Refined the ARES-6 experience
[WebKit-https.git] / PerformanceTests / ChangeLog
1 2017-03-06  Jon Davis  <jond@apple.com>
2
3         Refined the ARES-6 experience
4         https://bugs.webkit.org/show_bug.cgi?id=169205
5
6         Reviewed by Saam Barati.
7
8         * ARES-6/about.html: Copied from PerformanceTests/ARES-6/index.html.
9         * ARES-6/driver.js:
10         (Driver.prototype.start):
11         (Driver.prototype._updateIterations):
12         (Driver):
13         * ARES-6/index.html:
14         * ARES-6/stats.js:
15         (Stats.prototype.toString.span.span):
16         * ARES-6/styles.css:
17         (body):
18         (.swoop):
19         (p):
20         (.logo):
21         (.start):
22         (.test):
23         (.test:nth-child(odd)):
24         (.overall):
25         (.about h2):
26         (.button):
27         (.button:hover):
28         (.button.return:before):
29         (@keyframes fade-in):
30         (@keyframes scale-in):
31         (@keyframes flicker-in):
32         (@keyframes wipe-in):
33         (.flip): Deleted.
34         (.testing, .about): Deleted.
35         (.testing): Deleted.
36         (.about): Deleted.
37         (#about:target): Deleted.
38
39 2017-03-06  Jeremy Jones  <jeremyj@apple.com>
40
41         StandardCanPlayThrough.html does not work on iOS.
42         https://bugs.webkit.org/show_bug.cgi?id=169207
43
44         Reviewed by Jer Noble.
45
46         On iOS only, add muted=true and autoplay=allowed. The loading policy on iOS requires these attributes to load enough
47         data to trigger canPlayThrough.
48
49         * Media/StandardCanPlayThrough.html:
50
51 2017-02-13  Zalan Bujtas <zalan@apple.com>
52
53         Simple line layout: Add performance test with text-rendering: optimizeSpeed
54         https://bugs.webkit.org/show_bug.cgi?id=168248
55
56         Reviewed by Simon Fraser.
57
58         * Layout/simple-line-layout-with-varying-content-and-optimized-speed.html: Added.
59
60 2017-02-10  Zalan Bujtas  <zalan@apple.com>
61
62         Mail hangs when removing multiple rows from large table.
63         https://bugs.webkit.org/show_bug.cgi?id=168103
64         <rdar://problem/30090186>
65
66         Reviewed by Ryosuke Niwa.
67
68         * DOM/large-table-edit.html: Added.
69
70 2017-02-05  Filip Pizlo  <fpizlo@apple.com>
71
72         Beef up the ARES-6 explainer text
73         https://bugs.webkit.org/show_bug.cgi?id=167867
74
75         Reviewed by Saam Barati.
76         
77         Added appropriate links.
78
79         Added text that explains the subscores.
80
81         * ARES-6/index.html:
82
83 2017-02-06  Jer Noble  <jer.noble@apple.com>
84
85         Playback stalls when a SourceBuffer append causes frame eviction
86         https://bugs.webkit.org/show_bug.cgi?id=167834
87
88         Reviewed by Eric Carlson.
89
90         Add an in-page performance test measuring the amount of time required
91         to append a large amount of media data to a SourceBuffer, and then to
92         completely remove that data 30s at a time.
93
94         Add a microbenchmark for MediaTime which measures the amount of time
95         required to create a 1M entry std::map and traverse the map 1M times.
96
97         * Media/MSERemoveCodedFrames.html: Added.
98         * Media/media-source-loader.js:
99         (MediaSourceLoader.prototype.get duration):
100         * MediaTime/Configurations/Base.xcconfig: Added.
101         * MediaTime/Configurations/DebugRelease.xcconfig: Added.
102         * MediaTime/Makefile: Added.
103         * MediaTime/MediaTime.xcodeproj/project.pbxproj: Added.
104         * MediaTime/main.cpp: Added.
105         (performTest):
106         (test):
107         (main):
108         * Skipped:
109
110 2017-02-06  Saam Barati  <sbarati@apple.com>
111
112         Make ARES-6 work from the CLI again
113         https://bugs.webkit.org/show_bug.cgi?id=167895
114
115         Reviewed by Michael Saboff.
116
117         * ARES-6/driver.js:
118         (Driver.prototype.readyTrigger):
119         (Driver.prototype.disableTrigger):
120         * ARES-6/stats.js:
121         (Stats.prototype.toString.span.span):
122         (Stats.prototype.toString.return.span):
123         (Stats.prototype.toString):
124
125 2017-02-06  Ryan Haddad  <ryanhaddad@apple.com>
126
127         Skip ARES-6 on performance bots
128         https://bugs.webkit.org/show_bug.cgi?id=167863
129
130         Unreviewed test gardening.
131
132         ES6SampleBench was renamed to ARES-6, so this should be updated in PerformanceTests/Skipped as well.
133
134         * Skipped:
135
136 2017-02-05  Filip Pizlo  <fpizlo@apple.com>
137
138         Change ES6SampleBench into ARES-6
139         https://bugs.webkit.org/show_bug.cgi?id=167863
140
141         Rubber stamped by Saam Barati.
142         
143         This imports changes that Jonathan Davis (jond@apple.com) made to ES6SampleBench to make it look
144         super awesome. Its now called ARES-6.
145
146         * ARES-6: Copied from PerformanceTests/ES6SampleBench.
147         * ARES-6/ARES-6.svg: Added.
148         * ARES-6/driver.js:
149         (Driver):
150         (Driver.prototype.readyTrigger):
151         (Driver.prototype.disableTrigger):
152         (Driver.prototype.start):
153         (Driver.prototype._updateIterations):
154         * ARES-6/glue.js:
155         * ARES-6/index.html:
156         * ARES-6/results.js:
157         (Results.prototype.reportRunning):
158         (Results.prototype.reportDone):
159         * ARES-6/stats.js:
160         (Stats.prototype.toString.return.span):
161         (Stats.prototype.toString.span.span):
162         (Stats.prototype.toString): Deleted.
163         * ARES-6/style.css: Removed.
164         * ARES-6/styles.css: Added.
165         (html):
166         (body):
167         (header,):
168         (p a):
169         (p a:hover):
170         (.flip):
171         (header):
172         (h2,):
173         (p):
174         (.about h2):
175         (.logo):
176         (.start):
177         (.start.ready):
178         (.start.ready:hover):
179         (.start:after):
180         (.start.ready .testrun):
181         (.tests):
182         (.test):
183         (.overall h2,):
184         (.test .indicator):
185         (.test .indicator.running):
186         (@keyframes test-running):
187         (100%):
188         (.score):
189         (.score label):
190         (.score .value):
191         (.score .units):
192         (.score .margin):
193         (.overall):
194         (.overall .score):
195         (.overall .margin):
196         (.testing, .about):
197         (.testing):
198         (.about):
199         (#about:target):
200         (@media only screen and (max-width: 784px)):
201         * ARES-6/swoop.svg: Added.
202         * ES6SampleBench: Removed.
203         * ES6SampleBench/Air: Removed.
204         * ES6SampleBench/Air/README.md: Removed.
205         * ES6SampleBench/Air/airjs-tests.yaml: Removed.
206         * ES6SampleBench/Air/all.js: Removed.
207         * ES6SampleBench/Air/allocate_stack.js: Removed.
208         * ES6SampleBench/Air/arg.js: Removed.
209         * ES6SampleBench/Air/basic_block.js: Removed.
210         * ES6SampleBench/Air/benchmark.js: Removed.
211         * ES6SampleBench/Air/code.js: Removed.
212         * ES6SampleBench/Air/custom.js: Removed.
213         * ES6SampleBench/Air/frequented_block.js: Removed.
214         * ES6SampleBench/Air/insertion_set.js: Removed.
215         * ES6SampleBench/Air/inst.js: Removed.
216         * ES6SampleBench/Air/liveness.js: Removed.
217         * ES6SampleBench/Air/make_dist.sh: Removed.
218         * ES6SampleBench/Air/opcode.js: Removed.
219         * ES6SampleBench/Air/payload-airjs-ACLj8C.js: Removed.
220         * ES6SampleBench/Air/payload-gbemu-executeIteration.js: Removed.
221         * ES6SampleBench/Air/payload-imaging-gaussian-blur-gaussianBlur.js: Removed.
222         * ES6SampleBench/Air/payload-typescript-scanIdentifier.js: Removed.
223         * ES6SampleBench/Air/reg.js: Removed.
224         * ES6SampleBench/Air/stack_slot.js: Removed.
225         * ES6SampleBench/Air/stress-test.js: Removed.
226         * ES6SampleBench/Air/strip-hash.rb: Removed.
227         * ES6SampleBench/Air/symbols.js: Removed.
228         * ES6SampleBench/Air/test.html: Removed.
229         * ES6SampleBench/Air/test.js: Removed.
230         * ES6SampleBench/Air/tmp.js: Removed.
231         * ES6SampleBench/Air/tmp_base.js: Removed.
232         * ES6SampleBench/Air/util.js: Removed.
233         * ES6SampleBench/Basic: Removed.
234         * ES6SampleBench/Basic/ast.js: Removed.
235         * ES6SampleBench/Basic/basic-tests.yaml: Removed.
236         * ES6SampleBench/Basic/basic.js: Removed.
237         * ES6SampleBench/Basic/benchmark.js: Removed.
238         * ES6SampleBench/Basic/caseless_map.js: Removed.
239         * ES6SampleBench/Basic/lexer.js: Removed.
240         * ES6SampleBench/Basic/number.js: Removed.
241         * ES6SampleBench/Basic/parser.js: Removed.
242         * ES6SampleBench/Basic/random.js: Removed.
243         * ES6SampleBench/Basic/state.js: Removed.
244         * ES6SampleBench/Basic/stress-test.js: Removed.
245         * ES6SampleBench/Basic/test.html: Removed.
246         * ES6SampleBench/Basic/test.js: Removed.
247         * ES6SampleBench/Basic/util.js: Removed.
248         * ES6SampleBench/air_benchmark.js: Removed.
249         * ES6SampleBench/basic_benchmark.js: Removed.
250         * ES6SampleBench/cli.js: Removed.
251         * ES6SampleBench/driver.js: Removed.
252         * ES6SampleBench/glue.js: Removed.
253         * ES6SampleBench/index.html: Removed.
254         * ES6SampleBench/results.js: Removed.
255         * ES6SampleBench/stats.js: Removed.
256         * ES6SampleBench/style.css: Removed.
257
258 2017-02-05  Zalan Bujtas  <zalan@apple.com>
259
260         Simple line layout: Add performance test with varying string and font-size to minimize width-cache hit.
261         https://bugs.webkit.org/show_bug.cgi?id=167859
262         <rdar://problem/30368070>
263
264         Reviewed by Antti Koivisto.
265
266         This is in preparation for webkit.org/b/167843.
267
268         * Layout/simple-line-layout-with-varying-content.html: Added.
269
270 2017-01-16  Filip Pizlo  <fpizlo@apple.com>
271
272         Add framerate baselines to the splay-detail test
273         https://bugs.webkit.org/show_bug.cgi?id=167659
274
275         Reviewed by Geoffrey Garen.
276
277         * JetStream/Octane2/splay-detail.html:
278
279 2017-01-31  Carlos Alberto Lopez Perez  <clopez@igalia.com>
280
281         [EFL][GTK] Skip some IndexedDB tests that time out on the Perf bots.
