125         (.overall .score):
126         (.overall .margin):
127         (.testing, .about):
128         (.testing):
129         (.about):
130         (#about:target):
131         (@media only screen and (max-width: 784px)):
132         * ARES-6/swoop.svg: Added.
133         * ES6SampleBench: Removed.
134         * ES6SampleBench/Air: Removed.
135         * ES6SampleBench/Air/README.md: Removed.
136         * ES6SampleBench/Air/airjs-tests.yaml: Removed.
137         * ES6SampleBench/Air/all.js: Removed.
138         * ES6SampleBench/Air/allocate_stack.js: Removed.
139         * ES6SampleBench/Air/arg.js: Removed.
140         * ES6SampleBench/Air/basic_block.js: Removed.
141         * ES6SampleBench/Air/benchmark.js: Removed.
142         * ES6SampleBench/Air/code.js: Removed.
143         * ES6SampleBench/Air/custom.js: Removed.
144         * ES6SampleBench/Air/frequented_block.js: Removed.
145         * ES6SampleBench/Air/insertion_set.js: Removed.
146         * ES6SampleBench/Air/inst.js: Removed.
147         * ES6SampleBench/Air/liveness.js: Removed.
148         * ES6SampleBench/Air/make_dist.sh: Removed.
149         * ES6SampleBench/Air/opcode.js: Removed.
150         * ES6SampleBench/Air/payload-airjs-ACLj8C.js: Removed.
151         * ES6SampleBench/Air/payload-gbemu-executeIteration.js: Removed.
152         * ES6SampleBench/Air/payload-imaging-gaussian-blur-gaussianBlur.js: Removed.
153         * ES6SampleBench/Air/payload-typescript-scanIdentifier.js: Removed.
154         * ES6SampleBench/Air/reg.js: Removed.
155         * ES6SampleBench/Air/stack_slot.js: Removed.
156         * ES6SampleBench/Air/stress-test.js: Removed.
157         * ES6SampleBench/Air/strip-hash.rb: Removed.
158         * ES6SampleBench/Air/symbols.js: Removed.
159         * ES6SampleBench/Air/test.html: Removed.
160         * ES6SampleBench/Air/test.js: Removed.
161         * ES6SampleBench/Air/tmp.js: Removed.
162         * ES6SampleBench/Air/tmp_base.js: Removed.
163         * ES6SampleBench/Air/util.js: Removed.
164         * ES6SampleBench/Basic: Removed.
165         * ES6SampleBench/Basic/ast.js: Removed.
166         * ES6SampleBench/Basic/basic-tests.yaml: Removed.
167         * ES6SampleBench/Basic/basic.js: Removed.
168         * ES6SampleBench/Basic/benchmark.js: Removed.
169         * ES6SampleBench/Basic/caseless_map.js: Removed.
170         * ES6SampleBench/Basic/lexer.js: Removed.
171         * ES6SampleBench/Basic/number.js: Removed.
172         * ES6SampleBench/Basic/parser.js: Removed.
173         * ES6SampleBench/Basic/random.js: Removed.
174         * ES6SampleBench/Basic/state.js: Removed.
175         * ES6SampleBench/Basic/stress-test.js: Removed.
176         * ES6SampleBench/Basic/test.html: Removed.
177         * ES6SampleBench/Basic/test.js: Removed.
178         * ES6SampleBench/Basic/util.js: Removed.
179         * ES6SampleBench/air_benchmark.js: Removed.
180         * ES6SampleBench/basic_benchmark.js: Removed.
181         * ES6SampleBench/cli.js: Removed.
182         * ES6SampleBench/driver.js: Removed.
183         * ES6SampleBench/glue.js: Removed.
184         * ES6SampleBench/index.html: Removed.
185         * ES6SampleBench/results.js: Removed.
186         * ES6SampleBench/stats.js: Removed.
187         * ES6SampleBench/style.css: Removed.
188
189 2017-02-05  Zalan Bujtas  <zalan@apple.com>
190
191         Simple line layout: Add performance test with varying string and font-size to minimize width-cache hit.
192         https://bugs.webkit.org/show_bug.cgi?id=167859
193         <rdar://problem/30368070>
194
195         Reviewed by Antti Koivisto.
196
197         This is in preparation for webkit.org/b/167843.
198
199         * Layout/simple-line-layout-with-varying-content.html: Added.
200
201 2017-01-16  Filip Pizlo  <fpizlo@apple.com>
202
203         Add framerate baselines to the splay-detail test
204         https://bugs.webkit.org/show_bug.cgi?id=167659
205
206         Reviewed by Geoffrey Garen.
207
208         * JetStream/Octane2/splay-detail.html:
209
210 2017-01-31  Carlos Alberto Lopez Perez  <clopez@igalia.com>
211
212         [EFL][GTK] Skip some IndexedDB tests that time out on the Perf bots.
213
214         Unreviewed performance test gardening.
215
216         * Skipped:
217
218 2017-01-30  Zalan Bujtas  <zalan@apple.com>
219
220         Simple line layout: Small tweaks to improve performance.
221         https://bugs.webkit.org/show_bug.cgi?id=167611
222         <rdar://problem/30274294>
223
224         Reviewed by Simon Fraser.
225
226         * Layout/simple-line-layout-non-repeating-text.html: Added.
227
228 2017-01-13  Said Abou-Hallawa  <sabouhallawa@apple.com>
229
230         26 MotionMark performance tests failing
231         https://bugs.webkit.org/show_bug.cgi?id=166854
232
233         Reviewed by Ryosuke Niwa.
234         
235         Skip running MotionMark on WebKit perf bots.
236
237         * Skipped:
238
239 2017-01-06  Said Abou-Hallawa  <sabouhallawa@apple.com>
240
241         Rename the directory and the files of MotionMark from Animometer to MotionMark
242         https://bugs.webkit.org/show_bug.cgi?id=166659
243
244         Reviewed by Jon Lee.
245         
246         Rename the directory and the files of the benchmark to its new name.
247
248         * MotionMark/developer.html: Renamed from PerformanceTests/Animometer/developer.html.
249         * MotionMark/index.html: Renamed from PerformanceTests/Animometer/index.html.
250         * MotionMark/resources/debug-runner/d3.min.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/d3.min.js.
251         * MotionMark/resources/debug-runner/graph.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/graph.js.
252         * MotionMark/resources/debug-runner/motionmark.css: Renamed from PerformanceTests/Animometer/resources/debug-runner/animometer.css.
253         * MotionMark/resources/debug-runner/motionmark.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/animometer.js.
254         * MotionMark/resources/debug-runner/tests.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/tests.js.
255         * MotionMark/resources/extensions.js: Renamed from PerformanceTests/Animometer/resources/extensions.js.
256         * MotionMark/resources/runner/benchmark-runner.js: Renamed from PerformanceTests/Animometer/resources/runner/benchmark-runner.js.
257         * MotionMark/resources/runner/crystal.svg: Renamed from PerformanceTests/Animometer/resources/runner/crystal.svg.
258         * MotionMark/resources/runner/lines.svg: Renamed from PerformanceTests/Animometer/resources/runner/lines.svg.
259         * MotionMark/resources/runner/logo.svg: Renamed from PerformanceTests/Animometer/resources/runner/logo.svg.
260         * MotionMark/resources/runner/motionmark.css: Renamed from PerformanceTests/Animometer/resources/runner/animometer.css.
261         * MotionMark/resources/runner/motionmark.js: Renamed from PerformanceTests/Animometer/resources/runner/animometer.js.
262         * MotionMark/resources/runner/tests.js: Renamed from PerformanceTests/Animometer/resources/runner/tests.js.
263         * MotionMark/resources/statistics.js: Renamed from PerformanceTests/Animometer/resources/statistics.js.
264         * MotionMark/resources/strings.js: Renamed from PerformanceTests/Animometer/resources/strings.js.
265         * MotionMark/tests/3d/resources/webgl.js: Renamed from PerformanceTests/Animometer/tests/3d/resources/webgl.js.
266         * MotionMark/tests/3d/webgl.html: Renamed from PerformanceTests/Animometer/tests/3d/webgl.html.
267         * MotionMark/tests/bouncing-particles/bouncing-canvas-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-canvas-images.html.
268         * MotionMark/tests/bouncing-particles/bouncing-canvas-shapes.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html.
269         * MotionMark/tests/bouncing-particles/bouncing-css-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-css-images.html.
270         * MotionMark/tests/bouncing-particles/bouncing-css-shapes.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-css-shapes.html.
271         * MotionMark/tests/bouncing-particles/bouncing-svg-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-svg-images.html.
272         * MotionMark/tests/bouncing-particles/bouncing-svg-shapes.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-svg-shapes.html.
273         * MotionMark/tests/bouncing-particles/bouncing-tagged-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-tagged-images.html.
274         * MotionMark/tests/bouncing-particles/resources/bouncing-canvas-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-canvas-images.js.
275         * MotionMark/tests/bouncing-particles/resources/bouncing-canvas-particles.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-canvas-particles.js.
276         * MotionMark/tests/bouncing-particles/resources/bouncing-canvas-shapes.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-canvas-shapes.js.
277         * MotionMark/tests/bouncing-particles/resources/bouncing-css-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-css-images.js.
278         * MotionMark/tests/bouncing-particles/resources/bouncing-css-shapes.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-css-shapes.js.
279         * MotionMark/tests/bouncing-particles/resources/bouncing-svg-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-svg-images.js.
280         * MotionMark/tests/bouncing-particles/resources/bouncing-svg-particles.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-svg-particles.js.
281         * MotionMark/tests/bouncing-particles/resources/bouncing-svg-shapes.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js.
282         * MotionMark/tests/bouncing-particles/resources/bouncing-tagged-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-tagged-images.js.
283         * MotionMark/tests/bouncing-particles/resources/image1.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image1.jpg.
284         * MotionMark/tests/bouncing-particles/resources/image2.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image2.jpg.
285         * MotionMark/tests/bouncing-particles/resources/image3.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image3.jpg.
286         * MotionMark/tests/bouncing-particles/resources/image4.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image4.jpg.
287         * MotionMark/tests/bouncing-particles/resources/image5.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image5.jpg.
288         * MotionMark/tests/dom/compositing-transforms.html: Renamed from PerformanceTests/Animometer/tests/dom/compositing-transforms.html.
289         * MotionMark/tests/dom/focus.html: Renamed from PerformanceTests/Animometer/tests/dom/focus.html.
290         * MotionMark/tests/dom/leaves.html: Renamed from PerformanceTests/Animometer/tests/dom/leaves.html.
291         * MotionMark/tests/dom/particles.html: Renamed from PerformanceTests/Animometer/tests/dom/particles.html.
292         * MotionMark/tests/dom/resources/compositing-transforms.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/compositing-transforms.js.
293         * MotionMark/tests/dom/resources/dom-particles.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/dom-particles.js.
294         * MotionMark/tests/dom/resources/focus.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/focus.js.
295         * MotionMark/tests/dom/resources/leaves.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/leaves.js.
296         * MotionMark/tests/master/canvas-stage.html: Renamed from PerformanceTests/Animometer/tests/master/canvas-stage.html.
297         * MotionMark/tests/master/focus.html: Renamed from PerformanceTests/Animometer/tests/master/focus.html.
298         * MotionMark/tests/master/image-data.html: Renamed from PerformanceTests/Animometer/tests/master/image-data.html.
299         * MotionMark/tests/master/leaves.html: Renamed from PerformanceTests/Animometer/tests/master/leaves.html.
300         * MotionMark/tests/master/multiply.html: Renamed from PerformanceTests/Animometer/tests/master/multiply.html.
301         * MotionMark/tests/master/resources/canvas-stage.js: Renamed from PerformanceTests/Animometer/tests/master/resources/canvas-stage.js.
302         * MotionMark/tests/master/resources/canvas-tests.js: Renamed from PerformanceTests/Animometer/tests/master/resources/canvas-tests.js.
303         * MotionMark/tests/master/resources/compass.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/compass.svg.
304         * MotionMark/tests/master/resources/compass100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/compass100.png.
305         * MotionMark/tests/master/resources/console.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/console.svg.
306         * MotionMark/tests/master/resources/console100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/console100.png.
307         * MotionMark/tests/master/resources/contribute.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/contribute.svg.
308         * MotionMark/tests/master/resources/contribute100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/contribute100.png.
309         * MotionMark/tests/master/resources/debugger.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/debugger.svg.
310         * MotionMark/tests/master/resources/debugger100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/debugger100.png.
311         * MotionMark/tests/master/resources/focus.js: Renamed from PerformanceTests/Animometer/tests/master/resources/focus.js.
312         * MotionMark/tests/master/resources/image-data.js: Renamed from PerformanceTests/Animometer/tests/master/resources/image-data.js.
313         * MotionMark/tests/master/resources/inspector.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/inspector.svg.
314         * MotionMark/tests/master/resources/inspector100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/inspector100.png.
315         * MotionMark/tests/master/resources/layout.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/layout.svg.
316         * MotionMark/tests/master/resources/layout100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/layout100.png.
317         * MotionMark/tests/master/resources/leaves.js: Renamed from PerformanceTests/Animometer/tests/master/resources/leaves.js.
318         * MotionMark/tests/master/resources/multiply.js: Renamed from PerformanceTests/Animometer/tests/master/resources/multiply.js.
319         * MotionMark/tests/master/resources/particles.js: Renamed from PerformanceTests/Animometer/tests/master/resources/particles.js.
320         * MotionMark/tests/master/resources/performance.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/performance.svg.
321         * MotionMark/tests/master/resources/performance100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/performance100.png.
322         * MotionMark/tests/master/resources/script.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/script.svg.
323         * MotionMark/tests/master/resources/script100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/script100.png.
324         * MotionMark/tests/master/resources/shortcuts.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/shortcuts.svg.
325         * MotionMark/tests/master/resources/shortcuts100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/shortcuts100.png.
326         * MotionMark/tests/master/resources/standards.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/standards.svg.
327         * MotionMark/tests/master/resources/standards100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/standards100.png.
328         * MotionMark/tests/master/resources/storage.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/storage.svg.
329         * MotionMark/tests/master/resources/storage100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/storage100.png.
330         * MotionMark/tests/master/resources/styles.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/styles.svg.
331         * MotionMark/tests/master/resources/styles100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/styles100.png.
332         * MotionMark/tests/master/resources/svg-particles.js: Renamed from PerformanceTests/Animometer/tests/master/resources/svg-particles.js.
333         * MotionMark/tests/master/resources/text.js: Renamed from PerformanceTests/Animometer/tests/master/resources/text.js.
334         * MotionMark/tests/master/resources/timeline.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/timeline.svg.
335         * MotionMark/tests/master/resources/timeline100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/timeline100.png.
336         * MotionMark/tests/master/svg-particles.html: Renamed from PerformanceTests/Animometer/tests/master/svg-particles.html.
337         * MotionMark/tests/master/text.html: Renamed from PerformanceTests/Animometer/tests/master/text.html.
338         * MotionMark/tests/resources/main.js: Renamed from PerformanceTests/Animometer/tests/resources/main.js.
339         * MotionMark/tests/resources/math.js: Renamed from PerformanceTests/Animometer/tests/resources/math.js.
340         * MotionMark/tests/resources/stage.css: Renamed from PerformanceTests/Animometer/tests/resources/stage.css.
341         * MotionMark/tests/resources/star.svg: Renamed from PerformanceTests/Animometer/tests/resources/star.svg.
342         * MotionMark/tests/resources/yin-yang.png: Renamed from PerformanceTests/Animometer/tests/resources/yin-yang.png.
343         * MotionMark/tests/resources/yin-yang.svg: Renamed from PerformanceTests/Animometer/tests/resources/yin-yang.svg.
344         * MotionMark/tests/simple/resources/simple-canvas-paths.js: Renamed from PerformanceTests/Animometer/tests/simple/resources/simple-canvas-paths.js.
345         * MotionMark/tests/simple/resources/simple-canvas.js: Renamed from PerformanceTests/Animometer/tests/simple/resources/simple-canvas.js.
346         * MotionMark/tests/simple/resources/tiled-canvas-image.js: Renamed from PerformanceTests/Animometer/tests/simple/resources/tiled-canvas-image.js.
347         * MotionMark/tests/simple/simple-canvas-paths.html: Renamed from PerformanceTests/Animometer/tests/simple/simple-canvas-paths.html.
348         * MotionMark/tests/simple/tiled-canvas-image.html: Renamed from PerformanceTests/Animometer/tests/simple/tiled-canvas-image.html.
349         * MotionMark/tests/template/resources/template-canvas.js: Renamed from PerformanceTests/Animometer/tests/template/resources/template-canvas.js.
350         * MotionMark/tests/template/resources/template-css.js: Renamed from PerformanceTests/Animometer/tests/template/resources/template-css.js.
351         * MotionMark/tests/template/resources/template-svg.js: Renamed from PerformanceTests/Animometer/tests/template/resources/template-svg.js.
352         * MotionMark/tests/template/template-canvas.html: Renamed from PerformanceTests/Animometer/tests/template/template-canvas.html.
353         * MotionMark/tests/template/template-css.html: Renamed from PerformanceTests/Animometer/tests/template/template-css.html.
354         * MotionMark/tests/template/template-svg.html: Renamed from PerformanceTests/Animometer/tests/template/template-svg.html.
355
356 2016-12-13  Brady Eidson  <beidson@apple.com>
357
358         Add a PerfTest targeting IDBObjectStore.get().
359         https://bugs.webkit.org/show_bug.cgi?id=165816
360
361         Reviewed by Alex Christensen.
362
363         * IndexedDB/objectstore-get.html: Added.
364
365 2016-12-13  Brady Eidson  <beidson@apple.com>
366
367         Add a PerfTest targeting IDBIndex.get().
368         https://bugs.webkit.org/show_bug.cgi?id=165803
369
370         Reviewed by Alex Christensen.
371
372         * IndexedDB/index-get.html: Added.
373
374 2016-12-12  Brady Eidson  <beidson@apple.com>
375
376         More IndexedDB perf tests.
377         https://bugs.webkit.org/show_bug.cgi?id=165634
378
379         Reviewed by Sam Weinig.
380
381         * IndexedDB/index-multientry.html:
382         * IndexedDB/large-number-of-inserts-responsiveness.html: Added.
383         * IndexedDB/large-number-of-inserts.html: Added.
384         * IndexedDB/objectstore-cursor.html: Added. Runtime cut in 1/4th compared to r209672.
385
386         * resources/runner.js:
387
388 2016-12-12  Commit Queue  <commit-queue@webkit.org>
389
390         Unreviewed, rolling out r209672.
391         https://bugs.webkit.org/show_bug.cgi?id=165766
392
393         IndexedDB/objectstore-cursor.html perf test always times out
394         (Requested by ap on #webkit).
395
396         Reverted changeset:
397
398         "More IndexedDB perf tests."
399         https://bugs.webkit.org/show_bug.cgi?id=165634
400         http://trac.webkit.org/changeset/209672
401
402 2016-12-10  Brady Eidson  <beidson@apple.com>
403
404         More IndexedDB perf tests.
405         https://bugs.webkit.org/show_bug.cgi?id=165634
406
407         Reviewed by Sam Weinig.
408
409         * IndexedDB/index-multientry.html:
410         * IndexedDB/large-number-of-inserts-responsiveness.html: Added.
411         * IndexedDB/large-number-of-inserts.html: Added.
412         * IndexedDB/objectstore-cursor.html: Added.
413         
414         * resources/runner.js: Add "track responsiveness" functionality to PerfTestRunner.
415
416 2016-12-06  Filip Pizlo  <fpizlo@apple.com>
417
418         Concurrent GC should be stable enough to land enabled
419         https://bugs.webkit.org/show_bug.cgi?id=164990
420
421         Reviewed by Geoffrey Garen.
422         
423         Made CDjs more configurable and refined the "large.js" configuration. I was using that one and
424         the new "long.js" configuration to tune concurrent eden GCs.
425         
426         Added a new way of running Splay in browser, which using chartjs to plot the execution times of
427         2000 iterations. This includes the minified chartjs.
428
429         * JetStream/Octane2/splay-detail.html: Added.
430         * JetStream/cdjs/benchmark.js:
431         (benchmarkImpl):
432         (benchmark):
433         * JetStream/cdjs/long.js: Added.
434
435 2016-12-07  Brady Eidson  <beidson@apple.com>
436
437         Add IDB perf tests stressing key size.
438         https://bugs.webkit.org/show_bug.cgi?id=165567
439
440         Reviewed by Alex Christensen.
441
442         * IndexedDB/large-array-keys.html: Added.
443         * IndexedDB/large-binary-keys.html: Added.
444         * IndexedDB/large-string-keys.html: Added.
445
446 2016-12-05  Brady Eidson  <beidson@apple.com>
447
448         Add an IndexedDB perf test to PerformanceTests.
449         https://bugs.webkit.org/show_bug.cgi?id=165430
450
451         Reviewed by Alex Christensen.
452
453         * IndexedDB/index-multientry.html: Added.
454
455 2016-12-02  Filip Pizlo  <fpizlo@apple.com>
456
457         ES6SampleBench should report an average for Steady State so that all of the numbers are comparable
458         https://bugs.webkit.org/show_bug.cgi?id=165325
459
460         Reviewed by Saam Barati.
461
462         This makes all of the numbers that ES6SampleBench reports comparable to each other: they all speak of the time
463         it took to run an iteration of something.
464
465         * ES6SampleBench/results.js:
466         (Results.prototype.reportResult):
467
468 2016-11-18  Jer Noble  <jer.noble@apple.com>
469
470         Add some performance tests for various aspects of media loading and playback.
471         https://bugs.webkit.org/show_bug.cgi?id=164977
472
473         Reviewed by Darin Adler.
474
475         Add new tests for a few aspects of media playback; namely: how quickly  media elements fire
476         the "canplaythrough" event when all data is immediately available, how quickly HLS playback
477         switches resolutions, and at what playback rate MSE-backed video can play without dropping
478         frames. Skip these tests by default, as they're not runnable on all ports, and they require
479         a webserver (run-webkit-httpd) to complete on macOS.
480
481         * Media/HLSCanPlayThrough.html: Added.
482         * Media/HLSGearChange.html: Added.
483         * Media/MSECanPlayThrough.html: Added.