282
283         Unreviewed performance test gardening.
284
285         * Skipped:
286
287 2017-01-30  Zalan Bujtas  <zalan@apple.com>
288
289         Simple line layout: Small tweaks to improve performance.
290         https://bugs.webkit.org/show_bug.cgi?id=167611
291         <rdar://problem/30274294>
292
293         Reviewed by Simon Fraser.
294
295         * Layout/simple-line-layout-non-repeating-text.html: Added.
296
297 2017-01-13  Said Abou-Hallawa  <sabouhallawa@apple.com>
298
299         26 MotionMark performance tests failing
300         https://bugs.webkit.org/show_bug.cgi?id=166854
301
302         Reviewed by Ryosuke Niwa.
303         
304         Skip running MotionMark on WebKit perf bots.
305
306         * Skipped:
307
308 2017-01-06  Said Abou-Hallawa  <sabouhallawa@apple.com>
309
310         Rename the directory and the files of MotionMark from Animometer to MotionMark
311         https://bugs.webkit.org/show_bug.cgi?id=166659
312
313         Reviewed by Jon Lee.
314         
315         Rename the directory and the files of the benchmark to its new name.
316
317         * MotionMark/developer.html: Renamed from PerformanceTests/Animometer/developer.html.
318         * MotionMark/index.html: Renamed from PerformanceTests/Animometer/index.html.
319         * MotionMark/resources/debug-runner/d3.min.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/d3.min.js.
320         * MotionMark/resources/debug-runner/graph.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/graph.js.
321         * MotionMark/resources/debug-runner/motionmark.css: Renamed from PerformanceTests/Animometer/resources/debug-runner/animometer.css.
322         * MotionMark/resources/debug-runner/motionmark.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/animometer.js.
323         * MotionMark/resources/debug-runner/tests.js: Renamed from PerformanceTests/Animometer/resources/debug-runner/tests.js.
324         * MotionMark/resources/extensions.js: Renamed from PerformanceTests/Animometer/resources/extensions.js.
325         * MotionMark/resources/runner/benchmark-runner.js: Renamed from PerformanceTests/Animometer/resources/runner/benchmark-runner.js.
326         * MotionMark/resources/runner/crystal.svg: Renamed from PerformanceTests/Animometer/resources/runner/crystal.svg.
327         * MotionMark/resources/runner/lines.svg: Renamed from PerformanceTests/Animometer/resources/runner/lines.svg.
328         * MotionMark/resources/runner/logo.svg: Renamed from PerformanceTests/Animometer/resources/runner/logo.svg.
329         * MotionMark/resources/runner/motionmark.css: Renamed from PerformanceTests/Animometer/resources/runner/animometer.css.
330         * MotionMark/resources/runner/motionmark.js: Renamed from PerformanceTests/Animometer/resources/runner/animometer.js.
331         * MotionMark/resources/runner/tests.js: Renamed from PerformanceTests/Animometer/resources/runner/tests.js.
332         * MotionMark/resources/statistics.js: Renamed from PerformanceTests/Animometer/resources/statistics.js.
333         * MotionMark/resources/strings.js: Renamed from PerformanceTests/Animometer/resources/strings.js.
334         * MotionMark/tests/3d/resources/webgl.js: Renamed from PerformanceTests/Animometer/tests/3d/resources/webgl.js.
335         * MotionMark/tests/3d/webgl.html: Renamed from PerformanceTests/Animometer/tests/3d/webgl.html.
336         * MotionMark/tests/bouncing-particles/bouncing-canvas-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-canvas-images.html.
337         * MotionMark/tests/bouncing-particles/bouncing-canvas-shapes.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html.
338         * MotionMark/tests/bouncing-particles/bouncing-css-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-css-images.html.
339         * MotionMark/tests/bouncing-particles/bouncing-css-shapes.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-css-shapes.html.
340         * MotionMark/tests/bouncing-particles/bouncing-svg-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-svg-images.html.
341         * MotionMark/tests/bouncing-particles/bouncing-svg-shapes.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-svg-shapes.html.
342         * MotionMark/tests/bouncing-particles/bouncing-tagged-images.html: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/bouncing-tagged-images.html.
343         * MotionMark/tests/bouncing-particles/resources/bouncing-canvas-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-canvas-images.js.
344         * MotionMark/tests/bouncing-particles/resources/bouncing-canvas-particles.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-canvas-particles.js.
345         * MotionMark/tests/bouncing-particles/resources/bouncing-canvas-shapes.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-canvas-shapes.js.
346         * MotionMark/tests/bouncing-particles/resources/bouncing-css-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-css-images.js.
347         * MotionMark/tests/bouncing-particles/resources/bouncing-css-shapes.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-css-shapes.js.
348         * MotionMark/tests/bouncing-particles/resources/bouncing-svg-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-svg-images.js.
349         * MotionMark/tests/bouncing-particles/resources/bouncing-svg-particles.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-svg-particles.js.
350         * MotionMark/tests/bouncing-particles/resources/bouncing-svg-shapes.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-svg-shapes.js.
351         * MotionMark/tests/bouncing-particles/resources/bouncing-tagged-images.js: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/bouncing-tagged-images.js.
352         * MotionMark/tests/bouncing-particles/resources/image1.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image1.jpg.
353         * MotionMark/tests/bouncing-particles/resources/image2.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image2.jpg.
354         * MotionMark/tests/bouncing-particles/resources/image3.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image3.jpg.
355         * MotionMark/tests/bouncing-particles/resources/image4.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image4.jpg.
356         * MotionMark/tests/bouncing-particles/resources/image5.jpg: Renamed from PerformanceTests/Animometer/tests/bouncing-particles/resources/image5.jpg.
357         * MotionMark/tests/dom/compositing-transforms.html: Renamed from PerformanceTests/Animometer/tests/dom/compositing-transforms.html.
358         * MotionMark/tests/dom/focus.html: Renamed from PerformanceTests/Animometer/tests/dom/focus.html.
359         * MotionMark/tests/dom/leaves.html: Renamed from PerformanceTests/Animometer/tests/dom/leaves.html.
360         * MotionMark/tests/dom/particles.html: Renamed from PerformanceTests/Animometer/tests/dom/particles.html.
361         * MotionMark/tests/dom/resources/compositing-transforms.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/compositing-transforms.js.
362         * MotionMark/tests/dom/resources/dom-particles.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/dom-particles.js.
363         * MotionMark/tests/dom/resources/focus.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/focus.js.
364         * MotionMark/tests/dom/resources/leaves.js: Renamed from PerformanceTests/Animometer/tests/dom/resources/leaves.js.
365         * MotionMark/tests/master/canvas-stage.html: Renamed from PerformanceTests/Animometer/tests/master/canvas-stage.html.
366         * MotionMark/tests/master/focus.html: Renamed from PerformanceTests/Animometer/tests/master/focus.html.
367         * MotionMark/tests/master/image-data.html: Renamed from PerformanceTests/Animometer/tests/master/image-data.html.
368         * MotionMark/tests/master/leaves.html: Renamed from PerformanceTests/Animometer/tests/master/leaves.html.
369         * MotionMark/tests/master/multiply.html: Renamed from PerformanceTests/Animometer/tests/master/multiply.html.
370         * MotionMark/tests/master/resources/canvas-stage.js: Renamed from PerformanceTests/Animometer/tests/master/resources/canvas-stage.js.
371         * MotionMark/tests/master/resources/canvas-tests.js: Renamed from PerformanceTests/Animometer/tests/master/resources/canvas-tests.js.
372         * MotionMark/tests/master/resources/compass.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/compass.svg.
373         * MotionMark/tests/master/resources/compass100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/compass100.png.
374         * MotionMark/tests/master/resources/console.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/console.svg.
375         * MotionMark/tests/master/resources/console100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/console100.png.
376         * MotionMark/tests/master/resources/contribute.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/contribute.svg.
377         * MotionMark/tests/master/resources/contribute100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/contribute100.png.
378         * MotionMark/tests/master/resources/debugger.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/debugger.svg.
379         * MotionMark/tests/master/resources/debugger100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/debugger100.png.
380         * MotionMark/tests/master/resources/focus.js: Renamed from PerformanceTests/Animometer/tests/master/resources/focus.js.
381         * MotionMark/tests/master/resources/image-data.js: Renamed from PerformanceTests/Animometer/tests/master/resources/image-data.js.
382         * MotionMark/tests/master/resources/inspector.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/inspector.svg.
383         * MotionMark/tests/master/resources/inspector100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/inspector100.png.
384         * MotionMark/tests/master/resources/layout.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/layout.svg.
385         * MotionMark/tests/master/resources/layout100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/layout100.png.
386         * MotionMark/tests/master/resources/leaves.js: Renamed from PerformanceTests/Animometer/tests/master/resources/leaves.js.
387         * MotionMark/tests/master/resources/multiply.js: Renamed from PerformanceTests/Animometer/tests/master/resources/multiply.js.
388         * MotionMark/tests/master/resources/particles.js: Renamed from PerformanceTests/Animometer/tests/master/resources/particles.js.
389         * MotionMark/tests/master/resources/performance.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/performance.svg.
390         * MotionMark/tests/master/resources/performance100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/performance100.png.
391         * MotionMark/tests/master/resources/script.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/script.svg.
392         * MotionMark/tests/master/resources/script100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/script100.png.
393         * MotionMark/tests/master/resources/shortcuts.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/shortcuts.svg.
394         * MotionMark/tests/master/resources/shortcuts100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/shortcuts100.png.
395         * MotionMark/tests/master/resources/standards.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/standards.svg.
396         * MotionMark/tests/master/resources/standards100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/standards100.png.
397         * MotionMark/tests/master/resources/storage.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/storage.svg.
398         * MotionMark/tests/master/resources/storage100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/storage100.png.
399         * MotionMark/tests/master/resources/styles.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/styles.svg.
400         * MotionMark/tests/master/resources/styles100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/styles100.png.
401         * MotionMark/tests/master/resources/svg-particles.js: Renamed from PerformanceTests/Animometer/tests/master/resources/svg-particles.js.
402         * MotionMark/tests/master/resources/text.js: Renamed from PerformanceTests/Animometer/tests/master/resources/text.js.
403         * MotionMark/tests/master/resources/timeline.svg: Renamed from PerformanceTests/Animometer/tests/master/resources/timeline.svg.
404         * MotionMark/tests/master/resources/timeline100.png: Renamed from PerformanceTests/Animometer/tests/master/resources/timeline100.png.
405         * MotionMark/tests/master/svg-particles.html: Renamed from PerformanceTests/Animometer/tests/master/svg-particles.html.
406         * MotionMark/tests/master/text.html: Renamed from PerformanceTests/Animometer/tests/master/text.html.
407         * MotionMark/tests/resources/main.js: Renamed from PerformanceTests/Animometer/tests/resources/main.js.
408         * MotionMark/tests/resources/math.js: Renamed from PerformanceTests/Animometer/tests/resources/math.js.
409         * MotionMark/tests/resources/stage.css: Renamed from PerformanceTests/Animometer/tests/resources/stage.css.
410         * MotionMark/tests/resources/star.svg: Renamed from PerformanceTests/Animometer/tests/resources/star.svg.
411         * MotionMark/tests/resources/yin-yang.png: Renamed from PerformanceTests/Animometer/tests/resources/yin-yang.png.
412         * MotionMark/tests/resources/yin-yang.svg: Renamed from PerformanceTests/Animometer/tests/resources/yin-yang.svg.
413         * MotionMark/tests/simple/resources/simple-canvas-paths.js: Renamed from PerformanceTests/Animometer/tests/simple/resources/simple-canvas-paths.js.
414         * MotionMark/tests/simple/resources/simple-canvas.js: Renamed from PerformanceTests/Animometer/tests/simple/resources/simple-canvas.js.
415         * MotionMark/tests/simple/resources/tiled-canvas-image.js: Renamed from PerformanceTests/Animometer/tests/simple/resources/tiled-canvas-image.js.