484         * Media/MSEPlaybackRate.html: Added.
485         * Media/StandardCanPlayThrough.html: Added.
486         * Media/hls/1080p/iframe_index.m3u8: Added.
487         * Media/hls/1080p/prog_index.m3u8: Added.
488         * Media/hls/1080p/test.ts: Added.
489         * Media/hls/480p/iframe_index.m3u8: Added.
490         * Media/hls/480p/prog_index.m3u8: Added.
491         * Media/hls/480p/test.ts: Added.
492         * Media/hls/720p/iframe_index.m3u8: Added.
493         * Media/hls/720p/prog_index.m3u8: Added.
494         * Media/hls/720p/test.ts: Added.
495         * Media/hls/720p/test.ts.back: Added.
496         * Media/hls/index.m3u8: Added.
497         * Media/media-source-loader.js: Added.
498         (MediaSourceLoader):
499         (MediaSourceLoader.prototype.loadManifest):
500         (MediaSourceLoader.prototype.loadManifestSucceeded):
501         (MediaSourceLoader.prototype.loadManifestFailed):
502         (MediaSourceLoader.prototype.loadMediaData):
503         (MediaSourceLoader.prototype.loadMediaDataSucceeded):
504         (MediaSourceLoader.prototype.loadMediaDataFailed):
505         (MediaSourceLoader.prototype.get type):
506         (MediaSourceLoader.prototype.get duration):
507         (MediaSourceLoader.prototype.get initSegment):
508         (MediaSourceLoader.prototype.get mediaSegmentsLength):
509         (MediaSourceLoader.prototype.mediaSegments):
510         (MediaSourceLoader.prototype.get everyMediaSegment):
511         * Media/test-fragmented-video.json: Added.
512         * Media/test-fragmented-video.mp4: Added.
513         * Media/test.mp4: Added.
514         * Skipped:
515
516 2016-11-18  Filip Pizlo  <fpizlo@apple.com>
517
518         Concurrent GC should be able to run splay in debug mode and earley/raytrace in release mode with no perf regression
519         https://bugs.webkit.org/show_bug.cgi?id=164282
520
521         Reviewed by Geoffrey Garen and Oliver Hunt.
522         
523         CDjs is a fun benchmark for stressing concurrent GCs, but to really give the GC a good
524         workout you need to increase the amount of work that the test does. This adds a second
525         configuration of the benchmark that has more aircraft. It uses much more memory and causes us
526         to do more GCs and those GCs take longer.
527
528         * JetStream/cdjs/benchmark.js:
529         (benchmarkImpl):
530         (benchmark):
531         * JetStream/cdjs/large.js: Added.
532
533 2016-11-14  Filip Pizlo  <fpizlo@apple.com>
534
535         Unreviewed, revert unintended change.
536
537         * ES6SampleBench/Air/benchmark.js:
538
539 2016-10-03  Saam Barati  <sbarati@apple.com>
540
541         MapHash should speculate on the type of its child node
542         https://bugs.webkit.org/show_bug.cgi?id=161922
543
544         Reviewed by Filip Pizlo.
545
546         I gave the main generator a name so it's easier to see what
547         it is when using the sampling profiler.
548
549         * ES6SampleBench/Basic/ast.js:
550         (Basic.Program):
551
552 2016-09-19  Sergio Villar Senin  <svillar@igalia.com>
553
554         [css-grid] Remove the x2 computation of row sizes with indefinite heights
555         https://bugs.webkit.org/show_bug.cgi?id=162150
556
557         Reviewed by Darin Adler.
558
559         Added a new test case which checks the layout performance of grids inside other grids, i.e,
560         grids acting both as grid container and grid item.
561
562         * Layout/nested-grid.html: Added.
563
564 2016-09-09  Simon Fraser  <simon.fraser@apple.com>
565
566         Perf test Animation/css-accelerated-animation.html failing
567         https://bugs.webkit.org/show_bug.cgi?id=161795
568
569         Skip Animation/css-accelerated-animation.html since this only works as a MobileSafari
570         perf test.
571
572         * Skipped:
573
574 2016-09-08  Simon Fraser  <simon.fraser@apple.com>
575
576         Add a content-animation test with accelerated CSS animations
577         https://bugs.webkit.org/show_bug.cgi?id=161776
578
579         Reviewed by Dean Jackson.
580
581         Add a test that measures the frame rate of accelerated CSS animations.
582
583         This is like css-animation.html (which animates 'left' and 'top') but animates
584         transforms instead. In order to get animation in X and Y we make two nested elements
585         and set transformX() on one, and transformY() on the other.
586
587         * Animation/css-accelerated-animation.html: Added.
588
589 2016-08-23  Saam Barati  <sbarati@apple.com>
590
591         It should be easy to run ES6SampleBench from the jsc shell
592         https://bugs.webkit.org/show_bug.cgi?id=161085
593
594         Reviewed by Yusuke Suzuki.
595
596         This patch makes ES6 sample bench runnable through the `jsc` shell.
597         To do that, you need to be in the PerformanceTests/ES6SampleBench
598         directory and run `jsc cli.js`. To make this work, the benchmark
599         is now aware of being run in two modes: via the browser, and via
600         the shell. Each entry point will set a variable `isInBrowser`,
601         and the benchmark will do different things based on if that
602         variable is true or false.
603
604         * ES6SampleBench/Air/benchmark.js:
605         * ES6SampleBench/Air/stress-test.js:
606         * ES6SampleBench/Basic/benchmark.js:
607         (runBenchmark):
608         * ES6SampleBench/Basic/stress-test.js:
609         * ES6SampleBench/air_benchmark.js:
610         * ES6SampleBench/basic_benchmark.js:
611         * ES6SampleBench/cli.js: Added.
612         (return.doRun):
613         (makeBenchmarkRunner):
614         * ES6SampleBench/driver.js:
615         (Driver):
616         (Driver.prototype._recomputeSummary):
617         (Driver.prototype._iterate.window.setTimeout):
618         (Driver.prototype._iterate):
619         (Driver.prototype._updateIterations):
620         * ES6SampleBench/glue.js:
621         (reportResult):
622         * ES6SampleBench/index.html:
623         * ES6SampleBench/results.js:
624         (Results):
625         (Results.prototype.reportRunning):
626         (Results.prototype.reportDone):
627         (Results.prototype.reportError):
628         * ES6SampleBench/stats.js:
629         (Stats):
630         (Stats.prototype.toString):
631         (Stats.prototype._update):
632
633 2016-08-01  Said Abou-Hallawa  <sabouhallawa@apple.com>
634
635         Add an option to run the MotionMark tests with classic tile size (512x512)
636         https://bugs.webkit.org/show_bug.cgi?id=160371
637
638         Reviewed by Darin Adler.
639
640         Setting the body to large size (3000x3000) and overflow to scroll forces
641         the classic tile size (512x512). This option does not affect the canvas
642         tests because the canvas is displayed on one tile regardless of its size.
643
644         * Animometer/developer.html:
645         * Animometer/resources/debug-runner/animometer.css:
646         (body.showing-test-container.tiles-big):
647         (body.showing-test-container.tiles-classic):
648         * Animometer/resources/debug-runner/animometer.js:
649         * Animometer/resources/runner/animometer.js:
650         (window.benchmarkController.startBenchmark):
651
652 2016-06-29  Filip Pizlo  <fpizlo@apple.com>
653
654         Generators violate bytecode liveness validation
655         https://bugs.webkit.org/show_bug.cgi?id=159279
656
657         Reviewed by Yusuke Suzuki.
658
659         Add Basic to our test harness.
660
661         Also made some cosmetic changes to the benchmark harness.
662
663         * ES6SampleBench/Basic/basic-tests.yaml: Added.
664         * ES6SampleBench/Basic/stress-test.js: Added.
665         (preciseTime):
666         * ES6SampleBench/driver.js:
667         (Driver):
668         (Driver.prototype.start):
669         (Driver.prototype.reportError):
670         * ES6SampleBench/glue.js:
671         * ES6SampleBench/index.html:
672
673 2016-06-28  Filip Pizlo  <fpizlo@apple.com>
674
675         ES6SampleBench should have a harness
676         https://bugs.webkit.org/show_bug.cgi?id=159246
677
678         Reviewed by Saam Barati.
679         
680         This adds a simple web harness for ES6SampleBench. It runs Air and Basic 10 times for 200
681         iterations each and reports three metrics:
682         
683         First iteration: the time it takes for the first iteration to run. This is the first
684         iteration after the benchmark is loaded into the iframe, so it's representative of what
685         would happen if one of these workloads only ran for a short time.
686         
687         Worst 2%: of the last 199 iterations, the average of the worst 2% iterations. If code like
688         any of these workloads was used in an important event handler, you'd want that code to run
689         well in the worst case in addition to having great throughput.
690         
691         Steady state: the total of the last 199 iterations. This is representative of what would
692         happen if you ran code like this for a long time.
693         
694         The total score is the geomean of the firstIteration/worstCase/steadyState numbers of the
695         two benchmarks.
696         
697         The harness does statistics using Student's T-distribution confidence intervals.
698
699         * ES6SampleBench/Basic/benchmark.js:
700         (Benchmark):
701         * ES6SampleBench/air_benchmark.js: Added.
702         * ES6SampleBench/basic_benchmark.js: Added.
703         * ES6SampleBench/driver.js: Added.
704         (Driver):
705         (Driver.prototype.addBenchmark):
706         (Driver.prototype.start):
707         (Driver.prototype.reportResult):
708         (Driver.prototype.reportError):
709         (Driver.prototype._recomputeSummary.Geomean):
710         (Driver.prototype._recomputeSummary.Geomean.prototype.add):
711         (Driver.prototype._recomputeSummary.Geomean.prototype.get result):
712         (Driver.prototype._recomputeSummary):
713         (Driver.prototype._iterate):
714         (Driver.prototype._updateIterations):
715         * ES6SampleBench/glue.js: Added.
716         * ES6SampleBench/index.html: Added.
717         * ES6SampleBench/results.js: Added.
718         (Results):
719         (Results.prototype.get benchmark):
720         (Results.prototype.reset):
721         (Results.prototype.reportRunning):
722         (Results.prototype.reportDone):
723         (Results.prototype.reportResult.averageAbovePercentile):
724         (Results.prototype.reportResult):
725         (Results.prototype.reportError):
726         * ES6SampleBench/stats.js: Added.
727         (Stats):
728         (Stats.prototype.reset):
729         (Stats.prototype.add):
730         (Stats.prototype.get numIterations):
731         (Stats.prototype.valueForIteration):
732         (Stats.get result.tDist):
733         (Stats.prototype.get result):
734         (Stats.prototype.toString):
735         (Stats.prototype._update):
736         * ES6SampleBench/style.css: Added.
737         (body):
738         (body, th, tr):
739         (h1, h2, h3, h4):
740         (h1):
741         (h2):
742         (h3):
743         (hr):
744         (address):
745         (img):
746         (.underline):
747         (ol.loweralpha):
748         (ol.upperalpha):
749         (ol.lowerroman):
750         (ol.upperroman):
751         (ol.arabic):
752         (.banner-link:link, .banner-link:visited):
753         (:link, :visited):
754         (h4 :link, h4 :visited, h5 :link, h5 :visited, h6 :link, h6 :visited):
755         (.anchor:link, .anchor:visited):
756         (* > .anchor:link, * > .anchor:visited):
757         (span:hover .anchor):
758         (a.forbidden, span.forbidden):
759         (a.missing:hover):
760         (a.closed:link, a.closed:visited, span.closed):
761         (pre):
762         (div.code):
763         (div.code pre):
764         (dt):
765         (dd):
766         (dd:last-child):
767         (.site-logo):
768         (.site-logo .tagline):
769         (table):
770         (#contents):
771         (p):
772         (p:last-child):
773         (th):
774         (td):
775
776 2016-06-28  Jon Lee  <jonlee@apple.com>
777
778         Update focus test
779         https://bugs.webkit.org/show_bug.cgi?id=159242
780         rdar://problem/27070007
781
782         Reviewed by Dean Jackson.
783         Provisionally reviewed by Said Abou-Hallawa.
784
785         Move previous test to dom suite, and update the test for better reporting of frame rate, although
786         it uses a different rendering path.
787
788         * Animometer/resources/debug-runner/tests.js: Add to dom suite.
789         * Animometer/tests/dom/focus.html: Copied from PerformanceTests/Animometer/tests/master/focus.html.
790         * Animometer/tests/dom/resources/focus.js: Copied from PerformanceTests/Animometer/tests/master/resources/focus.js.
791         * Animometer/tests/master/focus.html: Remove center element.
792         * Animometer/tests/master/resources/focus.js: Use narrower size range with smaller particles. Remove the
793         container elements. Inline getBlurValue and getOpacityValue since they are only called once.
794
795 2016-06-28  Filip Pizlo  <fpizlo@apple.com>
796
797         Move Air.js and Basic into ES6SampleBench
798
799         Rubber stamped by Geoffrey Garen.
800         
801         We want to group our ES6 benchmarks together, and eventually, we want to give them a
802         common harness.
803
804         * Air.js: Removed.
805         * Air.js/README.md: Removed.
806         * Air.js/airjs-tests.yaml: Removed.
807         * Air.js/all.js: Removed.
808         * Air.js/allocate_stack.js: Removed.
809         * Air.js/arg.js: Removed.
810         * Air.js/basic_block.js: Removed.
811         * Air.js/benchmark.js: Removed.
812         * Air.js/code.js: Removed.
813         * Air.js/custom.js: Removed.
814         * Air.js/frequented_block.js: Removed.
815         * Air.js/insertion_set.js: Removed.
816         * Air.js/inst.js: Removed.
817         * Air.js/liveness.js: Removed.
818         * Air.js/make_dist.sh: Removed.
819         * Air.js/opcode.js: Removed.
820         * Air.js/payload-airjs-ACLj8C.js: Removed.
821         * Air.js/payload-gbemu-executeIteration.js: Removed.
822         * Air.js/payload-imaging-gaussian-blur-gaussianBlur.js: Removed.
823         * Air.js/payload-typescript-scanIdentifier.js: Removed.
824         * Air.js/reg.js: Removed.
825         * Air.js/stack_slot.js: Removed.
826         * Air.js/stress-test.js: Removed.
827         * Air.js/strip-hash.rb: Removed.
828         * Air.js/symbols.js: Removed.
829         * Air.js/test.html: Removed.
830         * Air.js/test.js: Removed.
831         * Air.js/tmp.js: Removed.
832         * Air.js/tmp_base.js: Removed.
833         * Air.js/util.js: Removed.
834         * Basic: Removed.
835         * Basic/ast.js: Removed.
836         * Basic/basic.js: Removed.
837         * Basic/benchmark.js: Removed.
838         * Basic/caseless_map.js: Removed.
839         * Basic/lexer.js: Removed.
840         * Basic/number.js: Removed.
841         * Basic/parser.js: Removed.
842         * Basic/random.js: Removed.
843         * Basic/state.js: Removed.
844         * Basic/test.html: Removed.
845         * Basic/test.js: Removed.
846         * Basic/util.js: Removed.
847         * ES6SampleBench: Added.
848         * ES6SampleBench/Air: Copied from PerformanceTests/Air.js.
849         * ES6SampleBench/Basic: Copied from PerformanceTests/Basic.
850         * Skipped:
851
852 2016-06-24  Filip Pizlo  <fpizlo@apple.com>
853
854         Add a ES6 generator benchmark
855         https://bugs.webkit.org/show_bug.cgi?id=159101
856
857         Rubber stamped by Keith Miller.
858         
859         This adds a Basic interpreter loosely based on ECMA-55:
860         
861         http://www.ecma-international.org/publications/files/ECMA-ST-WITHDRAWN/ECMA-55,%201st%20Edition,%20January%201978.pdf
862         
863         It includes a lexer that is a generator, a parser that uses regular expressions, and an AST
864         walk interpreter where the walking functions for statements are generators that call each
865         other with yield*. This enables the interpreter to look like an AST walk even though it can
866         yield at INPUT and PRINT statements.
867         
868         This also uses for-of, classes, Map, and WeakMap. It also uses deprecated-but-awesome RegExp
869         features like RegExp.lastMatch and RegExp.rightContext. I did it that way because this is
870         how I've always written lexers in dynamic languages; see offlineasm's lex() method in
871         parser.rb for example.
872         
873         The benchmark runs a handful of simple Basic programs. The longest-running one computes
874         prime numbers.
875         
876         Includes a command-line and web harness. On my machine it runs in 2-3 seconds.
877
878         * Basic: Added.
879         * Basic/ast.js: Added.
880         (Basic.NumberApply):
881         (Basic.Variable):
882         (Basic.Const):
883         (Basic.NumberPow):
884         (Basic.NumberMul):
885         (Basic.NumberDiv):
886         (Basic.NumberNeg):
887         (Basic.NumberAdd):
888         (Basic.NumberSub):
889         (Basic.StringVar):
890         (Basic.Equals):
891         (Basic.NotEquals):
892         (Basic.LessThan):
893         (Basic.GreaterThan):
894         (Basic.LessEqual):
895         (Basic.GreaterEqual):
896         (Basic.GoTo):
897         (Basic.GoSub):
898         (Basic.Def):
899         (Basic.Let):
900         (Basic.If):
901         (Basic.Return):
902         (Basic.Stop):
903         (Basic.On):
904         (sideState.shouldStop):
905         (Basic.For):
906         (Basic.Next):
907         (Basic.Print):
908         (Basic.Input):
909         (Basic.Read):
910         (Basic.Restore):
911         (Basic.Dim):
912         (Basic.Randomize):
913         (Basic.End):
914         (Basic.Program):
915         * Basic/basic.js: Added.
916         (prepare):
917         (simulate):
918         * Basic/benchmark.js: Added.
919         (Benchmark):
920         (Benchmark.prototype.runIteration.expect):
921         (Benchmark.prototype.runIteration):
922         (runBenchmark):
923         * Basic/caseless_map.js: Added.
924         (CaselessMap):
925         * Basic/lexer.js: Added.
926         (lex.consumeWhitespace):
927         (lex.consume):
928         (lex):
929         * Basic/lexer_test.js: Added.
930         * Basic/number.js: Added.
931         (NumberValue):
932         (NumberValue.prototype.get value):
933         (NumberValue.prototype.apply):
934         (NumberValue.prototype.leftApply):
935         (NumberValue.prototype.assign):
936         (NumberArray.):
937         (NumberArray):
938         (NumberArray.prototype.apply):
939         (NumberArray.prototype.leftApply):
940         (NumberFunction):
941         (NumberFunction.prototype.apply):
942         (NumberFunction.prototype.leftApply):
943         (NativeFunction):
944         (NativeFunction.prototype.apply):
945         (NativeFunction.prototype.leftApply):
946         * Basic/parser.js: Added.
947         (parse):
948         (parse.pushToken):
949         (parse.peekToken):
950         (parse.consumeKind):
951         (parse.consumeToken):
952         (parse.parseVariable):
953         (parse.parseNumericExpression.parsePrimary):
954         (parse.parseNumericExpression.parseFactor):
955         (parse.parseNumericExpression.parseTerm):
956         (parse.parseNumericExpression):
957         (parse.parseConstant):
958         (parse.parseStringExpression):
959         (parse.isStringExpression):
960         (parse.parseRelationalExpression):
961         (parse.parseNonNegativeInteger):
962         (parse.parseGoToStatement):
963         (parse.parseGoSubStatement):
964         (parse.parseStatement):
965         (parse.parseStatements):
966         * Basic/random.js: Added.
967         (createRNG):
968         (createRNGWithFixedSeed):
969         (createRNGWithRandomSeed):
970         * Basic/state.js: Added.
971         (State):
972         (State.prototype.getValue):
973         (State.prototype.getSideState):
974         (State.prototype.abort):
975         (State.prototype.validate):
976         * Basic/test.html: Added.
977         * Basic/test.js: Added.
978         * Basic/util.js: Added.
979         (this.performance.performance.now.currentTime):
980         (else.this.preciseTime.currentTime):
981         (else.currentTime):
982         * Skipped: Make sure that we don't run Basic yet.
983
984 2016-06-21  Jon Lee  <jonlee@apple.com>
985
986         Update canvas size when benchmark begins
987         https://bugs.webkit.org/show_bug.cgi?id=159010
988
989         Reviewed by Dean Jackson.
990
991         Previously we would evaluate the media queries and assign the canvas size to the test
992         on body.onload. Instead, do it once the user starts the benchmark.
993
994         * Animometer/resources/debug-runner/animometer.js: Move benchmarkController.determineCanvasSize()
995         to benchmarkController._startBenchmark.
996         * Animometer/resources/runner/animometer.js: Ditto.
997
998 2016-06-21  Jon Lee  <jonlee@apple.com>
999
1000         Add new timestamp option
1001         https://bugs.webkit.org/show_bug.cgi?id=159006
1002
1003         Reviewed by Dean Jackson.
1004
1005         Add a new option to take timestamps from the rAF callback.
1006
1007         * Animometer/developer.html: Added option.
1008         * Animometer/tests/resources/main.js:
1009         (Benchmark): Fall back to using the rAF timestamp if performance.now()
1010         is not available.
1011         (Benchmark._animateLoop): Update how timestamp is set. Prefer to use
1012         local var instead of accessing private var.
1013
1014 2016-06-21  Jon Lee  <jonlee@apple.com>
1015
1016         Improvements to Animometer benchmark
1017         https://bugs.webkit.org/show_bug.cgi?id=157738
1018
1019         Reviewed by Dean Jackson.
1020         Provisionally reviewed by Said Abou-Hallawa.
1021
1022         Update tests.
1023
1024         * Animometer/tests/master/text.html: Ensure only three text sizes for
1025         the three canvases.
1026         * Animometer/tests/master/focus.html: Reduce the text size to fit with smaller
1027         particle sizes.
1028         * Animometer/tests/master/resources/focus.js: Remove the quadratic distribution for
1029         particle sizes, and make it linear. Reduce the size variance. Shuffle the math to
1030         reduce some calculations per frame. Fix the placement of the particles which might
1031         otherwise be culled.
1032         * Animometer/tests/master/resources/image-data.js: Reduce the particle size
1033         to encourage larger scores.