416         * MotionMark/tests/simple/simple-canvas-paths.html: Renamed from PerformanceTests/Animometer/tests/simple/simple-canvas-paths.html.
417         * MotionMark/tests/simple/tiled-canvas-image.html: Renamed from PerformanceTests/Animometer/tests/simple/tiled-canvas-image.html.
418         * MotionMark/tests/template/resources/template-canvas.js: Renamed from PerformanceTests/Animometer/tests/template/resources/template-canvas.js.
419         * MotionMark/tests/template/resources/template-css.js: Renamed from PerformanceTests/Animometer/tests/template/resources/template-css.js.
420         * MotionMark/tests/template/resources/template-svg.js: Renamed from PerformanceTests/Animometer/tests/template/resources/template-svg.js.
421         * MotionMark/tests/template/template-canvas.html: Renamed from PerformanceTests/Animometer/tests/template/template-canvas.html.
422         * MotionMark/tests/template/template-css.html: Renamed from PerformanceTests/Animometer/tests/template/template-css.html.
423         * MotionMark/tests/template/template-svg.html: Renamed from PerformanceTests/Animometer/tests/template/template-svg.html.
424
425 2016-12-13  Brady Eidson  <beidson@apple.com>
426
427         Add a PerfTest targeting IDBObjectStore.get().
428         https://bugs.webkit.org/show_bug.cgi?id=165816
429
430         Reviewed by Alex Christensen.
431
432         * IndexedDB/objectstore-get.html: Added.
433
434 2016-12-13  Brady Eidson  <beidson@apple.com>
435
436         Add a PerfTest targeting IDBIndex.get().
437         https://bugs.webkit.org/show_bug.cgi?id=165803
438
439         Reviewed by Alex Christensen.
440
441         * IndexedDB/index-get.html: Added.
442
443 2016-12-12  Brady Eidson  <beidson@apple.com>
444
445         More IndexedDB perf tests.
446         https://bugs.webkit.org/show_bug.cgi?id=165634
447
448         Reviewed by Sam Weinig.
449
450         * IndexedDB/index-multientry.html:
451         * IndexedDB/large-number-of-inserts-responsiveness.html: Added.
452         * IndexedDB/large-number-of-inserts.html: Added.
453         * IndexedDB/objectstore-cursor.html: Added. Runtime cut in 1/4th compared to r209672.
454
455         * resources/runner.js:
456
457 2016-12-12  Commit Queue  <commit-queue@webkit.org>
458
459         Unreviewed, rolling out r209672.
460         https://bugs.webkit.org/show_bug.cgi?id=165766
461
462         IndexedDB/objectstore-cursor.html perf test always times out
463         (Requested by ap on #webkit).
464
465         Reverted changeset:
466
467         "More IndexedDB perf tests."
468         https://bugs.webkit.org/show_bug.cgi?id=165634
469         http://trac.webkit.org/changeset/209672
470
471 2016-12-10  Brady Eidson  <beidson@apple.com>
472
473         More IndexedDB perf tests.
474         https://bugs.webkit.org/show_bug.cgi?id=165634
475
476         Reviewed by Sam Weinig.
477
478         * IndexedDB/index-multientry.html:
479         * IndexedDB/large-number-of-inserts-responsiveness.html: Added.
480         * IndexedDB/large-number-of-inserts.html: Added.
481         * IndexedDB/objectstore-cursor.html: Added.
482         
483         * resources/runner.js: Add "track responsiveness" functionality to PerfTestRunner.
484
485 2016-12-06  Filip Pizlo  <fpizlo@apple.com>
486
487         Concurrent GC should be stable enough to land enabled
488         https://bugs.webkit.org/show_bug.cgi?id=164990
489
490         Reviewed by Geoffrey Garen.
491         
492         Made CDjs more configurable and refined the "large.js" configuration. I was using that one and
493         the new "long.js" configuration to tune concurrent eden GCs.
494         
495         Added a new way of running Splay in browser, which using chartjs to plot the execution times of
496         2000 iterations. This includes the minified chartjs.
497
498         * JetStream/Octane2/splay-detail.html: Added.
499         * JetStream/cdjs/benchmark.js:
500         (benchmarkImpl):
501         (benchmark):
502         * JetStream/cdjs/long.js: Added.
503
504 2016-12-07  Brady Eidson  <beidson@apple.com>
505
506         Add IDB perf tests stressing key size.
507         https://bugs.webkit.org/show_bug.cgi?id=165567
508
509         Reviewed by Alex Christensen.
510
511         * IndexedDB/large-array-keys.html: Added.
512         * IndexedDB/large-binary-keys.html: Added.
513         * IndexedDB/large-string-keys.html: Added.
514
515 2016-12-05  Brady Eidson  <beidson@apple.com>
516
517         Add an IndexedDB perf test to PerformanceTests.
518         https://bugs.webkit.org/show_bug.cgi?id=165430
519
520         Reviewed by Alex Christensen.
521
522         * IndexedDB/index-multientry.html: Added.
523
524 2016-12-02  Filip Pizlo  <fpizlo@apple.com>
525
526         ES6SampleBench should report an average for Steady State so that all of the numbers are comparable
527         https://bugs.webkit.org/show_bug.cgi?id=165325
528
529         Reviewed by Saam Barati.
530
531         This makes all of the numbers that ES6SampleBench reports comparable to each other: they all speak of the time
532         it took to run an iteration of something.
533
534         * ES6SampleBench/results.js:
535         (Results.prototype.reportResult):
536
537 2016-11-18  Jer Noble  <jer.noble@apple.com>
538
539         Add some performance tests for various aspects of media loading and playback.
540         https://bugs.webkit.org/show_bug.cgi?id=164977
541
542         Reviewed by Darin Adler.
543
544         Add new tests for a few aspects of media playback; namely: how quickly  media elements fire
545         the "canplaythrough" event when all data is immediately available, how quickly HLS playback
546         switches resolutions, and at what playback rate MSE-backed video can play without dropping
547         frames. Skip these tests by default, as they're not runnable on all ports, and they require
548         a webserver (run-webkit-httpd) to complete on macOS.
549
550         * Media/HLSCanPlayThrough.html: Added.
551         * Media/HLSGearChange.html: Added.
552         * Media/MSECanPlayThrough.html: Added.
553         * Media/MSEPlaybackRate.html: Added.
554         * Media/StandardCanPlayThrough.html: Added.
555         * Media/hls/1080p/iframe_index.m3u8: Added.
556         * Media/hls/1080p/prog_index.m3u8: Added.
557         * Media/hls/1080p/test.ts: Added.
558         * Media/hls/480p/iframe_index.m3u8: Added.
559         * Media/hls/480p/prog_index.m3u8: Added.
560         * Media/hls/480p/test.ts: Added.
561         * Media/hls/720p/iframe_index.m3u8: Added.
562         * Media/hls/720p/prog_index.m3u8: Added.
563         * Media/hls/720p/test.ts: Added.
564         * Media/hls/720p/test.ts.back: Added.
565         * Media/hls/index.m3u8: Added.
566         * Media/media-source-loader.js: Added.
567         (MediaSourceLoader):
568         (MediaSourceLoader.prototype.loadManifest):
569         (MediaSourceLoader.prototype.loadManifestSucceeded):
570         (MediaSourceLoader.prototype.loadManifestFailed):
571         (MediaSourceLoader.prototype.loadMediaData):
572         (MediaSourceLoader.prototype.loadMediaDataSucceeded):
573         (MediaSourceLoader.prototype.loadMediaDataFailed):
574         (MediaSourceLoader.prototype.get type):
575         (MediaSourceLoader.prototype.get duration):
576         (MediaSourceLoader.prototype.get initSegment):
577         (MediaSourceLoader.prototype.get mediaSegmentsLength):
578         (MediaSourceLoader.prototype.mediaSegments):
579         (MediaSourceLoader.prototype.get everyMediaSegment):
580         * Media/test-fragmented-video.json: Added.
581         * Media/test-fragmented-video.mp4: Added.
582         * Media/test.mp4: Added.
583         * Skipped:
584
585 2016-11-18  Filip Pizlo  <fpizlo@apple.com>
586
587         Concurrent GC should be able to run splay in debug mode and earley/raytrace in release mode with no perf regression
588         https://bugs.webkit.org/show_bug.cgi?id=164282
589
590         Reviewed by Geoffrey Garen and Oliver Hunt.
591         
592         CDjs is a fun benchmark for stressing concurrent GCs, but to really give the GC a good
593         workout you need to increase the amount of work that the test does. This adds a second
594         configuration of the benchmark that has more aircraft. It uses much more memory and causes us
595         to do more GCs and those GCs take longer.
596
597         * JetStream/cdjs/benchmark.js:
598         (benchmarkImpl):
599         (benchmark):
600         * JetStream/cdjs/large.js: Added.
601
602 2016-11-14  Filip Pizlo  <fpizlo@apple.com>
603
604         Unreviewed, revert unintended change.
605
606         * ES6SampleBench/Air/benchmark.js:
607
608 2016-10-03  Saam Barati  <sbarati@apple.com>
609
610         MapHash should speculate on the type of its child node
611         https://bugs.webkit.org/show_bug.cgi?id=161922
612
613         Reviewed by Filip Pizlo.
614
615         I gave the main generator a name so it's easier to see what
616         it is when using the sampling profiler.
617
618         * ES6SampleBench/Basic/ast.js:
619         (Basic.Program):
620
621 2016-09-19  Sergio Villar Senin  <svillar@igalia.com>
622
623         [css-grid] Remove the x2 computation of row sizes with indefinite heights
624         https://bugs.webkit.org/show_bug.cgi?id=162150
625
626         Reviewed by Darin Adler.
627
628         Added a new test case which checks the layout performance of grids inside other grids, i.e,
629         grids acting both as grid container and grid item.
630
631         * Layout/nested-grid.html: Added.
632
633 2016-09-09  Simon Fraser  <simon.fraser@apple.com>
634
635         Perf test Animation/css-accelerated-animation.html failing
636         https://bugs.webkit.org/show_bug.cgi?id=161795
637
638         Skip Animation/css-accelerated-animation.html since this only works as a MobileSafari
639         perf test.
640
641         * Skipped:
642
643 2016-09-08  Simon Fraser  <simon.fraser@apple.com>
644
645         Add a content-animation test with accelerated CSS animations
646         https://bugs.webkit.org/show_bug.cgi?id=161776
647
648         Reviewed by Dean Jackson.
649
650         Add a test that measures the frame rate of accelerated CSS animations.
651
652         This is like css-animation.html (which animates 'left' and 'top') but animates
653         transforms instead. In order to get animation in X and Y we make two nested elements
654         and set transformX() on one, and transformY() on the other.
655
656         * Animation/css-accelerated-animation.html: Added.
657
658 2016-08-23  Saam Barati  <sbarati@apple.com>
659
660         It should be easy to run ES6SampleBench from the jsc shell
661         https://bugs.webkit.org/show_bug.cgi?id=161085
662
663         Reviewed by Yusuke Suzuki.
664
665         This patch makes ES6 sample bench runnable through the `jsc` shell.
666         To do that, you need to be in the PerformanceTests/ES6SampleBench
667         directory and run `jsc cli.js`. To make this work, the benchmark
668         is now aware of being run in two modes: via the browser, and via
669         the shell. Each entry point will set a variable `isInBrowser`,
670         and the benchmark will do different things based on if that
671         variable is true or false.
672
673         * ES6SampleBench/Air/benchmark.js:
674         * ES6SampleBench/Air/stress-test.js:
675         * ES6SampleBench/Basic/benchmark.js:
676         (runBenchmark):
677         * ES6SampleBench/Basic/stress-test.js:
678         * ES6SampleBench/air_benchmark.js:
679         * ES6SampleBench/basic_benchmark.js:
680         * ES6SampleBench/cli.js: Added.
681         (return.doRun):
682         (makeBenchmarkRunner):
683         * ES6SampleBench/driver.js:
684         (Driver):
685         (Driver.prototype._recomputeSummary):
686         (Driver.prototype._iterate.window.setTimeout):
687         (Driver.prototype._iterate):
688         (Driver.prototype._updateIterations):
689         * ES6SampleBench/glue.js:
690         (reportResult):
691         * ES6SampleBench/index.html:
692         * ES6SampleBench/results.js:
693         (Results):
694         (Results.prototype.reportRunning):
695         (Results.prototype.reportDone):
696         (Results.prototype.reportError):
697         * ES6SampleBench/stats.js:
698         (Stats):
699         (Stats.prototype.toString):
700         (Stats.prototype._update):
701
702 2016-08-01  Said Abou-Hallawa  <sabouhallawa@apple.com>
703
704         Add an option to run the MotionMark tests with classic tile size (512x512)
705         https://bugs.webkit.org/show_bug.cgi?id=160371
706
707         Reviewed by Darin Adler.
708
709         Setting the body to large size (3000x3000) and overflow to scroll forces
710         the classic tile size (512x512). This option does not affect the canvas
711         tests because the canvas is displayed on one tile regardless of its size.
712
713         * Animometer/developer.html:
714         * Animometer/resources/debug-runner/animometer.css:
715         (body.showing-test-container.tiles-big):
716         (body.showing-test-container.tiles-classic):
717         * Animometer/resources/debug-runner/animometer.js:
718         * Animometer/resources/runner/animometer.js:
719         (window.benchmarkController.startBenchmark):
720
721 2016-06-29  Filip Pizlo  <fpizlo@apple.com>
722
723         Generators violate bytecode liveness validation
724         https://bugs.webkit.org/show_bug.cgi?id=159279
725
726         Reviewed by Yusuke Suzuki.
727
728         Add Basic to our test harness.
729
730         Also made some cosmetic changes to the benchmark harness.
731
732         * ES6SampleBench/Basic/basic-tests.yaml: Added.
733         * ES6SampleBench/Basic/stress-test.js: Added.
734         (preciseTime):
735         * ES6SampleBench/driver.js:
736         (Driver):
737         (Driver.prototype.start):
738         (Driver.prototype.reportError):
739         * ES6SampleBench/glue.js:
740         * ES6SampleBench/index.html:
741
742 2016-06-28  Filip Pizlo  <fpizlo@apple.com>
743
744         ES6SampleBench should have a harness
745         https://bugs.webkit.org/show_bug.cgi?id=159246
746
747         Reviewed by Saam Barati.
748         
749         This adds a simple web harness for ES6SampleBench. It runs Air and Basic 10 times for 200
750         iterations each and reports three metrics:
751         
752         First iteration: the time it takes for the first iteration to run. This is the first
753         iteration after the benchmark is loaded into the iframe, so it's representative of what
754         would happen if one of these workloads only ran for a short time.
755         
756         Worst 2%: of the last 199 iterations, the average of the worst 2% iterations. If code like
757         any of these workloads was used in an important event handler, you'd want that code to run
758         well in the worst case in addition to having great throughput.
759         
760         Steady state: the total of the last 199 iterations. This is representative of what would
761         happen if you ran code like this for a long time.
762         
763         The total score is the geomean of the firstIteration/worstCase/steadyState numbers of the
764         two benchmarks.
765         
766         The harness does statistics using Student's T-distribution confidence intervals.
767
768         * ES6SampleBench/Basic/benchmark.js:
769         (Benchmark):
770         * ES6SampleBench/air_benchmark.js: Added.
771         * ES6SampleBench/basic_benchmark.js: Added.
772         * ES6SampleBench/driver.js: Added.
773         (Driver):
774         (Driver.prototype.addBenchmark):
775         (Driver.prototype.start):
776         (Driver.prototype.reportResult):
777         (Driver.prototype.reportError):
778         (Driver.prototype._recomputeSummary.Geomean):
779         (Driver.prototype._recomputeSummary.Geomean.prototype.add):
780         (Driver.prototype._recomputeSummary.Geomean.prototype.get result):
781         (Driver.prototype._recomputeSummary):
782         (Driver.prototype._iterate):
783         (Driver.prototype._updateIterations):
784         * ES6SampleBench/glue.js: Added.
785         * ES6SampleBench/index.html: Added.
786         * ES6SampleBench/results.js: Added.
787         (Results):
788         (Results.prototype.get benchmark):
789         (Results.prototype.reset):
790         (Results.prototype.reportRunning):
791         (Results.prototype.reportDone):
792         (Results.prototype.reportResult.averageAbovePercentile):
793         (Results.prototype.reportResult):
794         (Results.prototype.reportError):
795         * ES6SampleBench/stats.js: Added.
796         (Stats):
797         (Stats.prototype.reset):
798         (Stats.prototype.add):
799         (Stats.prototype.get numIterations):
800         (Stats.prototype.valueForIteration):
801         (Stats.get result.tDist):
802         (Stats.prototype.get result):
803         (Stats.prototype.toString):
804         (Stats.prototype._update):
805         * ES6SampleBench/style.css: Added.
806         (body):
807         (body, th, tr):
808         (h1, h2, h3, h4):
809         (h1):
810         (h2):
811         (h3):
812         (hr):
813         (address):
814         (img):
815         (.underline):
816         (ol.loweralpha):
817         (ol.upperalpha):
818         (ol.lowerroman):
819         (ol.upperroman):
820         (ol.arabic):
821         (.banner-link:link, .banner-link:visited):
822         (:link, :visited):
823         (h4 :link, h4 :visited, h5 :link, h5 :visited, h6 :link, h6 :visited):
824         (.anchor:link, .anchor:visited):
825         (* > .anchor:link, * > .anchor:visited):
826         (span:hover .anchor):
827         (a.forbidden, span.forbidden):
828         (a.missing:hover):
829         (a.closed:link, a.closed:visited, span.closed):
830         (pre):
831         (div.code):
832         (div.code pre):
833         (dt):
834         (dd):
835         (dd:last-child):
836         (.site-logo):
837         (.site-logo .tagline):
838         (table):
839         (#contents):
840         (p):
841         (p:last-child):
842         (th):
843         (td):
844
845 2016-06-28  Jon Lee  <jonlee@apple.com>
846
847         Update focus test
848         https://bugs.webkit.org/show_bug.cgi?id=159242
849         rdar://problem/27070007
850
851         Reviewed by Dean Jackson.
852         Provisionally reviewed by Said Abou-Hallawa.
853
854         Move previous test to dom suite, and update the test for better reporting of frame rate, although
855         it uses a different rendering path.
856
857         * Animometer/resources/debug-runner/tests.js: Add to dom suite.
858         * Animometer/tests/dom/focus.html: Copied from PerformanceTests/Animometer/tests/master/focus.html.
859         * Animometer/tests/dom/resources/focus.js: Copied from PerformanceTests/Animometer/tests/master/resources/focus.js.
860         * Animometer/tests/master/focus.html: Remove center element.
861         * Animometer/tests/master/resources/focus.js: Use narrower size range with smaller particles. Remove the
862         container elements. Inline getBlurValue and getOpacityValue since they are only called once.
863
864 2016-06-28  Filip Pizlo  <fpizlo@apple.com>
865
866         Move Air.js and Basic into ES6SampleBench
867
868         Rubber stamped by Geoffrey Garen.
869         
870         We want to group our ES6 benchmarks together, and eventually, we want to give them a
871         common harness.
872
873         * Air.js: Removed.
874         * Air.js/README.md: Removed.
875         * Air.js/airjs-tests.yaml: Removed.
876         * Air.js/all.js: Removed.
877         * Air.js/allocate_stack.js: Removed.
878         * Air.js/arg.js: Removed.
879         * Air.js/basic_block.js: Removed.
880         * Air.js/benchmark.js: Removed.
881         * Air.js/code.js: Removed.
882         * Air.js/custom.js: Removed.
883         * Air.js/frequented_block.js: Removed.
884         * Air.js/insertion_set.js: Removed.
885         * Air.js/inst.js: Removed.
886         * Air.js/liveness.js: Removed.
887         * Air.js/make_dist.sh: Removed.
888         * Air.js/opcode.js: Removed.
889         * Air.js/payload-airjs-ACLj8C.js: Removed.
890         * Air.js/payload-gbemu-executeIteration.js: Removed.
891         * Air.js/payload-imaging-gaussian-blur-gaussianBlur.js: Removed.
892         * Air.js/payload-typescript-scanIdentifier.js: Removed.
893         * Air.js/reg.js: Removed.
894         * Air.js/stack_slot.js: Removed.
895         * Air.js/stress-test.js: Removed.
896         * Air.js/strip-hash.rb: Removed.
897         * Air.js/symbols.js: Removed.
898         * Air.js/test.html: Removed.
899         * Air.js/test.js: Removed.
900         * Air.js/tmp.js: Removed.
901         * Air.js/tmp_base.js: Removed.
902         * Air.js/util.js: Removed.
903         * Basic: Removed.
904         * Basic/ast.js: Removed.
905         * Basic/basic.js: Removed.
906         * Basic/benchmark.js: Removed.
907         * Basic/caseless_map.js: Removed.
908         * Basic/lexer.js: Removed.
909         * Basic/number.js: Removed.
910         * Basic/parser.js: Removed.
911         * Basic/random.js: Removed.
912         * Basic/state.js: Removed.
913         * Basic/test.html: Removed.
914         * Basic/test.js: Removed.
915         * Basic/util.js: Removed.
916         * ES6SampleBench: Added.
917         * ES6SampleBench/Air: Copied from PerformanceTests/Air.js.
918         * ES6SampleBench/Basic: Copied from PerformanceTests/Basic.
919         * Skipped:
920
921 2016-06-24  Filip Pizlo  <fpizlo@apple.com>
922
923         Add a ES6 generator benchmark
924         https://bugs.webkit.org/show_bug.cgi?id=159101
925
926         Rubber stamped by Keith Miller.
927         
928         This adds a Basic interpreter loosely based on ECMA-55:
929         
930         http://www.ecma-international.org/publications/files/ECMA-ST-WITHDRAWN/ECMA-55,%201st%20Edition,%20January%201978.pdf
931         
932         It includes a lexer that is a generator, a parser that uses regular expressions, and an AST
933         walk interpreter where the walking functions for statements are generators that call each
934         other with yield*. This enables the interpreter to look like an AST walk even though it can
935         yield at INPUT and PRINT statements.
936         
937         This also uses for-of, classes, Map, and WeakMap. It also uses deprecated-but-awesome RegExp
938         features like RegExp.lastMatch and RegExp.rightContext. I did it that way because this is
939         how I've always written lexers in dynamic languages; see offlineasm's lex() method in
940         parser.rb for example.
941         
942         The benchmark runs a handful of simple Basic programs. The longest-running one computes
943         prime numbers.
944         
945         Includes a command-line and web harness. On my machine it runs in 2-3 seconds.
946
947         * Basic: Added.
948         * Basic/ast.js: Added.