1034
1035 2016-06-21  Jon Lee  <jonlee@apple.com>
1036
1037         Improvements to Animometer benchmark
1038         https://bugs.webkit.org/show_bug.cgi?id=157738
1039
1040         Reviewed by Dean Jackson.
1041         Provisionally reviewed by Said Abou-Hallawa.
1042
1043         Include confidence interval for the final score, and store the canvas
1044         size in the serialization so that it is accurately shown in results.
1045
1046         * Animometer/developer.html: Add a "confidence" div.
1047         * Animometer/index.html: Ditto. Convert "mean" to "confidence".
1048         * Animometer/resources/debug-runner/animometer.js: Look at options, and
1049         if the configuration is included, update the body class based on it
1050         (similar to what we do when we first load the page). That way, if you
1051         drag-and-drop previous results in, that configuration is reflected in
1052         the dashboard. Show the full confidence interval.
1053
1054         * Animometer/resources/debug-runner/animometer.css:
1055         * Animometer/resources/debug-runner/animometer.js: Style update.
1056         * Animometer/resources/runner/animometer.css:
1057
1058         * Animometer/resources/runner/animometer.js:
1059         (_processData): Propagate the confidence interval values out and calculate
1060         the lower and upper bounds. For now, shortcut the aggregate calculation,
1061         since we only go through one iteration.
1062         (this._processData.calculateScore): Propagate the confidence interval out
1063         to be next to the score. Depending on the controller these values are
1064         calculated differently.
1065         (this._processData._getResultsProperty): Convenience function.
1066         (this._processData.get score): Refactor.
1067         (this._processData.get scoreLowerBound): Get the aggregate lower bound.
1068         (this._processData.get scoreUpperBound): Get the aggregate upper bound.
1069         (window.sectionsManager.setSectionScore):
1070         (window.benchmarkController._startBenchmark): When the benchmark starts, note
1071         the canvas size and add it to options. That way it will be automatically be
1072         serialized.
1073         (window.benchmarkController.showResults): Include the maximum deviation
1074         percentage.
1075         * Animometer/resources/runner/lines.svg: Make the background line up with the
1076         skew.
1077         * Animometer/resources/runner/tests.js:
1078         (Headers.details.text): Refactor.
1079         * Animometer/resources/statistics.js:
1080         (Statistics.largestDeviationPercentage): Convenience function to calculate
1081         the deviation percentage on either end and return the largest deviation.
1082         * Animometer/resources/strings.js:
1083
1084         Allow specifying a regression profile to use instead of taking the one
1085         with the lowest error.
1086
1087         Address an issue in the focus test when the regression calculated ends
1088         up overestimating the change point, causing a cascade of tougher
1089         ramps. The reason behind this is that at max complexity of an initial
1090         ramp, the frame length is very high, and it influences the second
1091         segment of the piecewise regression strongly, causing it to be very
1092         steep. As a result, the first segment, expected to be flat, ends up
1093         covering a higher range of complexity. That makes the change point
1094         much higher than it should be. To avoid this, we will add a sanity
1095         check on the maximum value of the ramp. If the regression's projected
1096         value at the maximum complexity of the current ramp is very slow (less
1097         than 20 fps), 1) reduce the maximum complexity by 20%, and 2) do not
1098         include the regression's change point in the change point estimator.
1099         That estimator is used as the midpoint of the next ramp, and including
1100         the change point from a poor regression can bake in the error. The
1101         controller already knows how to adjust for ramps that are too easy for
1102         the system.
1103
1104         * Animometer/resources/runner/animometer.js:
1105         (this._processData.findRegression): Takes a preferred profile and gives that to
1106         Regression.
1107         (this._processData.calculateScore): With the ramp controller, take the profile
1108         of the ramp that was used the most when calculating the ramp's regression. That
1109         profile is what is used for the test's score.
1110         * Animometer/resources/statistics.js:
1111         (Regression.Utilities.createClass): Update to take an options object which can
1112         specify a profile to calculate with. Otherwise it will continue to use both and
1113         select the one with the lower error.
1114         (_calculateRegression): Fix an issue where we claim 0 error if the regression
1115         calculation fails due to divide-by-zero. Instead reject that regression calculation
1116         by giving it Number.MAX_VALUE.
1117         * Animometer/resources/strings.js: New strings for marking a regression as flat
1118         or slope.
1119         * Animometer/tests/resources/main.js:
1120         (RampController): Rename the thresholds for clarity. Add a threshold that, if
1121         exceeded, will lower the maximum complexity of the next ramp.
1122         (tune): Relax the cdf check to consider whether the interval definitely falls in
1123         the desired frame length threshold.
1124         (processSamples): Include the profile in the ramp.
1125
1126         Update ramp controller test. Increase the length of the test to 30 seconds, and extend
1127         the interval to 120 ms during sampling. Improve the estimation of the ramp parameters.
1128
1129         * Animometer/developer.html: Change default to 30 seconds, and don't show the progress bar
1130         by default.
1131         * Animometer/resources/runner/animometer.js: Change default to 30 seconds.
1132         * Animometer/tests/resources/main.js: A number of improvements to the ramp controller, in
1133         the order in which they appear in the patch:
1134
1135         - With a longer test length use longer ramps with longer intervals to get more data at each
1136         complexity. Keep the 100 ms interval length during the ramp up phase since we don't need to
1137         spend more time there to find the right order of magnitude, but increase it during the
1138         ramps to 120 ms.
1139         - The ramp linearly interpolates the complexity to render based on its timestamp, but it would
1140         never sample the minimum complexity. Instead of lerping max to min complexity from time
1141         0 to t where t is the ramp length, instead lerp from 0 to (t - intervalSampleLength) so that
1142         we can have at least one interval sample at the min complexity for that ramp.
1143         - Some regression calculations only come out with one line segment rather than the two
1144         we expect. This could be due to a noisy ramp or the ramp's range is too narrow. If that's the
1145         case, influence the minimum complexity of the next ramp towards the lowest bound of 1, so that
1146         we ensure that at least part of the ramp is covering a complexity range that the system can
1147         handle at full 60.
1148         - Remove an assignment to interpolatedFrameLength since that is never subsequently used.
1149
1150         Update the format used to serialize the results for analysis.
1151
1152         Each data point used to have named properties for fields like complexity and frame rate.
1153         In addition the serialized numbers had rounding errors that took up many characters.
1154         Update the format by introducing a new data container called SampleData, which contains a
1155         field map. The map maps a string to an array index. Each data point is an array, so, to
1156         get a stat, use the field map to get the array index into the data point. This allows future
1157         versions to track other data, and reduces the size of the output string by two-thirds.
1158
1159         * Animometer/resources/extensions.js:
1160         (Utilities.toFixedNumber): Add convenience function that truncates the number to a fixed
1161         precision, and converts it back to a number.
1162         (SampleData): New class that contains sample data and a field map that maps properties to
1163         an array index.
1164         (get length): Number of data points.
1165         (addField): Add a field to the field map.
1166         (push): Add a data point.
1167         (sort): Sort the data.
1168         (slice): Return new SampleData object with sliced data.
1169         (forEach): Iterate over the data with provided function.
1170         (createDatum):
1171         (getFieldInDatum): Returns the data point associated with the field name by looking it up
1172         in the field map in the datum provided, which can be the datum object itself (an array) or
1173         an index into the data member variable.
1174         (setFieldInDatum): Sets the data point associated with the field name.
1175         (at): Returns the data point at the provided index.
1176         (toArray): Serializes the data where the field map serves as property names for each point.
1177
1178         * Animometer/resources/debug-runner/graph.js:
1179         (updateGraphData): Remove unused _testData. Convert the data to the old array format for the
1180         graph to use, since the old format was much easier to work with when displaying the graphs.
1181         (onGraphTypeChanged): For some controllers, no alternative score or mean is provided.
1182         * Animometer/resources/runner/animometer.css:
1183         * Animometer/resources/runner/animometer.js: Refactor to use SampleData. Update JSON output
1184         to only go to 3 digits of precision for purposes of reducing the data size.
1185         * Animometer/resources/strings.js: Add new strings to put into the field maps.
1186         * Animometer/tests/resources/main.js: Refactor to use SampleData.
1187
1188         * Animometer/developer.html:
1189         * Animometer/index.html: Restructure results table for both pages. Add charset attribute to
1190         tests.js include.
1191         * Animometer/resources/debug-runner/animometer.css: Clear out styles from release runner.
1192         * Animometer/resources/debug-runner/graph.js:
1193         (onGraphTypeChanged): Update score and mean if bootstrap results are available from the
1194         controller, since not all controllers do bootstrapping.
1195         * Animometer/resources/debug-runner/tests.js: Update header text.
1196         * Animometer/resources/runner/animometer.css: Include confidence interval in results.
1197         * Animometer/resources/runner/animometer.js:
1198         (ResultsTable._addHeader): Header contents can be HTML, so use innerHTML instead.
1199         (ResultsTable._addBody): Add tbody element.
1200         (ResultsTable._addTest): Allow a data cell to invoke a JS function to get its contents.
1201         (window.benchmarkController.showResults): Add table that includes tests' confidence intervals.
1202         * Animometer/resources/runner/tests.js:
1203         (Headers.details.text): Add new details table that includes bootstrap confidence interval.
1204         The interval can be asymmetric, but for simplicity, report the maximum deviation percentage
1205         on either side of the bootstrap median.
1206         * Animometer/resources/statistics.js:
1207         (bootstrap): Include the confidence percentage in the return object.
1208
1209         Report canvas size in results.
1210
1211         * Animometer/developer.html: Add markup to indicate whether a small, medium, or large
1212         canvas was used.
1213         * Animometer/index.html: Ditto.
1214         * Animometer/resources/debug-runner/animometer.js: Call determineCanvasSize().
1215         * Animometer/resources/runner/animometer.css: Update styles to set the canvas based on the
1216         body class size.
1217         * Animometer/resources/runner/animometer.js:
1218         (window.benchmarkController.initialize): Update styles to set the canvas based on the
1219         body class size.
1220         (window.benchmarkController.determineCanvasSize): Run various media queries and set the body
1221         class based on the size of the device.
1222
1223         * Animometer/developer.html: Refactor to include the main CSS file, and redo
1224         the layout so that it doesn't rely on flexbox.
1225         * Animometer/resources/debug-runner/animometer.css:
1226         * Animometer/resources/debug-runner/animometer.js:
1227         (updateDisplay): Since various parts of the script alter the body class, we can't
1228         replace the className directly. Instead, remove all display-based values and then add
1229         the one that was selected.
1230         * Animometer/resources/debug-runner/graph.js:
1231         (updateGraphData): To set the size of the graph, use window.innerHeight.
1232         * Animometer/resources/runner/animometer.js:
1233         (window.sectionsManager.showSection): Since various parts of the script alter the body
1234         class, we can't replace the className directly. Remove all of the section classes
1235         individually and then add the one desired.
1236         * Animometer/tests/resources/stage.css: Remove -apple-system as a font to use in the
1237         stage.
1238
1239 2016-06-12  Filip Pizlo  <fpizlo@apple.com>
1240
1241         Fix round-down goof in Air.js's ShuffleCustom.forEachArg
1242         https://bugs.webkit.org/show_bug.cgi?id=158674
1243
1244         Reviewed by Michael Saboff.
1245         
1246         x / 3 * 3 is not how you round down to multiples of 3 in JavaScript. You need to do
1247         Math.floor(x / 3) * 3 instead.
1248         
1249         This is a benign change, because having extra arguments to a Shuffle is not something we
1250         actually take advantage of yet. But I think it's best to match the original C++ code's
1251         intent.
1252
1253         * Air.js/custom.js:
1254         (const.ShuffleCustom.forEachArg):
1255
1256 2016-06-11  Filip Pizlo  <fpizlo@apple.com>
1257
1258         Unreviewed, skip Air.js. It got unskipped when I renamed it from JSAir.
1259
1260         * Skipped:
1261
1262 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1263
1264         Make it easy to package Air.js
1265         https://bugs.webkit.org/show_bug.cgi?id=158652
1266
1267         Reviewed by Benjamin Poulain.
1268         
1269         If you want to give Air.js to someone, you can now do:
1270         
1271         cd PerformanceTests/Air.js
1272         ./make_dist.sh
1273         
1274         This will create Air.js.tar.gz and a directory called Air.js. If you want to send someone
1275         the benchmark, just send either of those things and tell them to look at index.html.
1276         
1277         You may have to edit make_dist.sh for wherever you have Markdown.pl. You can get that from
1278         https://daringfireball.net/projects/markdown/
1279
1280         * Air.js/make_dist.sh: Added.
1281         * Air.js/test.html:
1282
1283 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1284
1285         Unreviewed, make the link to test.html into an actual link.
1286
1287         * Air.js/README.md:
1288
1289 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1290
1291         Fix some text in Air.js/README.md
1292         https://bugs.webkit.org/show_bug.cgi?id=158650
1293
1294         Reviewed by Benjamin Poulain.
1295         
1296         I read the text again and found bugs:
1297         
1298         - We never actually say how to run the benchmark. This change adds a blurb about how to run
1299           it.
1300
1301         - We both say that allocateStack is responsible for the bulk of the running time and that
1302           we haven't measured where the bulk of the time is spent. This changes the text to say that
1303           it was a goal to make allocateStack be the hottest part of the benchmark, but that we did
1304           not measure this.
1305
1306         * Air.js/README.md:
1307
1308 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1309
1310         Air.js should have some documentation
1311         https://bugs.webkit.org/show_bug.cgi?id=158648
1312
1313         Reviewed by Keith Miller.
1314         
1315         I want to be able to point people at a document if they want to know more about this
1316         benchmark.
1317
1318         * Air.js/README.md: Added.
1319
1320 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1321
1322         Rename JSAir to Air.js.
1323
1324         Rubber stamped by Mark Lam.
1325
1326         * Air.js: Copied from PerformanceTests/JSAir.
1327         * Air.js/airjs-tests.yaml: Copied from PerformanceTests/JSAir/jsair-tests.yaml.
1328         * Air.js/benchmark.js: Copied from PerformanceTests/JSAir/benchmark.js.
1329         (Benchmark):
1330         * Air.js/jsair-tests.yaml: Removed.
1331         * Air.js/payload-airjs-ACLj8C.js: Copied from PerformanceTests/JSAir/payload-jsair-ACLj8C.js.
1332         * Air.js/payload-jsair-ACLj8C.js: Removed.
1333         * Air.js/stress-test.js: Copied from PerformanceTests/JSAir/stress-test.js.
1334         * Air.js/test.html: Copied from PerformanceTests/JSAir/test.html.
1335         * Air.js/test.js: Copied from PerformanceTests/JSAir/test.js.
1336         * JSAir: Removed.
1337         * JSAir/all.js: Removed.
1338         * JSAir/allocate_stack.js: Removed.
1339         * JSAir/arg.js: Removed.
1340         * JSAir/basic_block.js: Removed.
1341         * JSAir/benchmark.js: Removed.
1342         * JSAir/code.js: Removed.
1343         * JSAir/custom.js: Removed.
1344         * JSAir/frequented_block.js: Removed.
1345         * JSAir/insertion_set.js: Removed.
1346         * JSAir/inst.js: Removed.
1347         * JSAir/jsair-tests.yaml: Removed.
1348         * JSAir/liveness.js: Removed.
1349         * JSAir/opcode.js: Removed.
1350         * JSAir/payload-gbemu-executeIteration.js: Removed.
1351         * JSAir/payload-imaging-gaussian-blur-gaussianBlur.js: Removed.
1352         * JSAir/payload-jsair-ACLj8C.js: Removed.
1353         * JSAir/payload-typescript-scanIdentifier.js: Removed.
1354         * JSAir/reg.js: Removed.
1355         * JSAir/stack_slot.js: Removed.
1356         * JSAir/stress-test.js: Removed.
1357         * JSAir/strip-hash.rb: Removed.
1358         * JSAir/symbols.js: Removed.
1359         * JSAir/test.html: Removed.
1360         * JSAir/test.js: Removed.
1361         * JSAir/tmp.js: Removed.
1362         * JSAir/tmp_base.js: Removed.
1363         * JSAir/util.js: Removed.
1364
1365 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1366
1367         JSC Stress Test failing: jsair-tests.yaml/test.js.ftl-eager-no-cjit
1368         https://bugs.webkit.org/show_bug.cgi?id=158571
1369
1370         Reviewed by Keith Miller.
1371         
1372         Introduce a different harness for run-jsc-stress-tests, which runs for a minimum of 10
1373         iterations and then tries to do any number of "bonus" iterations until it's been running for
1374         two seconds. Since this is the sort of test that isn't really meant to stress anything in
1375         particular, I think it's OK if it is time-limited in this way. The worst case is that some
1376         of its failures will be flaky, but I think that they would have been flaky anyway given the
1377         complexity of the test.
1378
1379         * JSAir/benchmark.js:
1380         (Benchmark):
1381         (Benchmark.prototype.runIteration):
1382         (benchmark): Deleted.
1383         * JSAir/jsair-tests.yaml:
1384         * JSAir/stress-test.js: Added.
1385         (preciseTime):
1386         * JSAir/test.html:
1387         * JSAir/test.js:
1388
1389 2016-06-09  Filip Pizlo  <fpizlo@apple.com>
1390
1391         Unreviewed, teach the perf bots not to run JSAir.
1392
1393         * Skipped:
1394
1395 2016-06-08  Filip Pizlo  <fpizlo@apple.com>
1396
1397         Use more ES6 features in JSAir
1398         https://bugs.webkit.org/show_bug.cgi?id=158497
1399
1400         Reviewed by Keith Miller.
1401         
1402         This improves JSAir with the following ES6 features suggested by JoePeck:
1403         
1404         - String interpolation.
1405         - Destructuring inside PatchCustom.
1406         - Default arguments.
1407         
1408         All of these things are on hot paths.
1409         
1410         Note that I didn't use string interpolation everywhere that I could, only in those places
1411         where it made the code more readable. In Ruby, I used the style that if the interpolation
1412         expression has any non-trivial stuff (like a ternary operator, a chain of calls, or embedded
1413         strings) then it's better to use regular strcat. I think that's what I carried over to here.
1414         
1415         Note that the previous change (Add result validation to JSAir) also made the Proxy code not
1416         dead, though it's not necessarily on the hot path. The Proxy isn't called into frequently
1417         but it's used from a function that is otherwise hot, so if calling into the Proxy prevents
1418         that function from being optimized then it will hurt so good.
1419         
1420         I also reenabled tail calls in a few places.
1421         
1422         This change doesn't seem to change the performance of the benchmark for us. That's expected
1423         since these ES6 features are cheap. Note that this claim doesn't include Proxy, which was
1424         added in a separate change and that change did make the benchmark overall more expensive.
1425
1426         * JSAir/allocate_stack.js:
1427         (allocateStack):
1428         * JSAir/arg.js:
1429         (Arg.createBitImm64):
1430         (Arg.createAddr):
1431         (Arg.createStack):
1432         (Arg.logScale):
1433         (Arg.createIndex):
1434         * JSAir/basic_block.js:
1435         (BasicBlock.get headerString):
1436         (BasicBlock.prototype.get if):
1437         (BasicBlock):
1438         * JSAir/benchmark.js:
1439         (benchmark):
1440         * JSAir/code.js:
1441         (Code):
1442         (Code.prototype.addBlock):
1443         (Code.prototype.addStackSlot):
1444         (Code.prototype.newTmp):
1445         (Code.prototype.get size):
1446         (Code.prototype.get blocks):
1447         (Code.prototype.get stackSlots):
1448         (Code.prototype.tmps):
1449         (Code.prototype.get callArgAreaSize):
1450         (Code.prototype.toString):
1451         * JSAir/custom.js:
1452         (const.PatchCustom.forEachArg):
1453         * JSAir/inst.js:
1454         (Inst):
1455         * JSAir/reg.js:
1456         (Reg.prototype.toString):
1457         * JSAir/util.js:
1458         (symbolName):
1459         (lowerSymbolName):
1460
1461 2016-06-07  Filip Pizlo  <fpizlo@apple.com>
1462
1463         Add result validation to JSAir
1464         https://bugs.webkit.org/show_bug.cgi?id=158493
1465
1466         Reviewed by Saam Barati.
1467         
1468         This adds the ability to hash a Code in a way that matches the C++ code's hashing of Code.
1469         This allows us to check if the Code that JSAir sees is the code that C++ saw. We use this to
1470         check the Code before and after allocateStack, and compare against hashes we got from C++.
1471         
1472         Doing this uncovered bugs. roundUpToMultipleOf wasn't doing anything. allocateStack was not
1473         allocating things correctly because I was concatting a Set to an Array, which doesn't really
1474         work. Now these bugs are fixed.
1475         
1476         The checking step adds to the running time so I reduced the number of iterations. The
1477         benchmark spends a decent amount of its time computing Code hashes; I think it's around 1/3
1478         total. This is probably OK. It's better to verify the results even if the running time is
1479         not all in the "core" of the algorithm.
1480         
1481         Also add a run-jsc-stress-tests yaml file to allow this to run as a test.
1482
1483         * JSAir/allocate_stack.js:
1484         * JSAir/arg.js:
1485         (Arg.createImm):
1486         (Arg.createBigImm):
1487         (Arg.createBitImm):
1488         (Arg.createBitImm64):
1489         (Arg.createWidth):
1490         (Arg.createSpecial):
1491         (Arg.prototype.get kind):
1492         (Arg.prototype.get isTmp):
1493         (Arg.prototype.get isImm):
1494         (Arg.prototype.get isSomeImm):
1495         (Arg.prototype.get isSomeBigImm):
1496         (Arg.prototype.get isCondition):
1497         (Arg.prototype.get isWidth):
1498         (Arg.prototype.get isSpecial):
1499         (Arg.prototype.get isAlive):
1500         (Arg.prototype.get tmp):
1501         (Arg.prototype.get value):
1502         (Arg.prototype.get lowValue):
1503         (Arg.prototype.get highValue):
1504         (Arg.prototype.get base):
1505         (Arg.prototype.get isGP):
1506         (Arg.prototype.get isFP):
1507         (Arg.prototype.isValidForm):
1508         (Arg.prototype.get isInvertible):
1509         (Arg.kindCode):
1510         (Arg.prototype.hash):
1511         (Arg.prototype.toString):
1512         (Arg):
1513         * JSAir/basic_block.js:
1514         (BasicBlock.get successorBlocks):
1515         * JSAir/benchmark.js:
1516         (benchmark):
1517         * JSAir/code.js:
1518         (Code.prototype.setFrameSize):
1519         (Code.prototype.hash):
1520         (Code.prototype.toString):
1521         (Code):
1522         * JSAir/inst.js:
1523         (Inst.prototype.get hasNonArgEffects):
1524         (Inst.prototype.hash):
1525         (Inst.prototype.toString):
1526         (Inst):
1527         * JSAir/jsair-tests.yaml: Added.