949         (Basic.NumberApply):
950         (Basic.Variable):
951         (Basic.Const):
952         (Basic.NumberPow):
953         (Basic.NumberMul):
954         (Basic.NumberDiv):
955         (Basic.NumberNeg):
956         (Basic.NumberAdd):
957         (Basic.NumberSub):
958         (Basic.StringVar):
959         (Basic.Equals):
960         (Basic.NotEquals):
961         (Basic.LessThan):
962         (Basic.GreaterThan):
963         (Basic.LessEqual):
964         (Basic.GreaterEqual):
965         (Basic.GoTo):
966         (Basic.GoSub):
967         (Basic.Def):
968         (Basic.Let):
969         (Basic.If):
970         (Basic.Return):
971         (Basic.Stop):
972         (Basic.On):
973         (sideState.shouldStop):
974         (Basic.For):
975         (Basic.Next):
976         (Basic.Print):
977         (Basic.Input):
978         (Basic.Read):
979         (Basic.Restore):
980         (Basic.Dim):
981         (Basic.Randomize):
982         (Basic.End):
983         (Basic.Program):
984         * Basic/basic.js: Added.
985         (prepare):
986         (simulate):
987         * Basic/benchmark.js: Added.
988         (Benchmark):
989         (Benchmark.prototype.runIteration.expect):
990         (Benchmark.prototype.runIteration):
991         (runBenchmark):
992         * Basic/caseless_map.js: Added.
993         (CaselessMap):
994         * Basic/lexer.js: Added.
995         (lex.consumeWhitespace):
996         (lex.consume):
997         (lex):
998         * Basic/lexer_test.js: Added.
999         * Basic/number.js: Added.
1000         (NumberValue):
1001         (NumberValue.prototype.get value):
1002         (NumberValue.prototype.apply):
1003         (NumberValue.prototype.leftApply):
1004         (NumberValue.prototype.assign):
1005         (NumberArray.):
1006         (NumberArray):
1007         (NumberArray.prototype.apply):
1008         (NumberArray.prototype.leftApply):
1009         (NumberFunction):
1010         (NumberFunction.prototype.apply):
1011         (NumberFunction.prototype.leftApply):
1012         (NativeFunction):
1013         (NativeFunction.prototype.apply):
1014         (NativeFunction.prototype.leftApply):
1015         * Basic/parser.js: Added.
1016         (parse):
1017         (parse.pushToken):
1018         (parse.peekToken):
1019         (parse.consumeKind):
1020         (parse.consumeToken):
1021         (parse.parseVariable):
1022         (parse.parseNumericExpression.parsePrimary):
1023         (parse.parseNumericExpression.parseFactor):
1024         (parse.parseNumericExpression.parseTerm):
1025         (parse.parseNumericExpression):
1026         (parse.parseConstant):
1027         (parse.parseStringExpression):
1028         (parse.isStringExpression):
1029         (parse.parseRelationalExpression):
1030         (parse.parseNonNegativeInteger):
1031         (parse.parseGoToStatement):
1032         (parse.parseGoSubStatement):
1033         (parse.parseStatement):
1034         (parse.parseStatements):
1035         * Basic/random.js: Added.
1036         (createRNG):
1037         (createRNGWithFixedSeed):
1038         (createRNGWithRandomSeed):
1039         * Basic/state.js: Added.
1040         (State):
1041         (State.prototype.getValue):
1042         (State.prototype.getSideState):
1043         (State.prototype.abort):
1044         (State.prototype.validate):
1045         * Basic/test.html: Added.
1046         * Basic/test.js: Added.
1047         * Basic/util.js: Added.
1048         (this.performance.performance.now.currentTime):
1049         (else.this.preciseTime.currentTime):
1050         (else.currentTime):
1051         * Skipped: Make sure that we don't run Basic yet.
1052
1053 2016-06-21  Jon Lee  <jonlee@apple.com>
1054
1055         Update canvas size when benchmark begins
1056         https://bugs.webkit.org/show_bug.cgi?id=159010
1057
1058         Reviewed by Dean Jackson.
1059
1060         Previously we would evaluate the media queries and assign the canvas size to the test
1061         on body.onload. Instead, do it once the user starts the benchmark.
1062
1063         * Animometer/resources/debug-runner/animometer.js: Move benchmarkController.determineCanvasSize()
1064         to benchmarkController._startBenchmark.
1065         * Animometer/resources/runner/animometer.js: Ditto.
1066
1067 2016-06-21  Jon Lee  <jonlee@apple.com>
1068
1069         Add new timestamp option
1070         https://bugs.webkit.org/show_bug.cgi?id=159006
1071
1072         Reviewed by Dean Jackson.
1073
1074         Add a new option to take timestamps from the rAF callback.
1075
1076         * Animometer/developer.html: Added option.
1077         * Animometer/tests/resources/main.js:
1078         (Benchmark): Fall back to using the rAF timestamp if performance.now()
1079         is not available.
1080         (Benchmark._animateLoop): Update how timestamp is set. Prefer to use
1081         local var instead of accessing private var.
1082
1083 2016-06-21  Jon Lee  <jonlee@apple.com>
1084
1085         Improvements to Animometer benchmark
1086         https://bugs.webkit.org/show_bug.cgi?id=157738
1087
1088         Reviewed by Dean Jackson.
1089         Provisionally reviewed by Said Abou-Hallawa.
1090
1091         Update tests.
1092
1093         * Animometer/tests/master/text.html: Ensure only three text sizes for
1094         the three canvases.
1095         * Animometer/tests/master/focus.html: Reduce the text size to fit with smaller
1096         particle sizes.
1097         * Animometer/tests/master/resources/focus.js: Remove the quadratic distribution for
1098         particle sizes, and make it linear. Reduce the size variance. Shuffle the math to
1099         reduce some calculations per frame. Fix the placement of the particles which might
1100         otherwise be culled.
1101         * Animometer/tests/master/resources/image-data.js: Reduce the particle size
1102         to encourage larger scores.
1103
1104 2016-06-21  Jon Lee  <jonlee@apple.com>
1105
1106         Improvements to Animometer benchmark
1107         https://bugs.webkit.org/show_bug.cgi?id=157738
1108
1109         Reviewed by Dean Jackson.
1110         Provisionally reviewed by Said Abou-Hallawa.
1111
1112         Include confidence interval for the final score, and store the canvas
1113         size in the serialization so that it is accurately shown in results.
1114
1115         * Animometer/developer.html: Add a "confidence" div.
1116         * Animometer/index.html: Ditto. Convert "mean" to "confidence".
1117         * Animometer/resources/debug-runner/animometer.js: Look at options, and
1118         if the configuration is included, update the body class based on it
1119         (similar to what we do when we first load the page). That way, if you
1120         drag-and-drop previous results in, that configuration is reflected in
1121         the dashboard. Show the full confidence interval.
1122
1123         * Animometer/resources/debug-runner/animometer.css:
1124         * Animometer/resources/debug-runner/animometer.js: Style update.
1125         * Animometer/resources/runner/animometer.css:
1126
1127         * Animometer/resources/runner/animometer.js:
1128         (_processData): Propagate the confidence interval values out and calculate
1129         the lower and upper bounds. For now, shortcut the aggregate calculation,
1130         since we only go through one iteration.
1131         (this._processData.calculateScore): Propagate the confidence interval out
1132         to be next to the score. Depending on the controller these values are
1133         calculated differently.
1134         (this._processData._getResultsProperty): Convenience function.
1135         (this._processData.get score): Refactor.
1136         (this._processData.get scoreLowerBound): Get the aggregate lower bound.
1137         (this._processData.get scoreUpperBound): Get the aggregate upper bound.
1138         (window.sectionsManager.setSectionScore):
1139         (window.benchmarkController._startBenchmark): When the benchmark starts, note
1140         the canvas size and add it to options. That way it will be automatically be
1141         serialized.
1142         (window.benchmarkController.showResults): Include the maximum deviation
1143         percentage.
1144         * Animometer/resources/runner/lines.svg: Make the background line up with the
1145         skew.
1146         * Animometer/resources/runner/tests.js:
1147         (Headers.details.text): Refactor.
1148         * Animometer/resources/statistics.js:
1149         (Statistics.largestDeviationPercentage): Convenience function to calculate
1150         the deviation percentage on either end and return the largest deviation.
1151         * Animometer/resources/strings.js:
1152
1153         Allow specifying a regression profile to use instead of taking the one
1154         with the lowest error.
1155
1156         Address an issue in the focus test when the regression calculated ends
1157         up overestimating the change point, causing a cascade of tougher
1158         ramps. The reason behind this is that at max complexity of an initial
1159         ramp, the frame length is very high, and it influences the second
1160         segment of the piecewise regression strongly, causing it to be very
1161         steep. As a result, the first segment, expected to be flat, ends up
1162         covering a higher range of complexity. That makes the change point
1163         much higher than it should be. To avoid this, we will add a sanity
1164         check on the maximum value of the ramp. If the regression's projected
1165         value at the maximum complexity of the current ramp is very slow (less
1166         than 20 fps), 1) reduce the maximum complexity by 20%, and 2) do not
1167         include the regression's change point in the change point estimator.
1168         That estimator is used as the midpoint of the next ramp, and including
1169         the change point from a poor regression can bake in the error. The
1170         controller already knows how to adjust for ramps that are too easy for
1171         the system.
1172
1173         * Animometer/resources/runner/animometer.js:
1174         (this._processData.findRegression): Takes a preferred profile and gives that to
1175         Regression.
1176         (this._processData.calculateScore): With the ramp controller, take the profile
1177         of the ramp that was used the most when calculating the ramp's regression. That
1178         profile is what is used for the test's score.
1179         * Animometer/resources/statistics.js:
1180         (Regression.Utilities.createClass): Update to take an options object which can
1181         specify a profile to calculate with. Otherwise it will continue to use both and
1182         select the one with the lower error.
1183         (_calculateRegression): Fix an issue where we claim 0 error if the regression
1184         calculation fails due to divide-by-zero. Instead reject that regression calculation
1185         by giving it Number.MAX_VALUE.
1186         * Animometer/resources/strings.js: New strings for marking a regression as flat
1187         or slope.
1188         * Animometer/tests/resources/main.js:
1189         (RampController): Rename the thresholds for clarity. Add a threshold that, if
1190         exceeded, will lower the maximum complexity of the next ramp.
1191         (tune): Relax the cdf check to consider whether the interval definitely falls in
1192         the desired frame length threshold.
1193         (processSamples): Include the profile in the ramp.
1194
1195         Update ramp controller test. Increase the length of the test to 30 seconds, and extend
1196         the interval to 120 ms during sampling. Improve the estimation of the ramp parameters.
1197
1198         * Animometer/developer.html: Change default to 30 seconds, and don't show the progress bar
1199         by default.
1200         * Animometer/resources/runner/animometer.js: Change default to 30 seconds.
1201         * Animometer/tests/resources/main.js: A number of improvements to the ramp controller, in
1202         the order in which they appear in the patch:
1203
1204         - With a longer test length use longer ramps with longer intervals to get more data at each
1205         complexity. Keep the 100 ms interval length during the ramp up phase since we don't need to
1206         spend more time there to find the right order of magnitude, but increase it during the
1207         ramps to 120 ms.
1208         - The ramp linearly interpolates the complexity to render based on its timestamp, but it would
1209         never sample the minimum complexity. Instead of lerping max to min complexity from time
1210         0 to t where t is the ramp length, instead lerp from 0 to (t - intervalSampleLength) so that
1211         we can have at least one interval sample at the min complexity for that ramp.
1212         - Some regression calculations only come out with one line segment rather than the two
1213         we expect. This could be due to a noisy ramp or the ramp's range is too narrow. If that's the
1214         case, influence the minimum complexity of the next ramp towards the lowest bound of 1, so that
1215         we ensure that at least part of the ramp is covering a complexity range that the system can
1216         handle at full 60.
1217         - Remove an assignment to interpolatedFrameLength since that is never subsequently used.
1218
1219         Update the format used to serialize the results for analysis.
1220
1221         Each data point used to have named properties for fields like complexity and frame rate.
1222         In addition the serialized numbers had rounding errors that took up many characters.
1223         Update the format by introducing a new data container called SampleData, which contains a
1224         field map. The map maps a string to an array index. Each data point is an array, so, to
1225         get a stat, use the field map to get the array index into the data point. This allows future
1226         versions to track other data, and reduces the size of the output string by two-thirds.
1227
1228         * Animometer/resources/extensions.js:
1229         (Utilities.toFixedNumber): Add convenience function that truncates the number to a fixed
1230         precision, and converts it back to a number.
1231         (SampleData): New class that contains sample data and a field map that maps properties to
1232         an array index.
1233         (get length): Number of data points.
1234         (addField): Add a field to the field map.
1235         (push): Add a data point.
1236         (sort): Sort the data.
1237         (slice): Return new SampleData object with sliced data.
1238         (forEach): Iterate over the data with provided function.
1239         (createDatum):
1240         (getFieldInDatum): Returns the data point associated with the field name by looking it up
1241         in the field map in the datum provided, which can be the datum object itself (an array) or
1242         an index into the data member variable.
1243         (setFieldInDatum): Sets the data point associated with the field name.
1244         (at): Returns the data point at the provided index.
1245         (toArray): Serializes the data where the field map serves as property names for each point.
1246
1247         * Animometer/resources/debug-runner/graph.js:
1248         (updateGraphData): Remove unused _testData. Convert the data to the old array format for the
1249         graph to use, since the old format was much easier to work with when displaying the graphs.
1250         (onGraphTypeChanged): For some controllers, no alternative score or mean is provided.
1251         * Animometer/resources/runner/animometer.css:
1252         * Animometer/resources/runner/animometer.js: Refactor to use SampleData. Update JSON output
1253         to only go to 3 digits of precision for purposes of reducing the data size.
1254         * Animometer/resources/strings.js: Add new strings to put into the field maps.
1255         * Animometer/tests/resources/main.js: Refactor to use SampleData.
1256
1257         * Animometer/developer.html:
1258         * Animometer/index.html: Restructure results table for both pages. Add charset attribute to
1259         tests.js include.
1260         * Animometer/resources/debug-runner/animometer.css: Clear out styles from release runner.
1261         * Animometer/resources/debug-runner/graph.js:
1262         (onGraphTypeChanged): Update score and mean if bootstrap results are available from the
1263         controller, since not all controllers do bootstrapping.
1264         * Animometer/resources/debug-runner/tests.js: Update header text.
1265         * Animometer/resources/runner/animometer.css: Include confidence interval in results.
1266         * Animometer/resources/runner/animometer.js:
1267         (ResultsTable._addHeader): Header contents can be HTML, so use innerHTML instead.
1268         (ResultsTable._addBody): Add tbody element.
1269         (ResultsTable._addTest): Allow a data cell to invoke a JS function to get its contents.
1270         (window.benchmarkController.showResults): Add table that includes tests' confidence intervals.
1271         * Animometer/resources/runner/tests.js:
1272         (Headers.details.text): Add new details table that includes bootstrap confidence interval.
1273         The interval can be asymmetric, but for simplicity, report the maximum deviation percentage
1274         on either side of the bootstrap median.
1275         * Animometer/resources/statistics.js:
1276         (bootstrap): Include the confidence percentage in the return object.
1277
1278         Report canvas size in results.
1279
1280         * Animometer/developer.html: Add markup to indicate whether a small, medium, or large
1281         canvas was used.
1282         * Animometer/index.html: Ditto.
1283         * Animometer/resources/debug-runner/animometer.js: Call determineCanvasSize().
1284         * Animometer/resources/runner/animometer.css: Update styles to set the canvas based on the
1285         body class size.
1286         * Animometer/resources/runner/animometer.js:
1287         (window.benchmarkController.initialize): Update styles to set the canvas based on the
1288         body class size.
1289         (window.benchmarkController.determineCanvasSize): Run various media queries and set the body
1290         class based on the size of the device.
1291
1292         * Animometer/developer.html: Refactor to include the main CSS file, and redo
1293         the layout so that it doesn't rely on flexbox.
1294         * Animometer/resources/debug-runner/animometer.css:
1295         * Animometer/resources/debug-runner/animometer.js:
1296         (updateDisplay): Since various parts of the script alter the body class, we can't
1297         replace the className directly. Instead, remove all display-based values and then add
1298         the one that was selected.
1299         * Animometer/resources/debug-runner/graph.js:
1300         (updateGraphData): To set the size of the graph, use window.innerHeight.
1301         * Animometer/resources/runner/animometer.js:
1302         (window.sectionsManager.showSection): Since various parts of the script alter the body
1303         class, we can't replace the className directly. Remove all of the section classes
1304         individually and then add the one desired.
1305         * Animometer/tests/resources/stage.css: Remove -apple-system as a font to use in the
1306         stage.
1307
1308 2016-06-12  Filip Pizlo  <fpizlo@apple.com>
1309
1310         Fix round-down goof in Air.js's ShuffleCustom.forEachArg
1311         https://bugs.webkit.org/show_bug.cgi?id=158674
1312
1313         Reviewed by Michael Saboff.
1314         
1315         x / 3 * 3 is not how you round down to multiples of 3 in JavaScript. You need to do
1316         Math.floor(x / 3) * 3 instead.
1317         
1318         This is a benign change, because having extra arguments to a Shuffle is not something we
1319         actually take advantage of yet. But I think it's best to match the original C++ code's
1320         intent.
1321
1322         * Air.js/custom.js:
1323         (const.ShuffleCustom.forEachArg):
1324
1325 2016-06-11  Filip Pizlo  <fpizlo@apple.com>
1326
1327         Unreviewed, skip Air.js. It got unskipped when I renamed it from JSAir.
1328
1329         * Skipped:
1330
1331 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1332
1333         Make it easy to package Air.js
1334         https://bugs.webkit.org/show_bug.cgi?id=158652
1335
1336         Reviewed by Benjamin Poulain.
1337         
1338         If you want to give Air.js to someone, you can now do:
1339         
1340         cd PerformanceTests/Air.js
1341         ./make_dist.sh
1342         
1343         This will create Air.js.tar.gz and a directory called Air.js. If you want to send someone
1344         the benchmark, just send either of those things and tell them to look at index.html.
1345         
1346         You may have to edit make_dist.sh for wherever you have Markdown.pl. You can get that from
1347         https://daringfireball.net/projects/markdown/
1348
1349         * Air.js/make_dist.sh: Added.
1350         * Air.js/test.html:
1351
1352 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1353
1354         Unreviewed, make the link to test.html into an actual link.
1355
1356         * Air.js/README.md:
1357
1358 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1359
1360         Fix some text in Air.js/README.md
1361         https://bugs.webkit.org/show_bug.cgi?id=158650
1362
1363         Reviewed by Benjamin Poulain.
1364         
1365         I read the text again and found bugs:
1366         
1367         - We never actually say how to run the benchmark. This change adds a blurb about how to run
1368           it.
1369
1370         - We both say that allocateStack is responsible for the bulk of the running time and that
1371           we haven't measured where the bulk of the time is spent. This changes the text to say that
1372           it was a goal to make allocateStack be the hottest part of the benchmark, but that we did
1373           not measure this.
1374
1375         * Air.js/README.md:
1376
1377 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1378
1379         Air.js should have some documentation
1380         https://bugs.webkit.org/show_bug.cgi?id=158648
1381
1382         Reviewed by Keith Miller.
1383         
1384         I want to be able to point people at a document if they want to know more about this
1385         benchmark.
1386
1387         * Air.js/README.md: Added.
1388
1389 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1390
1391         Rename JSAir to Air.js.
1392
1393         Rubber stamped by Mark Lam.
1394
1395         * Air.js: Copied from PerformanceTests/JSAir.
1396         * Air.js/airjs-tests.yaml: Copied from PerformanceTests/JSAir/jsair-tests.yaml.
1397         * Air.js/benchmark.js: Copied from PerformanceTests/JSAir/benchmark.js.
1398         (Benchmark):
1399         * Air.js/jsair-tests.yaml: Removed.
1400         * Air.js/payload-airjs-ACLj8C.js: Copied from PerformanceTests/JSAir/payload-jsair-ACLj8C.js.
1401         * Air.js/payload-jsair-ACLj8C.js: Removed.
1402         * Air.js/stress-test.js: Copied from PerformanceTests/JSAir/stress-test.js.
1403         * Air.js/test.html: Copied from PerformanceTests/JSAir/test.html.
1404         * Air.js/test.js: Copied from PerformanceTests/JSAir/test.js.
1405         * JSAir: Removed.
1406         * JSAir/all.js: Removed.
1407         * JSAir/allocate_stack.js: Removed.
1408         * JSAir/arg.js: Removed.
1409         * JSAir/basic_block.js: Removed.
1410         * JSAir/benchmark.js: Removed.
1411         * JSAir/code.js: Removed.
1412         * JSAir/custom.js: Removed.
1413         * JSAir/frequented_block.js: Removed.
1414         * JSAir/insertion_set.js: Removed.
1415         * JSAir/inst.js: Removed.
1416         * JSAir/jsair-tests.yaml: Removed.
1417         * JSAir/liveness.js: Removed.
1418         * JSAir/opcode.js: Removed.
1419         * JSAir/payload-gbemu-executeIteration.js: Removed.
1420         * JSAir/payload-imaging-gaussian-blur-gaussianBlur.js: Removed.
1421         * JSAir/payload-jsair-ACLj8C.js: Removed.
1422         * JSAir/payload-typescript-scanIdentifier.js: Removed.
1423         * JSAir/reg.js: Removed.
1424         * JSAir/stack_slot.js: Removed.
1425         * JSAir/stress-test.js: Removed.
1426         * JSAir/strip-hash.rb: Removed.
1427         * JSAir/symbols.js: Removed.
1428         * JSAir/test.html: Removed.
1429         * JSAir/test.js: Removed.
1430         * JSAir/tmp.js: Removed.
1431         * JSAir/tmp_base.js: Removed.
1432         * JSAir/util.js: Removed.
1433
1434 2016-06-10  Filip Pizlo  <fpizlo@apple.com>
1435
1436         JSC Stress Test failing: jsair-tests.yaml/test.js.ftl-eager-no-cjit
1437         https://bugs.webkit.org/show_bug.cgi?id=158571
1438
1439         Reviewed by Keith Miller.
1440         
1441         Introduce a different harness for run-jsc-stress-tests, which runs for a minimum of 10
1442         iterations and then tries to do any number of "bonus" iterations until it's been running for
1443         two seconds. Since this is the sort of test that isn't really meant to stress anything in
1444         particular, I think it's OK if it is time-limited in this way. The worst case is that some
1445         of its failures will be flaky, but I think that they would have been flaky anyway given the
1446         complexity of the test.
1447
1448         * JSAir/benchmark.js:
1449         (Benchmark):
1450         (Benchmark.prototype.runIteration):
1451         (benchmark): Deleted.
1452         * JSAir/jsair-tests.yaml:
1453         * JSAir/stress-test.js: Added.
1454         (preciseTime):
1455         * JSAir/test.html:
1456         * JSAir/test.js:
1457
1458 2016-06-09  Filip Pizlo  <fpizlo@apple.com>
1459
1460         Unreviewed, teach the perf bots not to run JSAir.
1461
1462         * Skipped:
1463
1464 2016-06-08  Filip Pizlo  <fpizlo@apple.com>
1465
1466         Use more ES6 features in JSAir
1467         https://bugs.webkit.org/show_bug.cgi?id=158497
1468
1469         Reviewed by Keith Miller.
1470         
1471         This improves JSAir with the following ES6 features suggested by JoePeck:
1472         
1473         - String interpolation.