1528         * JSAir/opcode.js:
1529         (Inst_forEachArg):
1530         (Inst_hasNonArgEffects):
1531         (opcodeCode):
1532         * JSAir/payload-gbemu-executeIteration.js:
1533         (createPayloadGbemuExecuteIteration):
1534         * JSAir/payload-imaging-gaussian-blur-gaussianBlur.js:
1535         (createPayloadImagingGaussianBlurGaussianBlur):
1536         * JSAir/payload-jsair-ACLj8C.js:
1537         (createPayloadJSAirACLj8C):
1538         * JSAir/payload-typescript-scanIdentifier.js:
1539         (createPayloadTypescriptScanIdentifier):
1540         * JSAir/reg.js:
1541         (Reg.prototype.get isReg):
1542         (Reg.prototype.hash):
1543         (Reg.prototype.toString):
1544         * JSAir/stack_slot.js:
1545         (StackSlot.prototype.setOffsetFromFP):
1546         (StackSlot.prototype.hash):
1547         (StackSlot.prototype.toString):
1548         * JSAir/symbols.js:
1549         (relCondCode):
1550         (resCondCode):
1551         (doubleCondCode):
1552         * JSAir/test.html:
1553         * JSAir/tmp.js:
1554         (Tmp.prototype.get isReg):
1555         (Tmp.prototype.hash):
1556         (Tmp.prototype.toString):
1557         * JSAir/util.js:
1558         (roundUpToMultipleOf):
1559         (symbolName):
1560
1561 2016-06-07  Filip Pizlo  <fpizlo@apple.com>
1562
1563         Implement Air::allocateStack() in ES6 to see how much of a bad idea that is
1564         https://bugs.webkit.org/show_bug.cgi?id=158318
1565
1566         Reviewed by Saam Barati.
1567         
1568         This adds a new benchmark for us to play with called JSAir. It's a complete ES6
1569         implementation of Air's allocateStack() phase along with all of Air needed to run that
1570         phase. This includes things like stack slots, registers, temporaries, basic blocks,
1571         instructions, and all of the code for iterating over, inspecting, and modifying those
1572         things.
1573         
1574         To make this work, JSC can now dump Air just before allocateStack() in the form of JS code
1575         that creates a Code object that matches exactly what C++ Air saw. This benchmark comprises
1576         four Air IRs:
1577         
1578         - Octane/gbemu's largest function, executeIteration.
1579         - Kraken/imaging-gaussian-blur's largest function in OSR entry mode, gaussuanBlur.
1580         - Octane/typescript's largest function that is actually hot, scanIdentifier.
1581         - JSAir's largest hot function, which is anonymous, so we call it by its hash (ACLj8C).
1582         
1583         This runs in about 2 seconds on my machine in JSC trunk. It includes both a commandline
1584         harness and a web harness. JSAir is almost exactly 100x slower in ES6 in WebKit than the
1585         C++ Air::allocateStack() phase on which it is based.
1586         
1587         JSAir uses the following ES6 features:
1588         
1589         - Symbol.
1590         - for-of.
1591         - arrow functions.
1592         - Map/Set.
1593         - let/const.
1594         - classes.
1595         
1596         All of these things are used in anger and should end up on the hot path. There is also code
1597         that uses Proxies, but it ends up being dead.
1598         
1599         We can improve this even more:
1600         I still need to add result validation: https://bugs.webkit.org/show_bug.cgi?id=158493
1601         I want to make it use more ES6 features: https://bugs.webkit.org/show_bug.cgi?id=158497
1602         
1603         * JSAir: Added.
1604         * JSAir/all.js: Added.
1605         * JSAir/allocate_stack.js: Added.
1606         (allocateStack.attemptAssignment):
1607         (allocateStack.assign):
1608         (allocateStack.interfere):
1609         (allocateStack.):
1610         (allocateStack):
1611         * JSAir/arg.js: Added.
1612         (Arg):
1613         (Arg.isAnyUse):
1614         (Arg.isColdUse):
1615         (Arg.isWarmUse):
1616         (Arg.cooled):
1617         (Arg.isEarlyUse):
1618         (Arg.isLateUse):
1619         (Arg.isAnyDef):
1620         (Arg.isEarlyDef):
1621         (Arg.isLateDef):
1622         (Arg.isZDef):
1623         (Arg.typeForB3Type):
1624         (Arg.widthForB3Type):
1625         (Arg.conservativeWidth):
1626         (Arg.minimumWidth):
1627         (Arg.bytes):
1628         (Arg.widthForBytes):
1629         (Arg.createTmp):
1630         (Arg.fromReg):
1631         (Arg.createImm):
1632         (Arg.createBigImm):
1633         (Arg.createBitImm):
1634         (Arg.createBitImm64):
1635         (Arg.createAddr):
1636         (Arg.createStack):
1637         (Arg.createCallArg):
1638         (Arg.createStackAddr):
1639         (Arg.isValidScale):
1640         (Arg.logScale):
1641         (Arg.createIndex):
1642         (Arg.createRelCond):
1643         (Arg.createResCond):
1644         (Arg.createDoubleCond):
1645         (Arg.createWidth):
1646         (Arg.prototype.get kind):
1647         (Arg.prototype.get isTmp):
1648         (Arg.prototype.get isImm):
1649         (Arg.prototype.get isBigImm):
1650         (Arg.prototype.get isBitImm):
1651         (Arg.prototype.get isBitImm64):
1652         (Arg.prototype.get isSomeImm):
1653         (Arg.prototype.get isAddr):
1654         (Arg.prototype.get isStack):
1655         (Arg.prototype.get isCallArg):
1656         (Arg.prototype.get isIndex):
1657         (Arg.prototype.get isMemory):
1658         (Arg.prototype.get isStackMemory):
1659         (Arg.prototype.get isRelCond):
1660         (Arg.prototype.get isResCond):
1661         (Arg.prototype.get isDoubleCond):
1662         (Arg.prototype.get isCondition):
1663         (Arg.prototype.get isWidth):
1664         (Arg.prototype.get isAlive):
1665         (Arg.prototype.get tmp):
1666         (Arg.prototype.get value):
1667         (Arg.prototype.get base):
1668         (Arg.prototype.get hasOffset):
1669         (Arg.prototype.get offset):
1670         (Arg.prototype.get stackSlot):
1671         (Arg.prototype.get index):
1672         (Arg.prototype.get scale):
1673         (Arg.prototype.get logScale):
1674         (Arg.prototype.get width):
1675         (Arg.prototype.get isGPTmp):
1676         (Arg.prototype.get isFPTmp):
1677         (Arg.prototype.get isGP):
1678         (Arg.prototype.get isFP):
1679         (Arg.prototype.get hasType):
1680         (Arg.prototype.get type):
1681         (Arg.prototype.isType):
1682         (Arg.prototype.isCompatibleType):
1683         (Arg.prototype.get isGPR):
1684         (Arg.prototype.get gpr):
1685         (Arg.prototype.get isFPR):
1686         (Arg.prototype.get fpr):
1687         (Arg.prototype.get isReg):
1688         (Arg.prototype.get reg):
1689         (Arg.isValidImmForm):
1690         (Arg.isValidBitImmForm):
1691         (Arg.isValidBitImm64Form):
1692         (Arg.isValidAddrForm):
1693         (Arg.isValidIndexForm):
1694         (Arg.prototype.isValidForm):
1695         (Arg.prototype.forEachTmpFast):
1696         (Arg.prototype.usesTmp):
1697         (Arg.prototype.forEachTmp):
1698         (Arg.prototype.is):
1699         (Arg.prototype.as):
1700         (Arg.prototype.forEachFast):
1701         (Arg.prototype.forEach):
1702         (Arg.extract):
1703         (Arg.forEachFast):
1704         (Arg.forEach):
1705         (Arg.prototype.get condition):
1706         (Arg.prototype.get isInvertible):
1707         (Arg.prototype.toString):
1708         * JSAir/basic_block.js: Added.
1709         (BasicBlock):
1710         (BasicBlock.prototype.get index):
1711         (BasicBlock.prototype.get size):
1712         (BasicBlock.prototype.Symbol.iterator):
1713         (BasicBlock.prototype.at):
1714         (BasicBlock.get last):
1715         (BasicBlock.get insts):
1716         (BasicBlock.get numSuccessors):
1717         (BasicBlock.get successors):
1718         (BasicBlock.get successorBlocks.):
1719         (BasicBlock.get successorBlocks):
1720         (BasicBlock.set get numPredecessors):
1721         (BasicBlock.get predecessors):
1722         (BasicBlock.get frequency):
1723         (BasicBlock.get headerString):
1724         (BasicBlock.get footerString):
1725         * JSAir/benchmark.js: Added.
1726         (benchmark):
1727         * JSAir/code.js: Added.
1728         (Code):
1729         (Code.prototype.addBlock):
1730         (Code.prototype.addStackSlot):
1731         (Code.prototype.newTmp):
1732         (Code.prototype.get size):
1733         (Code.prototype.at):
1734         (Code.prototype.Symbol.iterator):
1735         (Code.prototype.get blocks):
1736         (Code.prototype.get stackSlots):
1737         (Code.prototype.tmps):
1738         (Code.prototype.get callArgAreaSize):
1739         (Code.prototype.requestCallArgAreaSize):
1740         (Code.prototype.get frameSize):
1741         (Code.prototype.setFrameSize):
1742         (Code.prototype.toString):
1743         * JSAir/custom.js: Added.
1744         (const.ShuffleCustom.forEachArg):
1745         (const.ShuffleCustom.hasNonArgNonControlEffects):
1746         (const.PatchCustom.forEachArg):
1747         (const.PatchCustom.hasNonArgNonControlEffects):
1748         (const.CCallCustom.forEachArg):
1749         (const.CCallCustom.hasNonArgNonControlEffects):
1750         (const.ColdCCallCustom.forEachArg):
1751         (const.ColdCCallCustom.hasNonArgNonControlEffects):
1752         * JSAir/frequented_block.js: Added.
1753         (FrequentedBlock):
1754         (FrequentedBlock.prototype.toString):
1755         * JSAir/insertion_set.js: Added.
1756         (Insertion):
1757         (Insertion.prototype.get index):
1758         (Insertion.prototype.get element):
1759         (Insertion.prototype.lessThan):
1760         (InsertionSet):
1761         (InsertionSet.prototype.appendInsertion):
1762         (InsertionSet.prototype.append):
1763         (InsertionSet.prototype.execute):
1764         * JSAir/inst.js: Added.
1765         (Inst):
1766         (Inst.prototype.append):
1767         (Inst.prototype.clear):
1768         (Inst.prototype.get opcode):
1769         (Inst.prototype.get args):
1770         (Inst.prototype.visitArg):
1771         (Inst.prototype.forEachTmpFast):
1772         (Inst.prototype.forEachArg):
1773         (Inst.prototype.forEachTmp):
1774         (Inst.prototype.forEach):
1775         (Inst.forEachDef):
1776         (Inst.forEachDefWithExtraClobberedRegs):
1777         (Inst.prototype.get hasNonArgEffects):
1778         (Inst.prototype.toString):
1779         * JSAir/liveness.js: Added.
1780         (Liveness):
1781         (Liveness.prototype.get thing):
1782         (Liveness.prototype.get code):
1783         (Liveness.prototype.get liveAtHead):
1784         (Liveness.prototype.get liveAtTail):
1785         (Liveness.prototype.localCalc.LocalCalc):
1786         (Liveness.prototype.localCalc.LocalCalc.prototype.get liveSet):
1787         (Liveness.prototype.localCalc.LocalCalc.prototype.execute):
1788         (Liveness.prototype.localCalc):
1789         * JSAir/opcode.js: Added.
1790         (Inst_forEachArg):
1791         (Inst_hasNonArgEffects):
1792         * JSAir/payload-gbemu-executeIteration.js: Added.
1793         (createPayloadGbemuExecuteIteration):
1794         * JSAir/payload-imaging-gaussian-blur-gaussianBlur.js: Added.
1795         (createPayloadImagingGaussianBlurGaussianBlur):
1796         * JSAir/payload-jsair-ACLj8C.js: Added.
1797         (createPayloadJSAirACLj8C):
1798         * JSAir/payload-typescript-scanIdentifier.js: Added.
1799         (createPayloadTypescriptScanIdentifier):
1800         * JSAir/reg.js: Added.
1801         (Reg):
1802         (Reg.fromReg):
1803         (Reg.prototype.get index):
1804         (Reg.prototype.get type):
1805         (Reg.prototype.get name):
1806         (Reg.prototype.get isCalleeSave):
1807         (Reg.prototype.get isReg):
1808         (Reg.prototype.toString):
1809         (Reg.extract):
1810         (Reg.forEachFast):
1811         (Reg.forEach):
1812         (newGPR):
1813         (Reg.gprs.Reg.fprs.Reg.calleeSaveGPRs.Reg.calleeSaveFPRs.Reg.calleeSaves):
1814         * JSAir/stack_slot.js: Added.
1815         (StackSlot):
1816         (StackSlot.prototype.get byteSize):
1817         (StackSlot.prototype.get kind):
1818         (StackSlot.prototype.get isLocked):
1819         (StackSlot.prototype.get isSpill):
1820         (StackSlot.prototype.get index):
1821         (StackSlot.prototype.ensureSize):
1822         (StackSlot.prototype.get alignment):
1823         (StackSlot.prototype.get offsetFromFP):
1824         (StackSlot.prototype.setOffsetFromFP):
1825         (StackSlot.prototype.toString):
1826         (StackSlot.extract):
1827         (StackSlot.forEachFast):
1828         (StackSlot.forEach):
1829         * JSAir/symbols.js: Added.
1830         * JSAir/test.html: Added.
1831         * JSAir/test.js: Added.
1832         * JSAir/tmp.js: Added.
1833         (Tmp):
1834         (Tmp.fromReg):
1835         (Tmp.prototype.get index):
1836         (Tmp.prototype.get type):
1837         (Tmp.prototype.get isReg):
1838         (Tmp.prototype.toString):
1839         (Tmp.extract):
1840         (Tmp.forEachFast):
1841         (Tmp.forEach):
1842         * JSAir/tmp_base.js: Added.
1843         (TmpBase.prototype.get isGP):
1844         (TmpBase.prototype.get isFP):
1845         (TmpBase.prototype.get isGPR):
1846         (TmpBase.prototype.get isFPR):
1847         (TmpBase.prototype.get reg):
1848         (TmpBase.prototype.get gpr):
1849         (TmpBase.prototype.get fpr):
1850         (TmpBase):
1851         * JSAir/util.js: Added.
1852         (isRepresentableAsInt32):
1853         (addIndexed):
1854         (roundUpToMultipleOf):
1855         (symbolName):
1856         (mergeIntoSet):
1857         (nonEmptyRangesOverlap):
1858         (rangesOverlap):
1859         (removeAllMatching):
1860         (swap):
1861         (bubble):
1862         (bubbleSort):
1863         (currentTime):
1864         (else.currentTime):
1865
1866 2016-05-26  Yusuke Suzuki  <utatane.tea@gmail.com>
1867
1868         [JSC] Allow JSBench to use precise time
1869         https://bugs.webkit.org/show_bug.cgi?id=158050
1870
1871         Reviewed by Geoffrey Garen.
1872
1873         * JSBench/amazon-chrome-win/urem.js:
1874         (else.window.performance.window.performance.now.currentTimeInMS):
1875         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1876         (else.else.currentTimeInMS):
1877         (onload.cb):
1878         (onload):
1879         * JSBench/amazon-chrome/urem.js:
1880         (else.window.performance.window.performance.now.currentTimeInMS):
1881         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1882         (else.else.currentTimeInMS):
1883         (onload.cb):
1884         (onload):
1885         * JSBench/amazon-firefox-win/urm.js:
1886         (else.window.performance.window.performance.now.currentTimeInMS):
1887         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1888         (else.else.currentTimeInMS):
1889         (onload.cb):
1890         (onload):
1891         * JSBench/amazon-firefox/urm.js:
1892         (else.window.performance.window.performance.now.currentTimeInMS):
1893         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1894         (else.else.currentTimeInMS):
1895         (onload.cb):
1896         (onload):
1897         * JSBench/amazon-safari/urem.js:
1898         (else.window.performance.window.performance.now.currentTimeInMS):
1899         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1900         (else.else.currentTimeInMS):
1901         (onload.cb):
1902         (onload):
1903         * JSBench/facebook-chrome-win/urem.js:
1904         (else.window.performance.window.performance.now.currentTimeInMS):
1905         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1906         (else.else.currentTimeInMS):
1907         (onload.cb):
1908         (onload):
1909         * JSBench/facebook-chrome/urem.js:
1910         (else.window.performance.window.performance.now.currentTimeInMS):
1911         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1912         (else.else.currentTimeInMS):
1913         (onload.cb):
1914         (onload):
1915         * JSBench/facebook-firefox-win/urem.js:
1916         (else.window.performance.window.performance.now.currentTimeInMS):
1917         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1918         (else.else.currentTimeInMS):
1919         (onload.cb):
1920         (onload):
1921         * JSBench/facebook-firefox/urem.js:
1922         (else.window.performance.window.performance.now.currentTimeInMS):
1923         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1924         (else.else.currentTimeInMS):
1925         (onload.cb):
1926         (onload):
1927         * JSBench/facebook-safari/urem.js:
1928         (else.window.performance.window.performance.now.currentTimeInMS):
1929         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1930         (else.else.currentTimeInMS):
1931         (onload.cb):
1932         (onload):
1933         * JSBench/google-chrome-win/urem.js:
1934         (else.window.performance.window.performance.now.currentTimeInMS):
1935         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1936         (else.else.currentTimeInMS):
1937         (onload.cb):
1938         (onload):
1939         * JSBench/google-chrome/urem.js:
1940         (else.window.performance.window.performance.now.currentTimeInMS):
1941         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1942         (else.else.currentTimeInMS):
1943         (onload.cb):
1944         (onload):
1945         * JSBench/google-firefox-win/urem.js:
1946         (else.window.performance.window.performance.now.currentTimeInMS):
1947         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1948         (else.else.currentTimeInMS):
1949         (onload.cb):
1950         (onload):
1951         * JSBench/google-firefox/uem.js:
1952         (else.window.performance.window.performance.now.currentTimeInMS):
1953         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1954         (else.else.currentTimeInMS):
1955         (onload.cb):
1956         (onload):
1957         * JSBench/google-safari/urem.js:
1958         (else.window.performance.window.performance.now.currentTimeInMS):
1959         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1960         (else.else.currentTimeInMS):
1961         (onload.cb):
1962         (onload):
1963         * JSBench/harness.js:
1964         (runBenchmark.window.currentTimeInMS):
1965         (runBenchmark.else.window.currentTimeInMS):
1966         * JSBench/twitter-chrome-win/rem.js:
1967         (else.window.performance.window.performance.now.currentTimeInMS):
1968         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1969         (else.else.currentTimeInMS):
1970         (onload.cb):
1971         (onload):
1972         * JSBench/twitter-chrome/urem.js:
1973         (else.window.performance.window.performance.now.currentTimeInMS):
1974         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1975         (else.else.currentTimeInMS):
1976         (onload.cb):
1977         (onload):
1978         * JSBench/twitter-firefox-win/urem.js:
1979         (else.window.performance.window.performance.now.currentTimeInMS):
1980         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1981         (else.else.currentTimeInMS):
1982         (onload.cb):
1983         (onload):
1984         * JSBench/twitter-firefox/urem.js:
1985         (else.window.performance.window.performance.now.currentTimeInMS):
1986         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1987         (else.else.currentTimeInMS):
1988         (onload.cb):
1989         (onload):
1990         * JSBench/twitter-safari/urem.js:
1991         (else.window.performance.window.performance.now.currentTimeInMS):
1992         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1993         (else.else.currentTimeInMS):
1994         (onload.cb):
1995         (onload):
1996         * JSBench/yahoo-chrome-win/urem.js:
1997         (else.window.performance.window.performance.now.currentTimeInMS):
1998         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1999         (else.else.currentTimeInMS):
2000         (onload.cb):
2001         (onload):
2002         * JSBench/yahoo-chrome/urem.js:
2003         (else.window.performance.window.performance.now.currentTimeInMS):
2004         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2005         (else.else.currentTimeInMS):
2006         (onload.cb):
2007         (onload):
2008         * JSBench/yahoo-firefox-win/urem.js:
2009         (else.window.performance.window.performance.now.currentTimeInMS):
2010         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2011         (else.else.currentTimeInMS):
2012         (onload.cb):
2013         (onload):
2014         * JSBench/yahoo-firefox/urem.js:
2015         (else.window.performance.window.performance.now.currentTimeInMS):
2016         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2017         (else.else.currentTimeInMS):
2018         (onload.cb):
2019         (onload):
2020         * JSBench/yahoo-safari/urem.js:
2021         (else.window.performance.window.performance.now.currentTimeInMS):
2022         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2023         (else.else.currentTimeInMS):
2024         (onload.cb):
2025         (onload):
2026
2027 2016-05-25  Geoffrey Garen  <ggaren@apple.com>
2028
2029         replaceable own properties seem to ignore replacement after property caching
2030         https://bugs.webkit.org/show_bug.cgi?id=158091
2031
2032         Reviewed by Darin Adler.
2033
2034         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
2035         * MallocBench/MallocBench/Benchmark.cpp:
2036         * MallocBench/MallocBench/Interpreter.cpp:
2037         (Interpreter::doMallocOp):
2038         * MallocBench/MallocBench/Interpreter.h:
2039         * MallocBench/MallocBench/fastMallocLog.63316.ops: Added.
2040         * MallocBench/MallocBench/jetstream.cpp: Added.
2041         (benchmark_jetstream):
2042         * MallocBench/MallocBench/jetstream.h: Added.
2043
2044 2016-05-25  Keith Miller  <keith_miller@apple.com>
2045
2046         Unreviewed, add JSBench to the skipped list for now since it doesn't
2047         work currently.
2048
2049         * Skipped:
2050
2051 2016-05-17  Keith Miller  <keith_miller@apple.com>
2052         We should have JSBench in PerformanceTests
2053         https://bugs.webkit.org/show_bug.cgi?id=157952
2054
2055         Rubber-stamped by Saam Barati.
2056
2057         There are some slight changes to the layout of the test directory
2058         to make it work nicely with run-jsc-benchmarks. Before JSBench
2059         had each of the browser specific sub-tests in a sub-directory.
2060         These have been flattened e.g. amazon/safari/ has become
2061         amazon-safari/.
2062
2063         * JSBench/amazon-chrome-win/urem.html: Added.
2064         * JSBench/amazon-chrome-win/urem.js: Added.
2065         * JSBench/amazon-chrome/urem.html: Added.
2066         * JSBench/amazon-chrome/urem.js: Added.
2067         * JSBench/amazon-firefox-win/urm.html: Added.
2068         * JSBench/amazon-firefox-win/urm.js: Added.
2069         * JSBench/amazon-firefox/urm.html: Added.
2070         * JSBench/amazon-firefox/urm.js: Added.
2071         * JSBench/amazon-safari/urem.html: Added.
2072         * JSBench/amazon-safari/urem.js: Added.
2073         * JSBench/browsercheck.js: Added.
2074         * JSBench/facebook-chrome-win/urem.html: Added.
2075         * JSBench/facebook-chrome-win/urem.js: Added.
2076         * JSBench/facebook-chrome/urem.html: Added.
2077         * JSBench/facebook-chrome/urem.js: Added.
2078         * JSBench/facebook-firefox-win/urem.html: Added.
2079         * JSBench/facebook-firefox-win/urem.js: Added.
2080         * JSBench/facebook-firefox/urem.html: Added.
2081         * JSBench/facebook-firefox/urem.js: Added.
2082         * JSBench/facebook-safari/urem.html: Added.
2083         * JSBench/facebook-safari/urem.js: Added.
2084         * JSBench/google-chrome-win/urem.html: Added.
2085         * JSBench/google-chrome-win/urem.js: Added.
2086         * JSBench/google-chrome/urem.html: Added.
2087         * JSBench/google-chrome/urem.js: Added.
2088         * JSBench/google-firefox-win/urem.html: Added.
2089         * JSBench/google-firefox-win/urem.js: Added.
2090         * JSBench/google-firefox/uem.html: Added.
2091         * JSBench/google-firefox/uem.js: Added.
2092         * JSBench/google-safari/urem.html: Added.
2093         * JSBench/google-safari/urem.js: Added.
2094         * JSBench/harness.html: Added.
2095         * JSBench/harness.js: Added.
2096         * JSBench/harness.py: Added.
2097         * JSBench/index.html: Added.
2098         * JSBench/reload.html: Added.
2099         * JSBench/twitter-chrome-win/rem.html: Added.
2100         * JSBench/twitter-chrome-win/rem.js: Added.
2101         * JSBench/twitter-chrome/urem.html: Added.
2102         * JSBench/twitter-chrome/urem.js: Added.
2103         * JSBench/twitter-firefox-win/urem.html: Added.
2104         * JSBench/twitter-firefox-win/urem.js: Added.
2105         * JSBench/twitter-firefox/urem.html: Added.
2106         * JSBench/twitter-firefox/urem.js: Added.
2107         * JSBench/twitter-safari/urem.html: Added.
2108         * JSBench/twitter-safari/urem.js: Added.
2109         * JSBench/yahoo-chrome-win/urem.html: Added.
2110         * JSBench/yahoo-chrome-win/urem.js: Added.
2111         * JSBench/yahoo-chrome/urem.html: Added.
2112         * JSBench/yahoo-chrome/urem.js: Added.
2113         * JSBench/yahoo-firefox-win/urem.html: Added.
2114         * JSBench/yahoo-firefox-win/urem.js: Added.
2115         * JSBench/yahoo-firefox/urem.html: Added.
2116         * JSBench/yahoo-firefox/urem.js: Added.
2117         * JSBench/yahoo-safari/urem.html: Added.
2118         * JSBench/yahoo-safari/urem.js: Added.
2119
2120 2016-05-20  Ryosuke Niwa  <rniwa@webkit.org>
2121
2122         run-benchmark's results should contain Animometer's debug output
2123         https://bugs.webkit.org/show_bug.cgi?id=157941
2124
2125         Reviewed by Stephanie Lewis.
2126
2127         Made developer.html support the JSON generated by run-benchmark which stores everything under debugOutput.
2128
2129         * Animometer/resources/debug-runner/animometer.js:
2130         (Utilities.initialize): Unwrap debugOutput in the case run-benchmark's result JSON is used.
2131
2132 2016-05-18  Timothy Hatcher  <timothy@apple.com>
2133
2134         Make Animometer work in all browsers
2135         https://bugs.webkit.org/show_bug.cgi?id=157855
2136         rdar://problem/26338521
2137
2138         Reviewed by Jon Lee.
2139
2140         * Animometer/resources/runner/logo.svg: Use a symbol instead
2141         of putting the id on the root svg. This works in all browsers.
2142
2143 2016-05-07  Timothy Hatcher  <timothy@apple.com>
2144
2145         Update Animometer design
2146         https://bugs.webkit.org/show_bug.cgi?id=157449
2147
2148         Reviewed by Darin Adler.
2149
2150         * Animometer/developer.html: Stop including runner/animometer.css.
2151         * Animometer/index.html: Udpated elements to fit new styles.
2152
2153         * Animometer/resources/debug-runner/animometer.css:
2154         Copy the old runner/animometer.css to here so the debug runner is unaffected.
2155
2156         * Animometer/resources/runner/animometer.css:
2157         New styles for the design. Drop flex box for broad browser support.
2158
2159         * Animometer/resources/runner/animometer.js:
2160         (ResultsTable.clear): Use textContent.
2161         (window.sectionsManager.showSection): Add class to the body.
2162         (window.sectionsManager.setSectionScore): Use textContent, not innerHTML.
2163
2164         * Animometer/resources/runner/crystal.svg: Added.
2165         * Animometer/resources/runner/lines.svg: Added.
2166         * Animometer/resources/runner/logo.svg: Added.
2167
2168         * Animometer/resources/runner/tests.js: Shorten test name and title case them.
2169
2170 2016-05-06  Manuel Rego Casasnovas  <rego@igalia.com>
2171
2172         [css-grid] Unprefix CSS Grid Layout properties
2173         https://bugs.webkit.org/show_bug.cgi?id=157137
2174
2175         Reviewed by Simon Fraser.
2176
2177         Remove "-webkit" prefix from all the grid layout properties,
2178         including the display value.
2179         Update the source code to remove the prefix where it was used too.
2180
2181         * Layout/auto-grid-lots-of-data.html:
2182         * Layout/fixed-grid-lots-of-data.html:
2183         * Layout/fixed-grid-lots-of-stretched-data.html:
2184
2185 2016-04-25  Geoffrey Garen  <ggaren@apple.com>
2186
2187         bmalloc: Misc improvements to MallocBench
2188         https://bugs.webkit.org/show_bug.cgi?id=157004
2189
2190         Reviewed by Darin Adler.
2191
2192         * MallocBench/run-malloc-benchmarks: Added --memory and --memory_warning
2193         modes for focused memory testing.
2194
2195         * MallocBench/MallocBench/Benchmark.cpp:
2196         (Benchmark::printReport): Clarified output.
2197
2198         (Benchmark::currentMemoryBytes): Added compressed memory because top
2199         does the same. (It always happens to zero in the benchmarks we run. But
2200         this is good for sanity.)
2201
2202         * MallocBench/MallocBench/CommandLine.cpp: Moved up to 8 runs to reduce
2203         variance.
2204
2205         * MallocBench/MallocBench/alloc_free.cpp:
2206         (benchmark_alloc_free): Cycle a single allocation in order to stress
2207         the effect of merging on calls to madvise.
2208
2209         * MallocBench/MallocBench/big.cpp:
2210         (benchmark_big): Graduated to 8kB-128kB because medium tests up to 8 and
2211         our large allocator doesn't kick in until 64kB.
2212
2213         * MallocBench/MallocBench/medium.cpp:
2214         (benchmark_medium): Test all the way down to 1kB because our large
2215         allocator used to service 1kB allocations and 1kB is an interesting
2216         middle size where memory is unusually large but allocation throughput
2217         still matters.
2218
2219         * MallocBench/MallocBench/stress.cpp:
2220         (benchmark_stress): Reduced the churn count to match stress_aligned
2221         because this test was taking too long to complete.
2222
2223         * MallocBench/MallocBench/stress_aligned.cpp:
2224         (benchmark_stress_aligned): Our new large allocator can handle even
2225         more absurdly large values.
2226
2227 2016-04-25  Simon Fraser  <simon.fraser@apple.com>
2228
2229         Add a content animation test that uses SVG animation.
2230         https://bugs.webkit.org/show_bug.cgi?id=156827
2231
2232         * Animation/svg-animation.html: Added.
2233         * Skipped:
2234
2235 2016-04-22  Simon Fraser  <simon.fraser@apple.com>
2236
2237         Skip two content animation tests which are only meant for iOS testing.
2238
2239         * Animation/css-animation.html: Added.
2240         * Animation/raf-animation.html: Added.
2241
2242         * Skipped:
2243
2244 2016-04-20  Simon Fraser  <simon.fraser@apple.com>
2245
2246         Add content animation tests to benchmark_runner, and allow the runner to collect device data as part of the results
2247         https://bugs.webkit.org/show_bug.cgi?id=156827
2248
2249         Add two files missing from the previous commit.
2250
2251         * Animation/css-animation.html: Added.
2252         * Animation/raf-animation.html: Added.
2253
2254 2016-04-20  Simon Fraser  <simon.fraser@apple.com>
2255
2256         Add content animation tests to benchmark_runner, and allow the runner to collect device data as part of the results
2257         https://bugs.webkit.org/show_bug.cgi?id=156827
2258
2259         Reviewed by Stephanie Lewis.
2260
2261         Two new files for the content-animation suite.
2262
2263         After patching for use in the benchmark, these tests are a little different to other benchmarks.
2264         The test content itself does not collect performance data; the tests spit out a boilerplate JSON
2265         result with a placeholder for framerate, which is collected in native code.
2266
2267         When run as a benchmark, test completion involves a couple of bounces between the test and MobileSafari.
2268         Test completion is initiated by a setTimeout() in the test, at which point location.hash is set to "#done".
2269         MobileSafari detects that, and asynchronously dumps performance data. Once this is complete, MobileSafari
2270         changes the URL has to "#submit", which triggers the XHR to the test relay that indicates the test is
2271         complete.
2272
2273         * Animation/css-animation.html: Added.
2274         * Animation/raf-animation.html: Added.
2275
2276 2016-04-15  Jon Lee  <jonlee@apple.com>
2277
2278         Animometer test could report a NaN
2279         https://bugs.webkit.org/show_bug.cgi?id=156646
2280
2281         Reviewed by Darin Adler.
2282         Provisionally reviewed by Said Abou-Hallawa.
2283
2284         * Animometer/tests/resources/main.js:
2285         (didFinishInterval): The ramp controller has a first phase where it ramps up the complexity and
2286         reacts based on how well the system handles the load. The assumption was that it would handle at least
2287         1 particle easily. That is not always the case. As a result, an interpolation calculation could
2288         end up setting an upper bound of NaN.
2289
2290         This occurs because we never get out of the first tier, so this._lastTierComplexity is undefined.
2291         Now that we guarantee a minimum complexity of 1, modify the conditional to check for this._lastTierComplexity
2292         before interpolating the upper bound of the first ramp. In the case where the system struggles with
2293         1 particle, set it to currentComplexity.
2294
2295 2016-04-09  Jon Lee  <jonlee@apple.com>
2296
2297         Update Animometer to accommodate different screens
2298         https://bugs.webkit.org/show_bug.cgi?id=156449
2299
2300         Reviewed by Darin Adler.
2301         Provisionally reviewed by Said Abou-Hallawa.
2302
2303         * Animometer/index.html: Wrap button in a container to add padding at the bottom.
2304         * Animometer/resources/debug-runner/animometer.css:
2305         (@media screen and (min-device-width: 1800px)): Deleted.
2306         (@media screen and (min-width: 1800px)): Cannot use min-device-width since it may match incorrectly.
2307         (screen and (max-device-height: 414px) and (orientation: landscape)): Some devices swap device width
2308         and height with orientation change.
2309         * Animometer/resources/runner/animometer.css: Similar.
2310         (screen and (min-device-width: 1024px) and (orientation: landscape)):
2311         (screen and (max-device-height: 414px) and (orientation: landscape)):
2312         (.frame-container): On smaller iPhones, adding 1px prevents the navigation bars from appearing.
2313         (@media screen and (min-device-width: 768px) and (max-device-width: 1024px)): Deleted.
2314         (@media (min-device-height: 768px) and (max-device-height: 1024px)): Target iPad Airs and similar.
2315         (@media screen and (min-device-width: 1024px) and (max-device-width: 1366px)): Deleted.
2316         (@media screen and (max-device-width: 1024px) and (min-device-height: 1366px)): Target iPad Pro.
2317         (#results footer): Add padding below the button for testing again.
2318         * Animometer/tests/master/multiply.html: Remove the center text.
2319         * Animometer/tests/master/resources/text.js: Update the test so that in every frame the text moves.
2320         * Animometer/tests/master/text.html: Update the text sizing depending on the size of the device.
2321
2322 2016-04-08  Jon Lee  <jonlee@apple.com>
2323
2324         Have Animometer benchmark always start with complexity of 1
2325         https://bugs.webkit.org/show_bug.cgi?id=156432
2326
2327         Reviewed by Ryosuke Niwa.
2328
2329         * Animometer/tests/resources/main.js: Update the default Controller and RampController to
2330         set its minimum complexities to 1 instead of 0.
2331
2332 2016-04-08  Jon Lee  <jonlee@apple.com>
2333
2334         Fix SVG benchmark test
2335         https://bugs.webkit.org/show_bug.cgi?id=156410
2336
2337         Reviewed by Dean Jackson.
2338
2339         * Animometer/resources/extensions.js: Update Point.zero to be a static Point.
2340         * Animometer/tests/simple/resources/tiled-canvas-image.js:
2341         (Stage.call._setupTiles): Refactor.
2342         * Animometer/tests/master/resources/particles.js:
2343         (Particle.prototype.reset): Use Point.center.
2344         (complexity): We are not using a gradient background anymore, so remove the +1.
2345         * Animometer/tests/master/resources/svg-particles.js: Update to use SVG transform
2346         instead of CSS transform.
2347
2348 2016-04-06  Jon Lee  <jonlee@apple.com>
2349
2350         Update master benchmark with SVG test
2351         https://bugs.webkit.org/show_bug.cgi?id=156273
2352
2353         Reviewed by Dean Jackson.
2354         Provisionally reviewed by Said Abou-Hallawa.
2355
2356         Switch masks tests for SVG path test.
2357
2358         * Animometer/resources/debug-runner/tests.js: Move mask test here.
2359         * Animometer/resources/runner/tests.js: Add SVG test here.
2360         * Animometer/tests/dom/particles.html: Renamed from PerformanceTests/Animometer/tests/master/particles.html.
2361         * Animometer/tests/dom/resources/dom-particles.js: Renamed from PerformanceTests/Animometer/tests/master/resources/dom-particles.js.
2362
2363         * Animometer/tests/master/resources/particles.js: Add minPosition for bounds checking. Prevents particle from being
2364         partially obscured.
2365         * Animometer/tests/master/resources/svg-particles.js: Added.
2366         (Particle): The particle is either a path object or a rect using a path as a clip. The
2367         same path is used either way. For each particle create a linear gradient with a random
2368         rotation.
2369         (SVGParticlesStage): Look in #shapes to see how many different kinds of paths are available.
2370         This makes the test more generic in case other shapes need to be tested.
2371         * Animometer/tests/master/svg-particles.html: Added. Have two defs, one that houses each
2372         particle's gradient, and one that holds the shape templates.
2373
2374 2016-03-27  Hunseop Jeong  <hs85.jeong@samsung.com>
2375
2376         [EFL] REGRESSION(r188793): It made 200 layout tests and Bindings/event-target-wrapper.html performance test fail
2377         https://bugs.webkit.org/show_bug.cgi?id=148470 
2378
2379         Reviewed by Darin Adler.
2380
2381         * Skipped: Unskip the Bindings/event-target-wrapper test.
2382
2383 2016-03-23  Jon Lee  <jonlee@apple.com>
2384
2385         Update focus and leaves tests
2386         https://bugs.webkit.org/show_bug.cgi?id=155825
2387
2388         Reviewed by Simon Fraser.
2389
2390         Particles would improperly disappear, before they were fully off stage. Update the calculation.
2391         * Animometer/tests/dom/resources/leaves.js:
2392         * Animometer/tests/master/resources/leaves.js:
2393
2394         * Animometer/tests/master/focus.html: Update the center element styling.
2395         * Animometer/tests/master/resources/focus.js:
2396         (FocusElement.Utilities.createClass): Give enough space in the container for each particle for the blur
2397         to spread. It should be about 3x more space than the blur radius on each side.
2398         (Stage.call.initialize): Ditto for the center element.
2399
2400 2016-03-23  Jon Lee  <jonlee@apple.com>
2401
2402         Update support for other platforms
2403         https://bugs.webkit.org/show_bug.cgi?id=155824
2404
2405         Reviewed by Simon Fraser.
2406
2407         * Animometer/resources/debug-runner/animometer.css: Remove extraneous media query. All of it should
2408         be handled in the release stylesheet.
2409         * Animometer/resources/debug-runner/animometer.js: Add support for browsers that return an HTMLCollection
2410         when retrieving radio elements with a name, instead of the selected value.
2411         * Animometer/resources/runner/animometer.css: Adjust the queries for devices of varying widths.
2412         (.frame-container): Move overflow:hidden to #stage.
2413         * Animometer/tests/master/resources/stage.css: Consolidate into tests/resources/stage.css.
2414         * Animometer/tests/resources/stage.css:
2415         (body): Remove overflow:hidden.
2416         (#stage): Set overflow:hidden here.
2417         (#center-text): Moved from master/resources/stage.css.
2418
2419         Update stylesheet hrefs.
2420         * Animometer/tests/master/canvas-stage.html:
2421         * Animometer/tests/master/focus.html:
2422         * Animometer/tests/master/leaves.html:
2423         * Animometer/tests/master/multiply.html:
2424         * Animometer/tests/master/particles.html:
2425         * Animometer/tests/master/text.html: Adjust text size to fit in the device screen.
2426
2427 2016-03-23  Jon Lee  <jonlee@apple.com>
2428
2429         Make the benchmark require device to be in landscape orientation
2430         https://bugs.webkit.org/show_bug.cgi?id=155822
2431         rdar://problem/25258650
2432
2433         Reviewed by Ryosuke Niwa.
2434
2435         Check that the device is in landscape orientation. If not, disable the button to start it,
2436         and show a warning message.
2437
2438         * Animometer/developer.html: Add a message about how the browser should be set up. Include a
2439         warning message that appears if the orientation is incorrect on mobile devices.
2440         * Animometer/index.html: Ditto.
2441
2442         * Animometer/resources/debug-runner/animometer.css: Migrate the .hidden rule into the release
2443         stylesheet. Update the style to accommodate the new UI.
2444         * Animometer/resources/runner/animometer.css:
2445
2446         * Animometer/resources/runner/animometer.js:
2447         (window.benchmarkController.initialize): Add the orientation listener is needed.
2448         (window.benchmarkController.addOrientationListenerIfNecessary): Only mobile devices need this. Check
2449         to see for support of window.orientation.
2450         (window.benchmarkController._orientationChanged): Toggle the warning based on whether we match the
2451         landscape query. Set a state variable, which is needed for the debug harness. Call updateStartButtonState.
2452         (window.benchmarkController.updateStartButtonState):
2453         * Animometer/resources/debug-runner/animometer.js: Change _updateStartButtonState to return
2454         a boolean about whether at least one test is selected. That will be used in
2455         benchmarkController.updateStartButtonState(). Move the load event listener to the release version.
2456         (window.benchmarkController.updateStartButtonState): Override the release version, and also check
2457         that at least one test is selected.
2458
2459 2016-03-22  Geoffrey Garen  <ggaren@apple.com>
2460
2461         MallocBench: consolidate regression testing for aligned allocation
2462         https://bugs.webkit.org/show_bug.cgi?id=155762
2463
2464         Reviewed by Michael Saboff.
2465
2466         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
2467         * MallocBench/MallocBench/Benchmark.cpp:
2468         * MallocBench/MallocBench/memalign.cpp: Removed.
2469         * MallocBench/MallocBench/memalign.h: Removed. The stress_aligned test
2470         covers this and much more.
2471
2472         * MallocBench/MallocBench/stress_aligned.cpp:
2473         (benchmark_stress_aligned): Include specific tests for extreme sizes
2474         and alignments.
2475
2476 2016-03-21  Jon Lee  <jonlee@apple.com>
2477
2478         Update benchmark tests
2479         https://bugs.webkit.org/show_bug.cgi?id=155723
2480
2481         Reviewed by Darin Adler.
2482         Provisionally reviewed by Said Abou-Hallawa.
2483
2484         Add quadratic and bezier segments to the canvas path test.
2485
2486         * Animometer/resources/runner/tests.js: Some of the query strings are unnecessary.
2487         Rename the test.
2488         * Animometer/tests/master/resources/canvas-stage.js:
2489         (tune): Update to be able to take an array of possible constructors. Choose one randomly.
2490         * Animometer/tests/master/resources/canvas-tests.js:
2491         (CanvasLinePoint.Utilities.createClass): Move the point selection out to a separate
2492         function called randomPoint() for reuse. Move X_LOOPS, Y_LOOPS, and offsets into the
2493         class definition.
2494         (randomPoint): Scale the grid down a little bit so that the lines along the edge of the
2495         canvas are not cut off when the stroke size is thick.