1474         - Destructuring inside PatchCustom.
1475         - Default arguments.
1476         
1477         All of these things are on hot paths.
1478         
1479         Note that I didn't use string interpolation everywhere that I could, only in those places
1480         where it made the code more readable. In Ruby, I used the style that if the interpolation
1481         expression has any non-trivial stuff (like a ternary operator, a chain of calls, or embedded
1482         strings) then it's better to use regular strcat. I think that's what I carried over to here.
1483         
1484         Note that the previous change (Add result validation to JSAir) also made the Proxy code not
1485         dead, though it's not necessarily on the hot path. The Proxy isn't called into frequently
1486         but it's used from a function that is otherwise hot, so if calling into the Proxy prevents
1487         that function from being optimized then it will hurt so good.
1488         
1489         I also reenabled tail calls in a few places.
1490         
1491         This change doesn't seem to change the performance of the benchmark for us. That's expected
1492         since these ES6 features are cheap. Note that this claim doesn't include Proxy, which was
1493         added in a separate change and that change did make the benchmark overall more expensive.
1494
1495         * JSAir/allocate_stack.js:
1496         (allocateStack):
1497         * JSAir/arg.js:
1498         (Arg.createBitImm64):
1499         (Arg.createAddr):
1500         (Arg.createStack):
1501         (Arg.logScale):
1502         (Arg.createIndex):
1503         * JSAir/basic_block.js:
1504         (BasicBlock.get headerString):
1505         (BasicBlock.prototype.get if):
1506         (BasicBlock):
1507         * JSAir/benchmark.js:
1508         (benchmark):
1509         * JSAir/code.js:
1510         (Code):
1511         (Code.prototype.addBlock):
1512         (Code.prototype.addStackSlot):
1513         (Code.prototype.newTmp):
1514         (Code.prototype.get size):
1515         (Code.prototype.get blocks):
1516         (Code.prototype.get stackSlots):
1517         (Code.prototype.tmps):
1518         (Code.prototype.get callArgAreaSize):
1519         (Code.prototype.toString):
1520         * JSAir/custom.js:
1521         (const.PatchCustom.forEachArg):
1522         * JSAir/inst.js:
1523         (Inst):
1524         * JSAir/reg.js:
1525         (Reg.prototype.toString):
1526         * JSAir/util.js:
1527         (symbolName):
1528         (lowerSymbolName):
1529
1530 2016-06-07  Filip Pizlo  <fpizlo@apple.com>
1531
1532         Add result validation to JSAir
1533         https://bugs.webkit.org/show_bug.cgi?id=158493
1534
1535         Reviewed by Saam Barati.
1536         
1537         This adds the ability to hash a Code in a way that matches the C++ code's hashing of Code.
1538         This allows us to check if the Code that JSAir sees is the code that C++ saw. We use this to
1539         check the Code before and after allocateStack, and compare against hashes we got from C++.
1540         
1541         Doing this uncovered bugs. roundUpToMultipleOf wasn't doing anything. allocateStack was not
1542         allocating things correctly because I was concatting a Set to an Array, which doesn't really
1543         work. Now these bugs are fixed.
1544         
1545         The checking step adds to the running time so I reduced the number of iterations. The
1546         benchmark spends a decent amount of its time computing Code hashes; I think it's around 1/3
1547         total. This is probably OK. It's better to verify the results even if the running time is
1548         not all in the "core" of the algorithm.
1549         
1550         Also add a run-jsc-stress-tests yaml file to allow this to run as a test.
1551
1552         * JSAir/allocate_stack.js:
1553         * JSAir/arg.js:
1554         (Arg.createImm):
1555         (Arg.createBigImm):
1556         (Arg.createBitImm):
1557         (Arg.createBitImm64):
1558         (Arg.createWidth):
1559         (Arg.createSpecial):
1560         (Arg.prototype.get kind):
1561         (Arg.prototype.get isTmp):
1562         (Arg.prototype.get isImm):
1563         (Arg.prototype.get isSomeImm):
1564         (Arg.prototype.get isSomeBigImm):
1565         (Arg.prototype.get isCondition):
1566         (Arg.prototype.get isWidth):
1567         (Arg.prototype.get isSpecial):
1568         (Arg.prototype.get isAlive):
1569         (Arg.prototype.get tmp):
1570         (Arg.prototype.get value):
1571         (Arg.prototype.get lowValue):
1572         (Arg.prototype.get highValue):
1573         (Arg.prototype.get base):
1574         (Arg.prototype.get isGP):
1575         (Arg.prototype.get isFP):
1576         (Arg.prototype.isValidForm):
1577         (Arg.prototype.get isInvertible):
1578         (Arg.kindCode):
1579         (Arg.prototype.hash):
1580         (Arg.prototype.toString):
1581         (Arg):
1582         * JSAir/basic_block.js:
1583         (BasicBlock.get successorBlocks):
1584         * JSAir/benchmark.js:
1585         (benchmark):
1586         * JSAir/code.js:
1587         (Code.prototype.setFrameSize):
1588         (Code.prototype.hash):
1589         (Code.prototype.toString):
1590         (Code):
1591         * JSAir/inst.js:
1592         (Inst.prototype.get hasNonArgEffects):
1593         (Inst.prototype.hash):
1594         (Inst.prototype.toString):
1595         (Inst):
1596         * JSAir/jsair-tests.yaml: Added.
1597         * JSAir/opcode.js:
1598         (Inst_forEachArg):
1599         (Inst_hasNonArgEffects):
1600         (opcodeCode):
1601         * JSAir/payload-gbemu-executeIteration.js:
1602         (createPayloadGbemuExecuteIteration):
1603         * JSAir/payload-imaging-gaussian-blur-gaussianBlur.js:
1604         (createPayloadImagingGaussianBlurGaussianBlur):
1605         * JSAir/payload-jsair-ACLj8C.js:
1606         (createPayloadJSAirACLj8C):
1607         * JSAir/payload-typescript-scanIdentifier.js:
1608         (createPayloadTypescriptScanIdentifier):
1609         * JSAir/reg.js:
1610         (Reg.prototype.get isReg):
1611         (Reg.prototype.hash):
1612         (Reg.prototype.toString):
1613         * JSAir/stack_slot.js:
1614         (StackSlot.prototype.setOffsetFromFP):
1615         (StackSlot.prototype.hash):
1616         (StackSlot.prototype.toString):
1617         * JSAir/symbols.js:
1618         (relCondCode):
1619         (resCondCode):
1620         (doubleCondCode):
1621         * JSAir/test.html:
1622         * JSAir/tmp.js:
1623         (Tmp.prototype.get isReg):
1624         (Tmp.prototype.hash):
1625         (Tmp.prototype.toString):
1626         * JSAir/util.js:
1627         (roundUpToMultipleOf):
1628         (symbolName):
1629
1630 2016-06-07  Filip Pizlo  <fpizlo@apple.com>
1631
1632         Implement Air::allocateStack() in ES6 to see how much of a bad idea that is
1633         https://bugs.webkit.org/show_bug.cgi?id=158318
1634
1635         Reviewed by Saam Barati.
1636         
1637         This adds a new benchmark for us to play with called JSAir. It's a complete ES6
1638         implementation of Air's allocateStack() phase along with all of Air needed to run that
1639         phase. This includes things like stack slots, registers, temporaries, basic blocks,
1640         instructions, and all of the code for iterating over, inspecting, and modifying those
1641         things.
1642         
1643         To make this work, JSC can now dump Air just before allocateStack() in the form of JS code
1644         that creates a Code object that matches exactly what C++ Air saw. This benchmark comprises
1645         four Air IRs:
1646         
1647         - Octane/gbemu's largest function, executeIteration.
1648         - Kraken/imaging-gaussian-blur's largest function in OSR entry mode, gaussuanBlur.
1649         - Octane/typescript's largest function that is actually hot, scanIdentifier.
1650         - JSAir's largest hot function, which is anonymous, so we call it by its hash (ACLj8C).
1651         
1652         This runs in about 2 seconds on my machine in JSC trunk. It includes both a commandline
1653         harness and a web harness. JSAir is almost exactly 100x slower in ES6 in WebKit than the
1654         C++ Air::allocateStack() phase on which it is based.
1655         
1656         JSAir uses the following ES6 features:
1657         
1658         - Symbol.
1659         - for-of.
1660         - arrow functions.
1661         - Map/Set.
1662         - let/const.
1663         - classes.
1664         
1665         All of these things are used in anger and should end up on the hot path. There is also code
1666         that uses Proxies, but it ends up being dead.
1667         
1668         We can improve this even more:
1669         I still need to add result validation: https://bugs.webkit.org/show_bug.cgi?id=158493
1670         I want to make it use more ES6 features: https://bugs.webkit.org/show_bug.cgi?id=158497
1671         
1672         * JSAir: Added.
1673         * JSAir/all.js: Added.
1674         * JSAir/allocate_stack.js: Added.
1675         (allocateStack.attemptAssignment):
1676         (allocateStack.assign):
1677         (allocateStack.interfere):
1678         (allocateStack.):
1679         (allocateStack):
1680         * JSAir/arg.js: Added.
1681         (Arg):
1682         (Arg.isAnyUse):
1683         (Arg.isColdUse):
1684         (Arg.isWarmUse):
1685         (Arg.cooled):
1686         (Arg.isEarlyUse):
1687         (Arg.isLateUse):
1688         (Arg.isAnyDef):
1689         (Arg.isEarlyDef):
1690         (Arg.isLateDef):
1691         (Arg.isZDef):
1692         (Arg.typeForB3Type):
1693         (Arg.widthForB3Type):
1694         (Arg.conservativeWidth):
1695         (Arg.minimumWidth):
1696         (Arg.bytes):
1697         (Arg.widthForBytes):
1698         (Arg.createTmp):
1699         (Arg.fromReg):
1700         (Arg.createImm):
1701         (Arg.createBigImm):
1702         (Arg.createBitImm):
1703         (Arg.createBitImm64):
1704         (Arg.createAddr):
1705         (Arg.createStack):
1706         (Arg.createCallArg):
1707         (Arg.createStackAddr):
1708         (Arg.isValidScale):
1709         (Arg.logScale):
1710         (Arg.createIndex):
1711         (Arg.createRelCond):
1712         (Arg.createResCond):
1713         (Arg.createDoubleCond):
1714         (Arg.createWidth):
1715         (Arg.prototype.get kind):
1716         (Arg.prototype.get isTmp):
1717         (Arg.prototype.get isImm):
1718         (Arg.prototype.get isBigImm):
1719         (Arg.prototype.get isBitImm):
1720         (Arg.prototype.get isBitImm64):
1721         (Arg.prototype.get isSomeImm):
1722         (Arg.prototype.get isAddr):
1723         (Arg.prototype.get isStack):
1724         (Arg.prototype.get isCallArg):
1725         (Arg.prototype.get isIndex):
1726         (Arg.prototype.get isMemory):
1727         (Arg.prototype.get isStackMemory):
1728         (Arg.prototype.get isRelCond):
1729         (Arg.prototype.get isResCond):
1730         (Arg.prototype.get isDoubleCond):
1731         (Arg.prototype.get isCondition):
1732         (Arg.prototype.get isWidth):
1733         (Arg.prototype.get isAlive):
1734         (Arg.prototype.get tmp):
1735         (Arg.prototype.get value):
1736         (Arg.prototype.get base):
1737         (Arg.prototype.get hasOffset):
1738         (Arg.prototype.get offset):
1739         (Arg.prototype.get stackSlot):
1740         (Arg.prototype.get index):
1741         (Arg.prototype.get scale):
1742         (Arg.prototype.get logScale):
1743         (Arg.prototype.get width):
1744         (Arg.prototype.get isGPTmp):
1745         (Arg.prototype.get isFPTmp):
1746         (Arg.prototype.get isGP):
1747         (Arg.prototype.get isFP):
1748         (Arg.prototype.get hasType):
1749         (Arg.prototype.get type):
1750         (Arg.prototype.isType):
1751         (Arg.prototype.isCompatibleType):
1752         (Arg.prototype.get isGPR):
1753         (Arg.prototype.get gpr):
1754         (Arg.prototype.get isFPR):
1755         (Arg.prototype.get fpr):
1756         (Arg.prototype.get isReg):
1757         (Arg.prototype.get reg):
1758         (Arg.isValidImmForm):
1759         (Arg.isValidBitImmForm):
1760         (Arg.isValidBitImm64Form):
1761         (Arg.isValidAddrForm):
1762         (Arg.isValidIndexForm):
1763         (Arg.prototype.isValidForm):
1764         (Arg.prototype.forEachTmpFast):
1765         (Arg.prototype.usesTmp):
1766         (Arg.prototype.forEachTmp):
1767         (Arg.prototype.is):
1768         (Arg.prototype.as):
1769         (Arg.prototype.forEachFast):
1770         (Arg.prototype.forEach):
1771         (Arg.extract):
1772         (Arg.forEachFast):
1773         (Arg.forEach):
1774         (Arg.prototype.get condition):
1775         (Arg.prototype.get isInvertible):
1776         (Arg.prototype.toString):
1777         * JSAir/basic_block.js: Added.