2496         (CanvasQuadraticSegment): Added.
2497         (CanvasBezierSegment): Added.
2498         (SimpleCanvasStage): Pass in an array of the different segment types. Since line segments
2499         are short compared to the curved ones, make it twice as likely to render a line segment.
2500         (SimpleCanvasStage.animate): Update the drawing code so that we render all line segments.
2501
2502         Add a helper method that selects a random element from an array.
2503
2504         * Animometer/tests/resources/main.js:
2505         (Stage.randomElementInArray): Select a random element from the provided array.
2506
2507         * Animometer/tests/bouncing-particles/resources/bouncing-tagged-images.js: Refactor.
2508         * Animometer/tests/dom/resources/leaves.js: Ditto.
2509         * Animometer/tests/master/resources/dom-particles.js: Ditto.
2510         * Animometer/tests/master/resources/image-data.js: Ditto.
2511         * Animometer/tests/master/resources/leaves.js: Ditto.
2512         * Animometer/tests/simple/resources/simple-canvas-paths.js: Ditto.
2513
2514         Add canvas tests that includes all stroke and fill paths. This makes it possible to avoid having to
2515         include the full simple canvas suite for perf testing.
2516
2517         * Animometer/resources/debug-runner/tests.js: Add new tests. Move the canvas test into the Canvas
2518         suite.
2519         * Animometer/tests/simple/resources/simple-canvas-paths.js: Add a CanvasStroke and CanvasFill particle
2520         that random selects an object to render.
2521
2522         * Animometer/resources/debug-runner/tests.js: Move 3D suite before basic canvas suite.
2523         When updating the perf bot script, we will include the suites up to this one, but exclude
2524         the basic canvas suite.
2525
2526         Move compositing transforms test to HTML suite and remove the empty Miscellaneous suite.
2527
2528         * Animometer/resources/debug-runner/tests.js:
2529         * Animometer/tests/dom/compositing-transforms.html: Renamed from PerformanceTests/Animometer/tests/misc/compositing-transforms.html.
2530         * Animometer/tests/dom/resources/compositing-transforms.js: Renamed from PerformanceTests/Animometer/tests/misc/resources/compositing-transforms.js.
2531
2532         Clean up miscellaneous test suite. Add a canvas ellipse test, and remove the other
2533         canvas tests.
2534
2535         * Animometer/resources/debug-runner/tests.js: Add ellipse tests to the simple suite.
2536         * Animometer/tests/misc/canvas-electrons.html: Removed.
2537         * Animometer/tests/misc/canvas-stars.html: Removed.
2538         * Animometer/tests/misc/resources/canvas-electrons.js: Removed.
2539         * Animometer/tests/misc/resources/canvas-stars.js: Removed.
2540         * Animometer/tests/simple/resources/simple-canvas-paths.js: Add ellipse primitives.
2541
2542         Merge text tests together into one.
2543
2544         * Animometer/resources/runner/tests.js: Remove international.html.
2545         * Animometer/tests/master/international.html: Removed.
2546         * Animometer/tests/master/resources/text.js:
2547         (animate): Update styling. Manually calculate gradients for the shadow particles.
2548         Reduce the step size for y direction to avoid cutting text off at the margins. Increase
2549         step for x since there will be a little more room.
2550         * Animometer/tests/master/text.html: Add more translations and lay it out in a table.
2551
2552 2016-03-20  Jon Lee  <jonlee@apple.com>
2553
2554         Add a link to show debug data
2555         https://bugs.webkit.org/show_bug.cgi?id=155724
2556
2557         Reviewed by Simon Fraser.
2558
2559         * Animometer/developer.html: Attach the onclick handler to the score.
2560         * Animometer/index.html: Ditto.
2561
2562 2016-03-18  Dean Jackson  <dino@apple.com>
2563
2564         Add a basic WebGL test to Animometer
2565         https://bugs.webkit.org/show_bug.cgi?id=155475
2566         <rdar://problem/25156860>
2567
2568         Reviewed by Simon Fraser and Jon Lee.
2569
2570         Add a pretty simple WebGL test to Animometer
2571         which tries to draw a number of animated triangles.
2572         The result still needs a bit of tuning.
2573
2574         * Animometer/resources/debug-runner/tests.js: Add a new
2575         Suite for "3D" tests.
2576         * Animometer/tests/3d/resources/webgl.js: Added.
2577         * Animometer/tests/3d/webgl.html: Added.
2578
2579 2016-03-18  Jon Lee  <jonlee@apple.com>
2580
2581         Add support for statically linking to a specific test
2582         https://bugs.webkit.org/show_bug.cgi?id=155631
2583
2584         Rubber-stamped by Darin Adler.
2585
2586         * Animometer/developer.html: Update to call restartBenchmark instead so that it works for
2587         both the preset and the manually set test sessions.
2588         * Animometer/resources/debug-runner/animometer.css:
2589         (.tree .link): Style the "link" UI.
2590         * Animometer/resources/debug-runner/animometer.js: Move updateDisplay from suitesManager
2591         to optionsManager since this is a visual option and has nothing to do with the suites
2592         settings. Update the construction of the test UI by including a "link" after each test.
2593         Clicking on that link brings up a JS prompt with a URL and a query string with the current
2594         parameters of the controller, and selected test. Pasting this URL into the location bar will
2595         automatically start running the selected test.
2596         (suitesManager.suitesFromQueryString): Iterate through the Suites and tests and find the one
2597         that matches the provided parameters. Returns an object similar to the form in
2598         suitesManager.updateLocalStorageFromUI.
2599         (benchmarkController.initialize): After settings up the events and options, try parsing the
2600         query string and running the benchmark immediately. Otherwise, fall back to the form.
2601         (benchmarkController.startBenchmark): Store the options and suites into member variables for
2602         reuse in restartBenchmark.
2603         (benchmarkController.startBenchmarkImmediatelyIfEncoded): Convert the query string to an object.
2604         If that's successful, find the suite and test referenced in the query string. Start the benchmark if
2605         the search for the test succeeded.
2606         * Animometer/resources/extensions.js:
2607         (Utilities.stripNonASCIICharacters): Helper method to convert the name of the suite and test
2608         into a query-string-friendly version.
2609         (Utilities.convertObjectToQueryString): Helper method to convert an object to query string
2610         format.
2611         (Utilities.convertQueryStringToObject): Helper method to convert query string into an object
2612         with properties and values.
2613
2614 2016-03-16  Jon Lee  <jonlee@apple.com>
2615
2616         Add a new benchmark test
2617         https://bugs.webkit.org/show_bug.cgi?id=155570
2618
2619         Reviewed by Simon Fraser.
2620
2621         New Leaves test includes various image sizes and opacity.
2622
2623         * Animometer/resources/debug-runner/tests.js: Add it to the HTML test suite.
2624         * Animometer/tests/dom/leaves.html: Added.
2625         * Animometer/tests/dom/resources/leaves.js: Added. Override the
2626         (Particle.call.reset): Uses a range of sizes, and opacity.
2627         (Particle.call.animate): Opacity goes up then down. When it hits 0, reset the particle.
2628         (Particle.call.move): Set transform and opacity.
2629         * Animometer/tests/master/resources/leaves.js: Get rid of the closure so that it
2630         can be used in this test. Update the relative path so that it works in both the master
2631         and dom test suite.
2632
2633 2016-03-15  Simon Fraser  <simon.fraser@apple.com>
2634
2635         Add developer Animometer test that bounces P3-tagged images
2636         https://bugs.webkit.org/show_bug.cgi?id=155511
2637
2638         Reviewed by Tim Horton.
2639         
2640         Add a test for rendering performance of tagged images. The 5 images are tagged
2641         with the Display P3 colorspace.
2642
2643         * Animometer/resources/debug-runner/tests.js:
2644         * Animometer/tests/bouncing-particles/bouncing-tagged-images.html: Added.
2645         * Animometer/tests/bouncing-particles/resources/bouncing-tagged-images.js: Added.
2646         * Animometer/tests/bouncing-particles/resources/image1.jpg: Added.
2647         * Animometer/tests/bouncing-particles/resources/image2.jpg: Added.
2648         * Animometer/tests/bouncing-particles/resources/image3.jpg: Added.
2649         * Animometer/tests/bouncing-particles/resources/image4.jpg: Added.
2650         * Animometer/tests/bouncing-particles/resources/image5.jpg: Added.
2651
2652 2016-03-09  Jon Lee  <jonlee@apple.com>
2653
2654         Enhance existing Animometer tests
2655         https://bugs.webkit.org/show_bug.cgi?id=155261
2656
2657         Reviewed by Simon Fraser.
2658
2659         * Animometer/tests/master/resources/canvas-tests.js: Add a gradient to
2660         the filled circles. To expose more of the gradient, add another circle.
2661         (SimpleCanvasStage.call.animate): For each frame, create a gradient with
2662         undulating stop points and colors. Fill the circles twice; once with
2663         the solid color, and once with the gradient.
2664
2665         * Animometer/tests/master/resources/dom-particles.js: Refactor the
2666         emission variables into a separate stage for this test. Add a colorOffset
2667         variable to make the colors of each particle slightly different, since
2668         the ramp controller can add large numbers of particles all at once, which
2669         would otherwise get all the same color.
2670         * Animometer/tests/master/resources/particles.js:
2671         (initialize): Remove the code specific to the SVG mask test.
2672
2673         * Animometer/tests/master/resources/image-data.js:
2674         (initialize): Remove unused local variable.
2675         * Animometer/tests/master/resources/multiply.js:
2676         (initialize): Make the test harder by adding more total particles.
2677
2678 2016-03-09  Jon Lee  <jonlee@apple.com>
2679
2680         Add text tests
2681         https://bugs.webkit.org/show_bug.cgi?id=155257
2682
2683         Reviewed by Simon Fraser.
2684
2685         * Animometer/resources/extensions.js:
2686         (UnitBezier.Utilities.createClass): Add a class that computes Bezier points
2687         assuming that two of the control points are at (0,0) and (1,1). Taken from
2688         WebCore/platform/graphics/UnitBezier.h
2689         * Animometer/resources/runner/tests.js: Add a test for English text and one
2690         for international text.
2691         * Animometer/tests/master/international.html: Added.
2692         * Animometer/tests/master/resources/text.js: Added. The test assumes there is
2693         #template div which it will copy. The copies are placed behind the template, and
2694         are set with different colors each frame. They are moved around with CSS transform.
2695         * Animometer/tests/master/text.html: Added.
2696
2697         Remove the other text tests, since these ones cover the same techniques.
2698         * Animometer/tests/text/layering-text.html: Removed.
2699         * Animometer/tests/text/resources/layering-text.js: Removed.
2700         * Animometer/tests/text/resources/text-boxes.js: Removed.
2701         * Animometer/tests/text/text-boxes.html: Removed.
2702         * Animometer/resources/debug-runner/tests.js:
2703
2704 2016-03-09  Jon Lee  <jonlee@apple.com>
2705
2706         Add a new image test
2707         https://bugs.webkit.org/show_bug.cgi?id=155232
2708
2709         Reviewed by Dean Jackson.
2710         Provisionally reviewed by Said Abou-Hallawa.
2711
2712         The image test renders PNGs and moves them with translate and rotate
2713         transforms. Each particle has a lifetime, and when the lifetime ends or
2714         the particle goes offscreen, it resets itself somewhere on the stage.
2715
2716         * Animometer/resources/debug-runner/tests.js: Remove the CSS bouncing PNG
2717         images test, because this one tests the same technique.
2718         * Animometer/resources/runner/tests.js: Added here as "Leaves".
2719         * Animometer/tests/master/leaves.html: Added.
2720         * Animometer/tests/master/resources/particles.js: Refactor out the parts
2721         specific to the DOM particles test. Consequently make velocity a public
2722         member.
2723         (initialize): The options parameter is never used, so remove it.
2724         (animate): Remove unused local variable.
2725         * Animometer/tests/master/resources/leaves.js: Maintains a focal point that
2726         moves back and forth across the canvas. That point affects the velocity
2727         of the particles. When the focal point is on the leftmost side, it is
2728         between [-6, -2], and set in reset(). When the focal point is on the rightmost
2729         side, it will be from [2, 6].
2730         * Animometer/tests/master/resources/dom-particles.js: Move JS specific to
2731         this test here from particles.js.
2732
2733         New images.
2734         * Animometer/tests/master/resources/compass100.png: Added.
2735         * Animometer/tests/master/resources/console100.png: Added.
2736         * Animometer/tests/master/resources/contribute100.png: Added.
2737         * Animometer/tests/master/resources/debugger100.png: Added.
2738         * Animometer/tests/master/resources/inspector100.png: Added.
2739         * Animometer/tests/master/resources/layout100.png: Added.
2740         * Animometer/tests/master/resources/performance100.png: Added.
2741         * Animometer/tests/master/resources/script100.png: Added.
2742         * Animometer/tests/master/resources/shortcuts100.png: Added.
2743         * Animometer/tests/master/resources/standards100.png: Added.
2744         * Animometer/tests/master/resources/storage100.png: Added.
2745         * Animometer/tests/master/resources/styles100.png: Added.
2746         * Animometer/tests/master/resources/timeline100.png: Added.
2747
2748 2016-03-03  Jon Lee  <jonlee@apple.com>
2749
2750         Add ability to retrieve raw data from release harness
2751         https://bugs.webkit.org/show_bug.cgi?id=155026
2752
2753         Reviewed by Simon Fraser.
2754
2755         * Animometer/developer.html: Remove the special UI in the debug
2756         harness.
2757         * Animometer/resources/runner/animometer.css: Add styles for the
2758         overlay.
2759         * Animometer/resources/runner/animometer.js: Let 'j' show the JSON
2760         results, but only if the overlay doesn't exist. Add 'esc' key to dismiss
2761         the overlay.
2762         (window.benchmarkController.selectResults): Cycle the cases around
2763         so that the first 's' press selects both the benchmark score and
2764         the individual test scores.
2765         * Animometer/resources/debug-runner/animometer.css: Remove
2766         unneeded rules.
2767         * Animometer/resources/debug-runner/animometer.js: Make the same
2768         call to handleKeyPress.
2769
2770 2016-03-03  Jon Lee  <jonlee@apple.com>
2771
2772         Make sure multiply test particles have at least some opacity
2773         https://bugs.webkit.org/show_bug.cgi?id=155027
2774
2775         Reviewed by Simon Fraser.
2776
2777         * Animometer/tests/master/resources/multiply.js:
2778         (initialize): Have the elements spin a little faster to make it
2779         more obvious when the system is being stressed.
2780         (_addTile):
2781         (animate): Make sure each element being animated has at least 1%
2782         opacity.
2783
2784 2016-03-03  Jon Lee  <jonlee@apple.com>
2785
2786         Update image test
2787         https://bugs.webkit.org/show_bug.cgi?id=154962
2788
2789         Rubber-stamped by Darin Adler.
2790
2791         * Animometer/tests/master/image-data.html: Make each canvas have a compositing layer.
2792         * Animometer/tests/master/resources/image-data.js:
2793         (initialize): Update to load multiple images serially.
2794         (_loadImage):
2795         (tune): Use display instead of visibility. When showing an element anew, move it somewhere else.
2796         (_createTestElement): Set up the element, refactor out placement of element to _refreshElement.
2797         (_refreshElement): Place element in a tile grid.
2798         (animate):
2799         (_getRandomNeighboringPixelIndex): Use a more conservative distribution to make the effect last a little longer.
2800         * Animometer/tests/master/resources/compass.svg: Added.
2801         * Animometer/tests/master/resources/console.svg: Added.
2802         * Animometer/tests/master/resources/contribute.svg: Added.
2803         * Animometer/tests/master/resources/debugger.svg: Added.
2804         * Animometer/tests/master/resources/inspector.svg: Added.
2805         * Animometer/tests/master/resources/layout.svg: Added.
2806         * Animometer/tests/master/resources/performance.svg: Added.
2807         * Animometer/tests/master/resources/script.svg: Added.
2808         * Animometer/tests/master/resources/shortcuts.svg: Added.
2809         * Animometer/tests/master/resources/standards.svg: Added.
2810         * Animometer/tests/master/resources/storage.svg: Added.
2811         * Animometer/tests/master/resources/styles.svg: Added.
2812         * Animometer/tests/master/resources/timeline.svg: Added.
2813
2814 2016-03-02  Jon Lee  <jonlee@apple.com>
2815
2816         Add some new controllers, and refine tests
2817         https://bugs.webkit.org/show_bug.cgi?id=154914
2818
2819         Reviewed by Simon Fraser.
2820
2821         Improve tests.
2822
2823         * Animometer/tests/master/focus.html: Move each particle into a layer. Put the particle
2824         in a container element. Expand the container element by the max blur radius, and clip
2825         overflow. This way, when the blur is applied, it avoids causing layer resizes.
2826         * Animometer/tests/master/resources/focus.js:
2827         (FocusElement.Utilities.createClass): Create a container element, and put the particle
2828         inside.
2829         (hide): Set display:none.
2830         (show): Set display:block.
2831         (animate): Apply filters to the container element.
2832         (FocusStage.call.initialize): Instead of inserting and removing elements from the DOM,
2833         keep them in the stage, but set the display style instead. Use this._offsetIndex to
2834         keep track of which elements are displayed.
2835         (FocusStage.call.animate): Move some calculations around to avoid doing unneeded math.
2836         (FocusStage.call.getBlurValue): Make sure elements are always blurred.
2837         (FocusStage.call.getOpacityValue): Make sure elements have some opacity.
2838         * Animometer/tests/master/resources/multiply.js: Have the particles on the edge of the
2839         stage appear less black when the complexity gets large enough.
2840         (tune): Have this._distanceFactor calculate the factor. Avoid calculating square root each
2841         frame.
2842
2843 2016-03-02  Jon Lee  <jonlee@apple.com>
2844
2845         Add some new controllers, and refine tests
2846         https://bugs.webkit.org/show_bug.cgi?id=154914
2847
2848         Reviewed by Simon Fraser.
2849
2850         Add a controller that centers around 30 fps instead of 60 fps.
2851
2852         * Animometer/developer.html: Add a new option.
2853         * Animometer/resources/debug-runner/animometer.js:
2854         * Animometer/resources/runner/animometer.js:
2855         (this._processData.findRegression): When calculating the complexity-frameLength regression,
2856         check the controller, and use a 30 fps baseline if needed.
2857         * Animometer/resources/statistics.js:
2858         (Regression.Utilities.createClass): Update to allow clients to specify the baseline
2859         frame length.
2860         * Animometer/tests/resources/main.js:
2861         (tune): Override some of the constants in RampController. Move those constants out for
2862         easier reading.
2863
2864         Add a fixed controller, with no step.
2865
2866         * Animometer/developer.html: Add a controller that takes no step.
2867         * Animometer/resources/debug-runner/animometer.js:
2868         * Animometer/tests/resources/main.js:
2869         (Rotater.Utilities.createClass):
2870
2871         Switch to ramp controller as default.
2872
2873         * Animometer/developer.html: Increase the test length to 20 seconds.
2874         * Animometer/resources/debug-runner/animometer.js: Rename the "adjustment" field to "controller"
2875         since that is a more accurate description.
2876         * Animometer/resources/debug-runner/graph.js:
2877         * Animometer/resources/runner/animometer.js: Update preferences for release suite.
2878
2879 2016-02-26  Jon Lee  <jonlee@apple.com>
2880
2881         Address Dean's comments in 154673.
2882
2883         * Animometer/developer.html:
2884         * Animometer/resources/debug-runner/animometer.js:
2885         * Animometer/resources/debug-runner/graph.js:
2886
2887 2016-02-25  Jon Lee  <jonlee@apple.com>
2888
2889         Update animation benchmark and tests
2890         https://bugs.webkit.org/show_bug.cgi?id=154673
2891
2892         Reviewed by Dean Jackson.
2893
2894         Update test visuals.
2895
2896         * Animometer/tests/master/resources/canvas-tests.js:
2897         (CanvasLineSegment.Utilities.createClass): Line segments near the edge of the stage get clipped.
2898         Update circle position and radius to minimize impact.
2899         * Animometer/tests/master/resources/particles.js: Get rid of rotating gradient background, and
2900         have 3 locations for emitting particles.
2901         * Animometer/tests/master/resources/multiply.js: Update distance metric so that fringe tiles get
2902         more color.
2903         (initialize): Fix some of the math for laying out the tiles.
2904         (animate): When a tile isn't used, set visibility: hidden.
2905
2906 2016-02-24  Jon Lee  <jonlee@apple.com>
2907
2908         Update animation benchmark and tests
2909         https://bugs.webkit.org/show_bug.cgi?id=154673
2910
2911         Reviewed by Dean Jackson.
2912
2913         Update the ramp controller.
2914
2915         The controller refines the complexity interval to test across.
2916
2917         * Animometer/resources/statistics.js: Add functions that estimate cumulative distribution function.
2918         (Regression): For the flat regression, force the first segment to be at 60 fps.
2919         (valueAt): Add convenience function to return interpolated value based on the regression used.
2920         (_calculateRegression): Include the number of points included for both segments, and the piecewise
2921         errors.
2922         * Animometer/tests/resources/math.js: Make the Kalman estimator subclass Experiment, and allow it
2923         to be reset.
2924
2925         * Animometer/tests/resources/main.js: Initialize the tier such that it starts at 10^0 = 1.
2926         Increase the number of ramps. Maintain three FPS thresholds-- the frame rate of interest, a limit
2927         on the lowest FPS we care to go for later interpolation, and a minimum FPS threshold we want to
2928         aim for each ramp. Also keep three estimators: a running average of the change point, a minimum
2929         boundary for each ramp, and an estimator for all the frames within an interval. The first two
2930         are used to determine the parameters of the next ramp, and the latter allows us to refine the
2931         parameters.
2932         (update): During the tier phase, it is possible that the highest complexity possible for a test
2933         won't stress the system enough to trigger stopping the tier phase and transitioning to the ramps.
2934         If the complexity doesn't change when going to the next tier, we've maxed the test out, and move
2935         on. When the tier phase completed, turn off Controller.frameLengthEstimator, which estimates the
2936         FPS at each tier.
2937         (tune): At each interval, look at the confidence distribution of being on the 60 FPS side or the
2938         slow side. If the slowest FPS we achieve at the ramp's maximum complexity is not at least
2939         _fpsRampSlowThreshold, then increase the maximum complexity. If we ever achieve 60 FPS, increase
2940         the ramp's minimum complexity to that level. If, at an even lower complexity, a glitch causes the
2941         FPS to drop, we reset the minimum complexity.
2942
2943         Have the bootstrap calculation occur between tests. Clean up harness.
2944
2945         * Animometer/resources/debug-runner/animometer.js: Run bootstrap after a test has
2946         completed to avoid doing all of it at the end before showing the results. Clean up
2947         parameters being passed around.
2948         * Animometer/resources/debug-runner/tests.js:
2949         (text):
2950         * Animometer/resources/runner/animometer.js:
2951         (this._processData.calculateScore): Save the results to the same object holding the data.
2952         (this._processData._processData): In the case where a file is dragged, calculate the score
2953         serially. Grab the results object and move it to the results variable and remove it from
2954         the data object. This avoids serializing the results into the JSON.
2955         (this._processData.findRegression): Include the samples used for bootstrapping. Reduce the
2956         resample size to shorten the wait.
2957         * Animometer/resources/runner/benchmark-runner.js:
2958         * Animometer/resources/statistics.js:
2959         (bootstrap): Update how bootstrapData is sorted. In some regression results the mix of
2960         floats and integers causes an alphabetical sort to occur.
2961         * Animometer/resources/strings.js:
2962
2963         Add meta charset so that encodings between harness and test match.
2964
2965         * Animometer/tests/bouncing-particles/bouncing-canvas-images.html:
2966         * Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html:
2967         * Animometer/tests/bouncing-particles/bouncing-css-images.html:
2968         * Animometer/tests/bouncing-particles/bouncing-css-shapes.html:
2969         * Animometer/tests/bouncing-particles/bouncing-svg-images.html:
2970         * Animometer/tests/bouncing-particles/bouncing-svg-shapes.html:
2971         * Animometer/tests/master/canvas-stage.html:
2972         * Animometer/tests/master/focus.html:
2973         * Animometer/tests/master/image-data.html:
2974         * Animometer/tests/master/multiply.html:
2975         * Animometer/tests/master/particles.html:
2976         * Animometer/tests/misc/canvas-electrons.html:
2977         * Animometer/tests/misc/canvas-stars.html:
2978         * Animometer/tests/misc/compositing-transforms.html:
2979         * Animometer/tests/simple/simple-canvas-paths.html:
2980         * Animometer/tests/simple/tiled-canvas-image.html:
2981         * Animometer/tests/template/template-canvas.html:
2982         * Animometer/tests/template/template-css.html:
2983         * Animometer/tests/template/template-svg.html:
2984         * Animometer/tests/text/layering-text.html:
2985         * Animometer/tests/text/text-boxes.html:
2986
2987         Update test harness reporting.
2988
2989         * Animometer/developer.html: Add missing meta charset.
2990         * Animometer/index.html: Remove unnecessary utf-8 declaration.
2991         * Animometer/resources/debug-runner/animometer.css: Add convenience classes for
2992         formatting the results table.
2993         * Animometer/resources/debug-runner/animometer.js: Adjust which stats are shown.
2994         * Animometer/resources/debug-runner/tests.js: Display bootstrapping statistics.
2995         * Animometer/resources/strings.js: Move strings not used by the release harness.
2996
2997         Switch to a pseudo-random number generator.
2998
2999         * Animometer/resources/statistics.js: Add a Pseudo class, with a simple
3000         pseudo-random number generator.
3001         (_calculateRegression): Reset the generator before running bootstrap.
3002         (bootstrap): Deleted.
3003
3004         Replace Math.random with Pseudo.random.
3005         * Animometer/tests/master/resources/canvas-tests.js:
3006         * Animometer/tests/master/resources/focus.js:
3007         * Animometer/tests/master/resources/particles.js:
3008         * Animometer/tests/resources/main.js:
3009
3010         Use bootstrapping to get confidence interval in the breakpoint.
3011
3012         For the ramp controller, calculate the piecewise regression, and then use
3013         bootstrapping in order to find the 95% confidence interval. Use the raw data.
3014
3015         * Animometer/developer.html: Default to the complexity graph. Add a legend
3016         checkbox to toggle visibility of the bootstrap score and histogram.
3017         * Animometer/resources/debug-runner/animometer.css: Make some more space to show
3018         the old raw and average scores in the legend. Add new styles for the data.
3019         * Animometer/resources/debug-runner/graph.js:
3020         (_addRegressionLine): Allow passing an array for the variance bar tied to the
3021         regression line. Now |stdev| is |range|.
3022         (createComplexityGraph): Add bootstrap median, and overlay a histogram of
3023         the bootstrap samples. Switch raw samples from circles to X's.
3024         (onComplexityGraphOptionsChanged): Allow toggling of the bootstrap data.
3025         (onGraphTypeChanged): Move the regressions for the raw and average samples to the
3026         legend. In the subtitle use the bootstrap median, and include the 95% confidence
3027         interval.
3028         * Animometer/resources/runner/animometer.js:
3029         (this._processData.findRegression): Factor out the code that determines which
3030         samples to include when calculating the piecewise regression. For series that have
3031         many samples, or a wider range of recorded complexities, throw away the 2.5%
3032         lowest and highest samples before calculating the regression. Keep all samples
3033         if the number of samples to regress is small or the range of complexities is
3034         narrow.
3035         (this._processData._calculateScore): Factor out regression calculation to
3036         findRegression(). Bootstrap the change point of the regression. The score is the
3037         median.
3038         * Animometer/resources/statistics.js:
3039         (_calculateRegression): Correct an issue in the calculation of the regression, where
3040         the denominator can be 0.
3041         (bootstrap): Template for bootstrapping. Create a bootstrap sample array, Create
3042         re-samples by random selection with replacement. Return the 95% confidence samples,
3043         the bootstrap median, mean, and the data itself.
3044         * Animometer/resources/strings.js: Add bootstrap.
3045         * Animometer/tests/resources/main.js:
3046         (processSamples): Don't prematurely cut the sample data.
3047
3048         Fix graph drawing.
3049
3050         * Animometer/resources/debug-runner/animometer.js: Add spacing in the JSON output.
3051         Multiple tests output a lot of JSON and can hang when selecting JSON with no whitespace.
3052         * Animometer/resources/debug-runner/animometer.css:
3053         (#complexity-graph .series.raw circle): Update the color.
3054         * Animometer/resources/debug-runner/graph.js: Use the FPS axis instead of the
3055         complexity axis, which can vary in domain. For determining the complexity domain,
3056         only use samples after samplingTimeOffset.
3057
3058         Allow dropping results JSON.
3059
3060         * Animometer/developer.html: Add a button.
3061         * Animometer/resources/debug-runner/animometer.css:
3062         * Animometer/resources/debug-runner/animometer.js: Read the data and go straight
3063         to the dashboard. With JSON output, write out only the options and the raw data.
3064
3065         Teach the harness to evaluate the samples and determine the test score.
3066
3067         This will allow us to update how the score is calculated separately from the samples recorded.
3068         This also prepares the harness to be able to accept JSON of prior runs.
3069
3070         * Animometer/resources/strings.js: Clean up and remove unneeded strings and reduce some of the
3071         hierarchy.
3072         * Animometer/resources/debug-runner/tests.js: Update to use the new strings.
3073
3074         * Animometer/tests/resources/main.js: Allow all controllers to show a complexity-FPS graph.
3075         (_processComplexitySamples): Factor out some of the sample processing done in the ramp
3076         controller for the benefit of the other controllers. |complexitySamples| contains a list of
3077         samples. Sort the samples by complexity. Optionally remove the top x% of samples.
3078         Group them, and calculate distribution of samples within the same complexity, and add those as
3079         new entries into |complexityAverageSamples|.
3080         (Controller.processSamples): Move the code responsible for determining the complexity and FPS
3081         scores out to ResultsDashboard. The structure of the data returned by the controller is:
3082
3083         {
3084             controller: [time-regression, time-regression, ...], // optional, data specific to controller
3085             marks: [...],
3086             samples: {                    // all of the sample data
3087                 controller: [...],
3088                 complexity: [...],        // processed from controller samples
3089                 complexityAverage: [...], // processed from complexity samples
3090             }
3091         }
3092
3093         (AdaptiveController.processSamples): Adding the target frame length is no longer necessary; we
3094         now pass the test options to the graph.
3095         (Regression): Move to statistics.js.
3096         * Animometer/resources/statistics.js: Move Regression to here. Add a check if the sampling range
3097         only contains one sample, since we cannot calculate a regression from one sample point.
3098
3099         Teach the test harness to evaluate the data.
3100         * Animometer/resources/runner/animometer.js:
3101         (ResultsDashboard): Store the options used to run the test and the computed results/score separately
3102         from the data. The results are stored as:
3103
3104         {
3105             score: /* geomean of iteration score */,
3106             iterationsResults: [
3107                 {
3108                     score: /* geomean of tests */,
3109                     testsResults: {
3110                         suiteName: {
3111                             testName: {
3112                                 controller: {
3113                                     average:
3114                                     concern:
3115                                     stdev:
3116                                     percent:
3117                                 },
3118                                 frameLength: { ... },
3119                                 complexity: {
3120                                     complexity:
3121                                     stdev:
3122                                     segment1:
3123                                     segment2:
3124                                 },
3125                                 complexityAverage: { ... }
3126                             },
3127                             testName: { ... },
3128                         },
3129                         ... next suite ...
3130                     }
3131                 },
3132                 { ...next iteration... }
3133             ]
3134         }
3135
3136         * Animometer/resources/debug-runner/animometer.js: Pass options around instead of relying
3137         on what was selected in the form. This will later allow for dropping previous results, and
3138         using those runs' options when calculating scores.
3139         (ResultsTable._addGraphButton): Simplify button action by using attached test data.
3140         * Animometer/resources/debug-runner/graph.js: Refactor to use the data.
3141
3142         Consolidate JS files, and move statistics out to a separate JS.
3143
3144         Preparation for having the Controller only handle recording and storage of the samples,
3145         and leave the evaluation of the test score out to the harness. Move Experiment to
3146         a new statistics.js, where Regression will also eventually go. Get rid of algorithm.js
3147         and move it to utilities.js since the Heap is used only for Experiments.
3148
3149         * Animometer/tests/resources/algorithm.js: Removed. Heap is in utilities.js.
3150         * Animometer/tests/resources/sampler.js: Removed. Experiment is in statistics.js,
3151         Sampler in main.js.
3152         * Animometer/tests/resources/main.js: Move Sampler here.
3153         * Animometer/resources/statistics.js: Added. Move Statistics and Experiment here.
3154         * Animometer/resources/extensions.js: Move Heap here. Attach static method to create
3155         a max or min heap to Heap, instead of a new Algorithm object.
3156
3157         Update JS files.
3158         * Animometer/developer.html:
3159         * Animometer/index.html:
3160         * Animometer/tests/bouncing-particles/bouncing-canvas-images.html:
3161         * Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html:
3162         * Animometer/tests/bouncing-particles/bouncing-css-images.html:
3163         * Animometer/tests/bouncing-particles/bouncing-css-shapes.html:
3164         * Animometer/tests/bouncing-particles/bouncing-svg-images.html:
3165         * Animometer/tests/bouncing-particles/bouncing-svg-shapes.html:
3166         * Animometer/tests/master/canvas-stage.html:
3167         * Animometer/tests/master/focus.html:
3168         * Animometer/tests/master/image-data.html:
3169         * Animometer/tests/master/multiply.html:
3170         * Animometer/tests/master/particles.html:
3171         * Animometer/tests/misc/canvas-electrons.html:
3172         * Animometer/tests/misc/canvas-stars.html:
3173         * Animometer/tests/misc/compositing-transforms.html:
3174         * Animometer/tests/simple/simple-canvas-paths.html:
3175         * Animometer/tests/simple/tiled-canvas-image.html:
3176         * Animometer/tests/template/template-canvas.html:
3177         * Animometer/tests/template/template-css.html:
3178         * Animometer/tests/template/template-svg.html:
3179         * Animometer/tests/text/layering-text.html:
3180         * Animometer/tests/text/text-boxes.html:
3181
3182         Fix the cursor in the graph analysis when the min
3183         complexity is not 0.
3184
3185         * Animometer/resources/debug-runner/graph.js:
3186         (_addRegression):
3187         (createComplexityGraph):
3188
3189 2016-02-23  Geoffrey Garen  <ggaren@apple.com>
3190
3191         Fix some issues in MallocBench
3192         https://bugs.webkit.org/show_bug.cgi?id=154600
3193
3194         Reviewed by Oliver Hunt.
3195
3196         * MallocBench/MallocBench.xcodeproj/project.pbxproj: Make nimlang.ops
3197         a part of the build so that it copies to the right place for execution.
3198
3199         * MallocBench/MallocBench/Interpreter.cpp:
3200         (Interpreter::Interpreter): Don't try to open .ops files for writing
3201         because we're only going to read and we might need extra permissions to
3202         write.
3203
3204         (Interpreter::~Interpreter): Give more context when opening a file fails
3205         to help with debugging.
3206
3207         * MallocBench/MallocBench/stress.cpp:
3208         (benchmark_stress): Reduce iterations to complete in less than 10 seconds.
3209
3210 2016-02-23  Michael Saboff  <msaboff@apple.com>
3211
3212         Unreviewed change to revert extraneous changes made part of change set 196955.
3213
3214         * MallocBench/MallocBench/Interpreter.cpp:
3215         (Interpreter::doMallocOp):
3216         (Interpreter::Thread::switchTo):
3217         (writeData): Deleted.
3218
3219 2016-02-19  Michael Saboff  <msaboff@apple.com>
3220
3221         MallocBench: Added recording for nimlang website, new recording details and added new options
3222         https://bugs.webkit.org/show_bug.cgi?id=154485
3223
3224         Reviewed by Geoff Garen.
3225
3226         Added new capabilities to MallocBench.  These include:
3227             Added a recording of http://nim-lang.org/docs/lib.html.
3228             Added thread id to the recording and the ability to playback switching threads in MallocBench
3229             Added aligned allocations to recordings and the ability to playback
3230             Added --use-thread-id option to honor recorded thread ids
3231             Added --detailed-report to output remaining allocations by size after playback
3232             Added --no-warmup to not run the warm up iteration
3233
3234         Changed the way that options are passed down to the benchmarks.  Instead of passing individual
3235         boolean or numeric option values, just pass a reference the CommandLine itself.  Each benchmark
3236         can access the options that are appropriate.  The Benchmark class also uses the options for
3237         is parallel, run counts and warm up.
3238
3239         Added thread id and aligned malloc to the Op by noticing that structure padding and Opcode allowed
3240         for another 32 bits of data.  Breaking that unused 32 bits into a 16 bit thread id value and a
3241         16 bit log base 2 of the alignment for aligned malloc allowed for existing recordings to playback
3242         without any incompatibilities.
3243
3244         Threaded operation is simulated by creating threads as needed.  As long as the next Op's thread id
3245         is the same as the last, operation continues as normal.  When the next Op has a different thread id,
3246         we switch to that thread using the shared Op stream to continue playing back.  There is a mutex to
3247         assure that only one thread is really running at a time and a condition variable used to wait
3248         that the current thread id matches each block thread's thread id.  This doesn't simulate true
3249         concurrent threading, but is instead plays back Ops recorded for multiple thread faithfully.
3250
3251         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
3252         * MallocBench/MallocBench/Benchmark.cpp:
3253         (deallocateHeap):
3254         (Benchmark::Benchmark):
3255         (Benchmark::runOnce):
3256         (Benchmark::run):
3257         * MallocBench/MallocBench/Benchmark.h:
3258         (Benchmark::isValid):
3259         * MallocBench/MallocBench/CommandLine.cpp:
3260         (CommandLine::printUsage):
3261         * MallocBench/MallocBench/CommandLine.h:
3262         (CommandLine::isValid):
3263         (CommandLine::benchmarkName):
3264         (CommandLine::isParallel):
3265         (CommandLine::useThreadID):
3266         (CommandLine::detailedReport):
3267         (CommandLine::warmUp):
3268         (CommandLine::heapSize):
3269         (CommandLine::runs):
3270         * MallocBench/MallocBench/Interpreter.cpp:
3271         (Interpreter::Interpreter):
3272         (Interpreter::run):
3273         (Interpreter::readOps):
3274         (Interpreter::doOnSameThread):
3275         (Interpreter::switchToThread):
3276         (Interpreter::detailedReport):
3277         (compute2toPower):
3278         (writeData):
3279         (Interpreter::doMallocOp):
3280         (Interpreter::Thread::Thread):
3281         (Interpreter::Thread::stop):
3282         (Interpreter::Thread::~Thread):
3283         (Interpreter::Thread::runThread):
3284         (Interpreter::Thread::waitToRun):
3285         (Interpreter::Thread::switchTo):
3286         * MallocBench/MallocBench/Interpreter.h:
3287         (Interpreter::Thread::isMainThread):
3288         * MallocBench/MallocBench/alloc_free.cpp: Added.
3289         (benchmark_alloc_free):
3290         * MallocBench/MallocBench/alloc_free.h: Added.
3291         * MallocBench/MallocBench/balloon.cpp:
3292         (benchmark_balloon):
3293         * MallocBench/MallocBench/balloon.h:
3294         * MallocBench/MallocBench/big.cpp:
3295         (benchmark_big):
3296         * MallocBench/MallocBench/big.h:
3297         * MallocBench/MallocBench/churn.cpp:
3298         (benchmark_churn):
3299         * MallocBench/MallocBench/churn.h:
3300         * MallocBench/MallocBench/facebook.cpp:
3301         (benchmark_facebook):
3302         * MallocBench/MallocBench/facebook.h:
3303         * MallocBench/MallocBench/flickr.cpp:
3304         (benchmark_flickr):
3305         (benchmark_flickr_memory_warning):
3306         * MallocBench/MallocBench/flickr.h:
3307         * MallocBench/MallocBench/fragment.cpp:
3308         (validate):
3309         (benchmark_fragment):
3310         (benchmark_fragment_iterate):
3311         * MallocBench/MallocBench/fragment.h:
3312         * MallocBench/MallocBench/list.cpp:
3313         (benchmark_list_allocate):
3314         (benchmark_list_traverse):
3315         * MallocBench/MallocBench/list.h:
3316         * MallocBench/MallocBench/main.cpp:
3317         (main):
3318         * MallocBench/MallocBench/medium.cpp:
3319         (benchmark_medium):
3320         * MallocBench/MallocBench/medium.h:
3321         * MallocBench/MallocBench/memalign.cpp:
3322         (test):
3323         (benchmark_memalign):
3324         * MallocBench/MallocBench/memalign.h:
3325         * MallocBench/MallocBench/message.cpp:
3326         (benchmark_message_one):
3327         (benchmark_message_many):
3328         * MallocBench/MallocBench/message.h:
3329         * MallocBench/MallocBench/nimlang.cpp: Added.
3330         (benchmark_nimlang):
3331         * MallocBench/MallocBench/nimlang.h: Added.
3332         * MallocBench/MallocBench/nimlang.ops: Added.
3333         * MallocBench/MallocBench/realloc.cpp:
3334         (benchmark_realloc):
3335         * MallocBench/MallocBench/realloc.h:
3336         * MallocBench/MallocBench/reddit.cpp:
3337         (benchmark_reddit):
3338         (benchmark_reddit_memory_warning):
3339         * MallocBench/MallocBench/reddit.h:
3340         * MallocBench/MallocBench/stress.cpp:
3341         (deallocate):
3342         (benchmark_stress):
3343         * MallocBench/MallocBench/stress.h:
3344         * MallocBench/MallocBench/stress_aligned.cpp:
3345         (benchmark_stress_aligned):
3346         * MallocBench/MallocBench/stress_aligned.h:
3347         * MallocBench/MallocBench/theverge.cpp:
3348         (benchmark_theverge):
3349         (benchmark_theverge_memory_warning):
3350         * MallocBench/MallocBench/theverge.h:
3351         * MallocBench/MallocBench/tree.cpp:
3352         (benchmark_tree_allocate):
3353         (benchmark_tree_traverse):
3354         (benchmark_tree_churn):
3355         * MallocBench/MallocBench/tree.h:
3356         * MallocBench/run-malloc-benchmarks:
3357
3358 2016-02-11  Jon Lee  <jonlee@apple.com>
3359
3360         Fix a missing refactoring.
3361
3362         * Animometer/tests/master/resources/multiply.js:
3363         (animate): Move to Utilities.lerp.
3364         (_lerp): Deleted.
3365
3366 2016-02-10  Jon Lee  <jonlee@apple.com>
3367
3368         Add new benchmark tests.
3369         https://bugs.webkit.org/show_bug.cgi?id=154063
3370
3371         Provisionally reviewed by Said Abou-Hallawa.
3372
3373         Add tests for get/put image data, filters, opacity, and css transforms.
3374
3375         * Animometer/resources/runner/benchmark-runner.js:
3376         (_runBenchmarkAndRecordResults): Update the body background color to match that of
3377         the stage.
3378         (this._runNextIteration): Clear the background color style for the results page.
3379         * Animometer/resources/runner/tests.js:
3380         * Animometer/tests/master/focus.html: Added.
3381         * Animometer/tests/master/image-data.html: Added.
3382         * Animometer/tests/master/multiply.html: Added.
3383         * Animometer/tests/master/resources/focus.js: Added.
3384         * Animometer/tests/master/resources/image-data.js: Added.
3385         * Animometer/tests/master/resources/multiply.js: Added.
3386         * Animometer/tests/master/resources/stage.css: Move common styles out.
3387         * Animometer/tests/resources/main.js: Update Stage.randomBool to use Math.random.
3388         Add Stage.randomSign for randomly setting a direction. Add the notion of the
3389         current timestamp of the test to Benchmark, since some animations cycle through
3390         colors and rely on an incremental counter like the time.
3391
3392 2016-02-09  Said Abou-Hallawa  <sabouhallawa@apple.com>
3393
3394         Add internal benchmark tests for CSS mix-blend-modes and filters
3395         https://bugs.webkit.org/show_bug.cgi?id=154058
3396
3397         Provisionally reviewed by Jon Lee.
3398
3399         * Animometer/resources/debug-runner/tests.js: Include the new tests in the
3400         "HTML suite" of the debug runner.
3401
3402         * Animometer/resources/extensions.js:
3403         (Utilities.browserPrefix):
3404         (Utilities.setElementPrefixedProperty): Utility functions to allow setting
3405         prefixed style properties.
3406
3407         * Animometer/tests/bouncing-particles/resources/bouncing-css-shapes.js:
3408         Set the mix-blend-mode and the filter to some random values if the options
3409         of the test requested that.
3410
3411         * Animometer/tests/bouncing-particles/resources/bouncing-particles.js:
3412         (parseShapeParameters): Parse the url options "blend" and "filter" and set
3413         the corresponding flags.
3414
3415         * Animometer/tests/resources/main.js:
3416         (randomStyleMixBlendMode):
3417         (randomStyleFilter): Return random mix-blend-mode and filter.
3418
3419 2016-02-08  Jon Lee  <jonlee@apple.com>
3420
3421         Add a ramp controller
3422         https://bugs.webkit.org/show_bug.cgi?id=154028
3423
3424         Provisionally reviewed by Said Abou-Hallawa.
3425
3426         Enhance the graph to include a complexity-fps graph, in addition
3427         to the time graph.
3428
3429         * Animometer/developer.html: Add a ramp option.
3430         * Animometer/resources/debug-runner/animometer.css: Update the style.
3431         * Animometer/resources/strings.js: Flatten the Strings.text constants.
3432         * Animometer/resources/debug-runner/animometer.js:
3433         (ResultsTable.call._addGraphButton): Refactor.
3434         (ResultsTable.call._addTest): Add regression data.
3435         (benchmarkController): Add a form that allows the user to switch between the two forms,
3436         Add a form that allows the user to toggle different data. Hide certain header columns
3437         depending on the selected controller.
3438         * Animometer/resources/debug-runner/graph.js: Add the complexity regressions.
3439         * Animometer/resources/debug-runner/tests.js: Add headers for the ramp results.
3440         * Animometer/resources/runner/animometer.js:
3441         (ResultsTable): If a header is disabled don't include them in _flattenedHeaders.
3442         * Animometer/tests/resources/main.js:
3443         (Controller): Allow options to specify the capacity for sample arrays.
3444         (Regression): A piecewise regression that tries to fit a slope and a flat profile.
3445         (_calculateRegression): Options can fix the slope and bias when calculating the minimal
3446         error. Sweep across the samples in time (which could be backward depending on the controller)
3447         and calculate the intersection point.
3448         (RampController): This controller assumes that the target frame rate is below
3449         58 FPS. It runs in two stages. The first stage quickly determines the order of
3450         magnitude of objects needed to stress the system by the setting the complexity
3451         to increasingly difficult tiers. Perform a series of ramps descending from a
3452         high-water mark of complexity. The complexity needed to reach the target frame
3453         length is done by performing a piecewise regression on each ramp, and track a
3454         running average of these values. For the next ramp, make that running average
3455         the center of the ramp. With a minimum complexity of 0, the high-water mark is
3456         twice that average. The score is based on the highest complexity that can
3457         reach 60 fps.
3458
3459 2016-02-08  Jon Lee  <jonlee@apple.com>
3460
3461         Address Said's comments on the benchmark, and do some clean up.
3462
3463         * Animometer/developer.html:
3464         * Animometer/resources/debug-runner/animometer.css: Add styles for averages.
3465         * Animometer/resources/debug-runner/animometer.js: Use the right
3466         Strings constants.
3467         * Animometer/resources/debug-runner/graph.js:
3468         (_addRegressionLine): Add missing code to draw the line and standard
3469         deviation highlight.
3470         (onGraphTypeChanged): Remove unneeded variables
3471         (onTimeGraphOptionsChanged):
3472         * Animometer/resources/runner/benchmark-runner.js:
3473         (_runBenchmarkAndRecordResults): Rename samplers to suiteResults and
3474         _suitesSamplers to _suitesResults.
3475         * Animometer/tests/resources/main.js:
3476         (results): Call processSamples().
3477         (update): Change sampling timestamp comparison.
3478         (_animateLoop): Move shouldStop call to before the update.
3479         * Animometer/tests/resources/sampler.js:
3480         (process): Rename to processSamples().
3481
3482 2016-02-07  Jon Lee  <jonlee@apple.com>
3483
3484         Teach Controller to measure intervals, and turn off the frame length estimator.
3485
3486         * Animometer/tests/resources/main.js: Default interval length is 100 ms.
3487         (start): Set the first interval.
3488         (_measureAndResetInterval): Reports the average frame length of the interval that just
3489         completed, and sets up the next interval.
3490         (update): If there is no length, then just use the estimator per frame, otherwise the
3491         estimator measures per interval. Add a didFinishInterval for subclasses to process
3492         prior to recording the sample. Update tune() to include whether an interval had
3493         finished.
3494         (StepController): Step controllers don't measure on an interval basis.
3495
3496 2016-02-07  Jon Lee  <jonlee@apple.com>
3497
3498         Minor refactoring. Rename Controller._estimator to Controller._frameLengthEstimator
3499         and switch the parameters for start(), update(), and tune(), so that the timestamp
3500         is first and stage is second.
3501
3502         * Animometer/tests/resources/main.js:
3503
3504 2016-02-07  Jon Lee  <jonlee@apple.com>
3505
3506         Move ResultsTable functionality not needed for release tests out.
3507         Move reporting of score and mean to selection of the time-based graph.
3508
3509         * Animometer/developer.html: Rename graph-options to time-graph-options.
3510         * Animometer/resources/debug-runner/animometer.js:
3511         (DeveloperResultsTable): Moved from runner/animometer.js. Switch from mean
3512         values to "average" objects which can hold stdev. Move graph button and
3513         calculation of noisy measurements here. Sophisticated header processing
3514         is not needed in release suite.
3515         (populateTable): Use DeveloperResultsTable.
3516         * Animometer/resources/debug-runner/graph.js: Pull time graph creation to
3517         its own function, and add a new onGraphTypeChanged handler in preparation
3518         of a complexity graph to be added later.
3519         * Animometer/resources/runner/animometer.js:
3520         (ResultsTable): Simplify to just handle test names and scores.
3521
3522 2016-02-07  Jon Lee  <jonlee@apple.com>
3523
3524         Tests: reuse objects already made.
3525
3526         Avoid thrash of object creation and removal by maintaining an index that
3527         moves along the array as the adjust values change. If the tune value
3528         requires more objects than the maximum size of the object array, then create
3529         new objects. This means that the object array size never decreases.
3530
3531         * Animometer/tests/master/resources/canvas-stage.js: Maintain a separate
3532         offsetIndex. For these tests, we want to avoid drawing the oldest objects,
3533         so the scene will draw the object at offsetIndex to the end of the array.
3534         (tune): Reverse the logic since "removal" of objects is much simpler and
3535         involves simply changing the offsetIndex.
3536         (animate): Update the for loop to draw from offsetIndex to the end.
3537         (complexity): Update the definition.
3538         * Animometer/tests/master/resources/canvas-tests.js: Maintain a separate
3539         offsetIndex. For these tests, we want to avoid drawing the newest objects,
3540         so the scene will draw the object at index 0 to the object at offsetIndex.
3541         (SimpleCanvasStage.animate): Fly-by removal of local stage variable,
3542         which is unneeded. Update the for loop to draw from offsetIndex to the end.
3543         * Animometer/tests/simple/resources/simple-canvas-paths.js:
3544         (SimpleCanvasStage.animate): Update the for loop to draw from 0 to
3545         offsetIndex.
3546         * Animometer/tests/simple/resources/simple-canvas.js:
3547         (tune): Update logic. Here, offsetIndex represents the boundary of the last
3548         index to render.
3549         (animate): Update the for loop to draw from 0 to offsetIndex.
3550         (complexity): Update the definition.
3551
3552 2016-02-07  Jon Lee  <jonlee@apple.com>
3553
3554         Tests: refactor and update styles.
3555
3556         * Animometer/tests/resources/main.js: Add helper methods that return
3557         a color that hue rotates based on the date, and a counter value that
3558         increases based on the date. Fix randomInt() to not bias against the min and
3559         max values.
3560
3561         * Animometer/tests/master/resources/canvas-tests.js: Use new helper methods.
3562         * Animometer/tests/master/resources/dom-particles.js: Ditto.
3563         * Animometer/tests/master/resources/particles.js: Ditto.
3564         * Animometer/tests/simple/resources/simple-canvas-paths.js: Refactor to
3565         use a rotating color instead of a random color. The fast switching of color
3566         is too vivid to watch.
3567
3568         * Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js:
3569         (BouncingSvgParticlesStage.call.createGradient): Fix the gradient so
3570         that the last stop is located at the end.
3571
3572 2016-02-07  Jon Lee  <jonlee@apple.com>
3573
3574         Refactor tune() to not return the complexity of the scene.
3575
3576         We have stage.complexity() now, so returning the complexity through tune
3577         is unnecessary.
3578
3579         * Animometer/tests/bouncing-particles/resources/bouncing-particles.js:
3580         * Animometer/tests/master/resources/canvas-stage.js:
3581         * Animometer/tests/master/resources/particles.js:
3582         * Animometer/tests/misc/resources/canvas-electrons.js:
3583         * Animometer/tests/misc/resources/canvas-stars.js:
3584         * Animometer/tests/resources/main.js:
3585         * Animometer/tests/simple/resources/simple-canvas.js:
3586         * Animometer/tests/simple/resources/tiled-canvas-image.js:
3587         * Animometer/tests/template/resources/template-canvas.js:
3588         * Animometer/tests/template/resources/template-css.js:
3589         * Animometer/tests/template/resources/template-svg.js:
3590         * Animometer/tests/text/resources/layering-text.js:
3591
3592 2016-02-07  Jon Lee  <jonlee@apple.com>
3593
3594         Make the fixed controller a step controller instead. Halfway through the test
3595         it will bump up the complexity 4-fold. Calculate the step timestamp using options
3596         instead of a separate parameter to the Controller constructor.
3597
3598         * Animometer/developer.html: Change value to "step"
3599         * Animometer/resources/debug-runner/animometer.js:
3600         (window.suitesManager.updateEditsElementsState): Show number inputs when set to "step".
3601         * Animometer/tests/resources/main.js:
3602         (update): Provide a hook for subclasses to tune.
3603         (StepController): Maintain a flag determining whether we've stepped, and the time
3604         we should step.
3605         (Benchmark): Use the new StepController.
3606
3607 2016-02-07  Jon Lee  <jonlee@apple.com>
3608
3609         Adjust the FPS graph scale.
3610
3611         Instead of making the FPS graph linearly scale, scale it based on the frame length,
3612         but show the data in terms of FPS. Because it is inversely proportional, and most
3613         of the data never gets below 20, concentrate the axis from 20-60 FPS, since otherwise
3614         over half of the available graph space ends up blank.
3615
3616         This means we should convert all of the FPS data to frame length data.
3617
3618         * Animometer/resources/debug-runner/graph.js: Update the domain to be based on
3619         frame length in milliseconds instead of FPS. Update the cursor to consider all of the
3620         values being shown, and then pick the min and max values to represent the length of the
3621         cursor.
3622         * Animometer/resources/runner/animometer.js:
3623         * Animometer/resources/strings.js:
3624         * Animometer/tests/resources/main.js:
3625         (processSamples): Add the ability to only sample a range of the data instead of everything
3626         after an offset index. Update sampler to record the frame lengths instead of the frame
3627         rate.
3628
3629 2016-02-07  Jon Lee  <jonlee@apple.com>
3630
3631         Add option to use different methods for retrieving a timestamp.
3632
3633         * Animometer/developer.html: Add performance.now and Date.now options.
3634         * Animometer/resources/runner/animometer.js: Default to performance.now.
3635         (window.benchmarkController.startBenchmark):
3636         * Animometer/tests/resources/main.js: Tie the desired method to _getTimestamp.
3637         (run): Use _getTimestamp.
3638         (_animateLoop): Ditto.
3639
3640 2016-02-07  Jon Lee  <jonlee@apple.com>
3641
3642         Allow adding any number of markers to the graph. The markers can be labeled
3643         and contain timestamp and sample index data. Make it a part of the controller
3644         rather than keeping it in the sampler.
3645
3646         * Animometer/resources/debug-runner/animometer.css: Add styles for markers
3647         * Animometer/resources/debug-runner/graph.js: Create the markers and add
3648         text labels.
3649         * Animometer/resources/runner/animometer.js: Assume the samplingTimeOffset
3650         is just one of the marks provided.
3651         * Animometer/resources/strings.js: Add Strings.json.marks.
3652         * Animometer/tests/resources/main.js:
3653         (Controller): Keep marks here. They are keyed by the marker name, so no two
3654         markers should have the same name.
3655         (recordFirstSample): Refactor to use mark.
3656         (mark): Allows for arbitrary data if needed later. The timestamp maintained
3657         is relative to the absolute start timestamp.
3658         (containsMark): Checks whether a mark with a specific comment exists.
3659         (processSamples): Removes the _startTimestamp offset from the marks before
3660         setting it in results.
3661         * Animometer/tests/resources/sampler.js: Remove marks.
3662
3663 2016-02-07  Jon Lee  <jonlee@apple.com>
3664
3665         Get rid of options member variable in Benchmark.
3666
3667         Options are only needed when initializing the stage or benchmark, so there's no
3668         need to also keep a reference to it.
3669
3670         * Animometer/tests/resources/main.js: Get rid of options variable in Benchmark.
3671         Pass options to Controllers and Stages.
3672         (Controller.Utilities.createClass):
3673         (Benchmark.Utilities.createClass):
3674         (get options): Deleted.
3675
3676         * Animometer/tests/bouncing-particles/resources/bouncing-canvas-images.js:
3677         * Animometer/tests/bouncing-particles/resources/bouncing-canvas-particles.js:
3678         * Animometer/tests/bouncing-particles/resources/bouncing-canvas-shapes.js:
3679         * Animometer/tests/bouncing-particles/resources/bouncing-css-images.js:
3680         * Animometer/tests/bouncing-particles/resources/bouncing-css-shapes.js:
3681         * Animometer/tests/bouncing-particles/resources/bouncing-particles.js:
3682         * Animometer/tests/bouncing-particles/resources/bouncing-svg-images.js:
3683         * Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js:
3684         * Animometer/tests/master/resources/canvas-stage.js:
3685         * Animometer/tests/master/resources/canvas-tests.js:
3686         * Animometer/tests/master/resources/particles.js:
3687         * Animometer/tests/misc/resources/canvas-electrons.js:
3688         * Animometer/tests/misc/resources/canvas-stars.js:
3689         * Animometer/tests/misc/resources/compositing-transforms.js:
3690         * Animometer/tests/simple/resources/simple-canvas-paths.js:
3691         * Animometer/tests/simple/resources/tiled-canvas-image.js:
3692         * Animometer/tests/template/resources/template-canvas.js:
3693         * Animometer/tests/template/resources/template-css.js:
3694         * Animometer/tests/template/resources/template-svg.js:
3695         * Animometer/tests/text/resources/layering-text.js:
3696
3697 2016-02-07  Jon Lee  <jonlee@apple.com>
3698
3699         Update how the benchmark is run
3700         https://bugs.webkit.org/show_bug.cgi?id=153960
3701
3702         Provisionally reviewed by Said Abou-Hallawa.
3703
3704         Introduce the notion of a Controller. It is responsible for recording, updating,
3705         and processing the statistics and complexity of the benchmark. This allows
3706         plugging in different Controllers.
3707
3708         This strips most of the functionality from Animator and BenchmarkState, so fold
3709         what's left into Benchmark. Now, Benchmarks only own a stage and a controller, but
3710         are responsible for driving the animation loop.
3711
3712         Rewrite Animator._shouldRequestAnotherFrame into two different Controllers. One
3713         maintains a fixed complexity, and the other adapts the complexity to meet a
3714         fixed FPS.
3715
3716         Fix the Kalman estimator to be modeled on a scalar variable with no model.
3717
3718         * Animometer/tests/resources/main.js: Remove BenchmarkState and Animator, and
3719         replace it with a Controller. Add a FixedController and refactor the previous controller
3720         to an AdaptiveController.
3721
3722         (Controller): Controllers own the estimator and the sampler. When a new frame is
3723         displayed, the animation loop calls update(). The estimator and sampler record
3724         stats, then tune. Samplers can track multiple series of data. The basic controller
3725         tracks timestamp, complexity, and estimated frame rate.
3726                 The Kalman estimation is based on the frame length rather than the frame
3727         rate. Because FPS is inverse proportional to frame length, in the case where the measured
3728         frame length is very small, the FPS ends up being a wildly large number (in the order of
3729         600-1000 "FPS"), and it pulls the estimator up drastically enough that it takes a while
3730         for it to settle back down. Using frame length reduces the impact of these spikes.
3731                 Converging the estimation takes enough time to avoid initializing it immediately
3732         when the benchmark starts. Instead, the benchmark runs for a brief period of time (100ms)
3733         before running it in earnest. Allow controllers an opportunity to set the complexity
3734         before starting recording.
3735                 When the benchmark is complete, the controller has an opportunity to process
3736         the samples. The default implementation calculates the raw FPS based on the time
3737         difference of the samples, and calculates the complexity score. This is moved from
3738         Benchmark.processSamples.
3739
3740         (Controller): Initialize timestamps. These are at first relative to the start of the
3741         benchmark, but are offset by the absolute start time during start(). By default maintain
3742         3 data series, but subclasses can override.
3743         (start): Calls recordFirstSample() for subclasses to override if needed.
3744         (recordFirstSample): For basic controller, start sampling at the beginning.
3745         (update): Update the frame length estimator and sample.
3746         (shouldStop): Checks that the time is before _endTimestamp.
3747         (results): Returns the processed samples.
3748         (processSamples): Iterate through the sample data and collate them. Include scores.
3749
3750         (FixedComplexityController): Controller that tunes the stage to the desired complexity
3751         prior to starting, and keeps it at that complexity.
3752
3753         (AdaptiveController): Have the estimator estimate the interval frame rate instead of the
3754         raw frame rate.
3755                 The previous version of this controller ignored the frame that came after the
3756         adjustment. The raw FPS show that whatever noise the scene change adds is negligible
3757         compared to the noise of the system overall. Stop ignoring that frame and include all
3758         frames in the measurements.
3759
3760         (Benchmark): Remove dependency on animator, and instantiate a runner based on what is
3761         selected. Most of the loop's functionality is in Controller, so remove here.
3762         (Benchmark.run): Remove start() since it is only called from run(), and fold it in here.
3763         (Benchmark._animateLoop): Fold in from Animator.animateLoop. Let the benchmark run for
3764         a brief period before calling Controller.start().
3765
3766         * Animometer/tests/resources/math.js: Fix the Kalman estimator. The filter estimates
3767         a scalar variable, and makes basic assumptions regarding the model. As a result
3768         none of the linear algebra classes are needed, so remove Matrix, Vector3, and Matrix3.
3769         (SimpleKalmanEstimator): Calculate the gain based on the provided process and
3770         measurement errors.
3771         (KalmanEstimator): Deleted.
3772         (IdentityEstimator): Deleted.
3773         (PIDController): Refactor to use the Utilities.createClass() helper.
3774
3775         The Kalman filter algorithm is explained here http://greg.czerniak.info/guides/kalman1/.
3776         The state, represented by a scalar, is the estimated frame length. There is no user
3777         transition of the state, and the state is the same as the measurement. With this model,
3778         the estimation error converges, so calculate the gain ahead of time.
3779
3780         * Animometer/developer.html: Remove fixed-after-warmup since it is not useful.
3781         Replace the option to toggle the estimator, and make it possible to customize the
3782         estimator's error parameters. Show raw FPS by default, and remove interval FPS,
3783         which will be shown instead of the filtered raw FPS.
3784         * Animometer/resources/debug-runner/animometer.css: Put the header behind the graph.
3785         Remove #intervalFPS rules; move the color to #filteredFPS.
3786         * Animometer/resources/debug-runner/graph.js:
3787         (updateGraphData): Update the hr style to force the layout to be calculated
3788         correctly. Change the tick format to be in terms of seconds, since the timestamps
3789         are in milliseconds. Remove interval data.
3790         * Animometer/resources/runner/animometer.js:
3791         (window.benchmarkController.startBenchmark): Set Kalman parameters.
3792         * Animometer/resources/runner/benchmark-runner.js:
3793         (_runBenchmarkAndRecordResults): When a benchmark completes, expect it to return
3794         the final data, rather than passing a sampler from the controller. This avoids
3795         needing to expose the sampler variable in the benchmark.
3796         * Animometer/tests/resources/sampler.js:
3797         (process): Move the setting of the target frame rate to AdaptiveController.
3798
3799 2016-02-06  Jon Lee  <jonlee@apple.com>
3800
3801         Code clean up: Move Rotater function closer to Stage static methods.
3802         The Rotater is used together with those methods; keep them close.
3803
3804         * Animometer/tests/resources/main.js:
3805
3806 2016-02-06  Jon Lee  <jonlee@apple.com>
3807
3808         Update the JS includes due to ResultsTable move.
3809
3810         * Animometer/developer.html:
3811         * Animometer/index.html:
3812
3813 2016-02-06  Jon Lee  <jonlee@apple.com>
3814
3815         Move createElement and createSVGElement to Utilities.
3816
3817         * Animometer/resources/extensions.js:
3818         (Utilities.createElement): Added.
3819         (Utilities.createSVGElement): Added.
3820         (DocumentExtension.createElement): Deleted.
3821         (DocumentExtension.createSvgElement): Deleted.
3822
3823         * Animometer/resources/debug-runner/animometer.js:
3824         * Animometer/resources/runner/animometer.js:
3825         * Animometer/tests/bouncing-particles/resources/bouncing-svg-images.js:
3826         * Animometer/tests/bouncing-particles/resources/bouncing-svg-particles.js:
3827         * Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js:
3828
3829 2016-02-06  Jon Lee  <jonlee@apple.com>
3830
3831         Add a convenience function for creating a class.
3832
3833         The pattern for creating a class is common enough to add as a Utilities
3834         helper function. It also makes it easy to collapse class definitions when
3835         editing.
3836