1778         (BasicBlock):
1779         (BasicBlock.prototype.get index):
1780         (BasicBlock.prototype.get size):
1781         (BasicBlock.prototype.Symbol.iterator):
1782         (BasicBlock.prototype.at):
1783         (BasicBlock.get last):
1784         (BasicBlock.get insts):
1785         (BasicBlock.get numSuccessors):
1786         (BasicBlock.get successors):
1787         (BasicBlock.get successorBlocks.):
1788         (BasicBlock.get successorBlocks):
1789         (BasicBlock.set get numPredecessors):
1790         (BasicBlock.get predecessors):
1791         (BasicBlock.get frequency):
1792         (BasicBlock.get headerString):
1793         (BasicBlock.get footerString):
1794         * JSAir/benchmark.js: Added.
1795         (benchmark):
1796         * JSAir/code.js: Added.
1797         (Code):
1798         (Code.prototype.addBlock):
1799         (Code.prototype.addStackSlot):
1800         (Code.prototype.newTmp):
1801         (Code.prototype.get size):
1802         (Code.prototype.at):
1803         (Code.prototype.Symbol.iterator):
1804         (Code.prototype.get blocks):
1805         (Code.prototype.get stackSlots):
1806         (Code.prototype.tmps):
1807         (Code.prototype.get callArgAreaSize):
1808         (Code.prototype.requestCallArgAreaSize):
1809         (Code.prototype.get frameSize):
1810         (Code.prototype.setFrameSize):
1811         (Code.prototype.toString):
1812         * JSAir/custom.js: Added.
1813         (const.ShuffleCustom.forEachArg):
1814         (const.ShuffleCustom.hasNonArgNonControlEffects):
1815         (const.PatchCustom.forEachArg):
1816         (const.PatchCustom.hasNonArgNonControlEffects):
1817         (const.CCallCustom.forEachArg):
1818         (const.CCallCustom.hasNonArgNonControlEffects):
1819         (const.ColdCCallCustom.forEachArg):
1820         (const.ColdCCallCustom.hasNonArgNonControlEffects):
1821         * JSAir/frequented_block.js: Added.
1822         (FrequentedBlock):
1823         (FrequentedBlock.prototype.toString):
1824         * JSAir/insertion_set.js: Added.
1825         (Insertion):
1826         (Insertion.prototype.get index):
1827         (Insertion.prototype.get element):
1828         (Insertion.prototype.lessThan):
1829         (InsertionSet):
1830         (InsertionSet.prototype.appendInsertion):
1831         (InsertionSet.prototype.append):
1832         (InsertionSet.prototype.execute):
1833         * JSAir/inst.js: Added.
1834         (Inst):
1835         (Inst.prototype.append):
1836         (Inst.prototype.clear):
1837         (Inst.prototype.get opcode):
1838         (Inst.prototype.get args):
1839         (Inst.prototype.visitArg):
1840         (Inst.prototype.forEachTmpFast):
1841         (Inst.prototype.forEachArg):
1842         (Inst.prototype.forEachTmp):
1843         (Inst.prototype.forEach):
1844         (Inst.forEachDef):
1845         (Inst.forEachDefWithExtraClobberedRegs):
1846         (Inst.prototype.get hasNonArgEffects):
1847         (Inst.prototype.toString):
1848         * JSAir/liveness.js: Added.
1849         (Liveness):
1850         (Liveness.prototype.get thing):
1851         (Liveness.prototype.get code):
1852         (Liveness.prototype.get liveAtHead):
1853         (Liveness.prototype.get liveAtTail):
1854         (Liveness.prototype.localCalc.LocalCalc):
1855         (Liveness.prototype.localCalc.LocalCalc.prototype.get liveSet):
1856         (Liveness.prototype.localCalc.LocalCalc.prototype.execute):
1857         (Liveness.prototype.localCalc):
1858         * JSAir/opcode.js: Added.
1859         (Inst_forEachArg):
1860         (Inst_hasNonArgEffects):
1861         * JSAir/payload-gbemu-executeIteration.js: Added.
1862         (createPayloadGbemuExecuteIteration):
1863         * JSAir/payload-imaging-gaussian-blur-gaussianBlur.js: Added.
1864         (createPayloadImagingGaussianBlurGaussianBlur):
1865         * JSAir/payload-jsair-ACLj8C.js: Added.
1866         (createPayloadJSAirACLj8C):
1867         * JSAir/payload-typescript-scanIdentifier.js: Added.
1868         (createPayloadTypescriptScanIdentifier):
1869         * JSAir/reg.js: Added.
1870         (Reg):
1871         (Reg.fromReg):
1872         (Reg.prototype.get index):
1873         (Reg.prototype.get type):
1874         (Reg.prototype.get name):
1875         (Reg.prototype.get isCalleeSave):
1876         (Reg.prototype.get isReg):
1877         (Reg.prototype.toString):
1878         (Reg.extract):
1879         (Reg.forEachFast):
1880         (Reg.forEach):
1881         (newGPR):
1882         (Reg.gprs.Reg.fprs.Reg.calleeSaveGPRs.Reg.calleeSaveFPRs.Reg.calleeSaves):
1883         * JSAir/stack_slot.js: Added.
1884         (StackSlot):
1885         (StackSlot.prototype.get byteSize):
1886         (StackSlot.prototype.get kind):
1887         (StackSlot.prototype.get isLocked):
1888         (StackSlot.prototype.get isSpill):
1889         (StackSlot.prototype.get index):
1890         (StackSlot.prototype.ensureSize):
1891         (StackSlot.prototype.get alignment):
1892         (StackSlot.prototype.get offsetFromFP):
1893         (StackSlot.prototype.setOffsetFromFP):
1894         (StackSlot.prototype.toString):
1895         (StackSlot.extract):
1896         (StackSlot.forEachFast):
1897         (StackSlot.forEach):
1898         * JSAir/symbols.js: Added.
1899         * JSAir/test.html: Added.
1900         * JSAir/test.js: Added.
1901         * JSAir/tmp.js: Added.
1902         (Tmp):
1903         (Tmp.fromReg):
1904         (Tmp.prototype.get index):
1905         (Tmp.prototype.get type):
1906         (Tmp.prototype.get isReg):
1907         (Tmp.prototype.toString):
1908         (Tmp.extract):
1909         (Tmp.forEachFast):
1910         (Tmp.forEach):
1911         * JSAir/tmp_base.js: Added.
1912         (TmpBase.prototype.get isGP):
1913         (TmpBase.prototype.get isFP):
1914         (TmpBase.prototype.get isGPR):
1915         (TmpBase.prototype.get isFPR):
1916         (TmpBase.prototype.get reg):
1917         (TmpBase.prototype.get gpr):
1918         (TmpBase.prototype.get fpr):
1919         (TmpBase):
1920         * JSAir/util.js: Added.
1921         (isRepresentableAsInt32):
1922         (addIndexed):
1923         (roundUpToMultipleOf):
1924         (symbolName):
1925         (mergeIntoSet):
1926         (nonEmptyRangesOverlap):
1927         (rangesOverlap):
1928         (removeAllMatching):
1929         (swap):
1930         (bubble):
1931         (bubbleSort):
1932         (currentTime):
1933         (else.currentTime):
1934
1935 2016-05-26  Yusuke Suzuki  <utatane.tea@gmail.com>
1936
1937         [JSC] Allow JSBench to use precise time
1938         https://bugs.webkit.org/show_bug.cgi?id=158050
1939
1940         Reviewed by Geoffrey Garen.
1941
1942         * JSBench/amazon-chrome-win/urem.js:
1943         (else.window.performance.window.performance.now.currentTimeInMS):
1944         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1945         (else.else.currentTimeInMS):
1946         (onload.cb):
1947         (onload):
1948         * JSBench/amazon-chrome/urem.js:
1949         (else.window.performance.window.performance.now.currentTimeInMS):
1950         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1951         (else.else.currentTimeInMS):
1952         (onload.cb):
1953         (onload):
1954         * JSBench/amazon-firefox-win/urm.js:
1955         (else.window.performance.window.performance.now.currentTimeInMS):
1956         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1957         (else.else.currentTimeInMS):
1958         (onload.cb):
1959         (onload):
1960         * JSBench/amazon-firefox/urm.js:
1961         (else.window.performance.window.performance.now.currentTimeInMS):
1962         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
1963         (else.else.currentTimeInMS):
1964         (onload.cb):
1965         (onload):
1966         * JSBench/amazon-safari/urem.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/facebook-chrome-win/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/facebook-chrome/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/facebook-firefox-win/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/facebook-firefox/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/facebook-safari/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/google-chrome-win/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/google-chrome/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/google-firefox-win/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/google-firefox/uem.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         * JSBench/google-safari/urem.js:
2027         (else.window.performance.window.performance.now.currentTimeInMS):
2028         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2029         (else.else.currentTimeInMS):
2030         (onload.cb):
2031         (onload):
2032         * JSBench/harness.js:
2033         (runBenchmark.window.currentTimeInMS):
2034         (runBenchmark.else.window.currentTimeInMS):
2035         * JSBench/twitter-chrome-win/rem.js:
2036         (else.window.performance.window.performance.now.currentTimeInMS):
2037         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2038         (else.else.currentTimeInMS):
2039         (onload.cb):
2040         (onload):
2041         * JSBench/twitter-chrome/urem.js:
2042         (else.window.performance.window.performance.now.currentTimeInMS):
2043         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2044         (else.else.currentTimeInMS):
2045         (onload.cb):
2046         (onload):
2047         * JSBench/twitter-firefox-win/urem.js:
2048         (else.window.performance.window.performance.now.currentTimeInMS):
2049         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2050         (else.else.currentTimeInMS):
2051         (onload.cb):
2052         (onload):
2053         * JSBench/twitter-firefox/urem.js:
2054         (else.window.performance.window.performance.now.currentTimeInMS):
2055         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2056         (else.else.currentTimeInMS):
2057         (onload.cb):
2058         (onload):
2059         * JSBench/twitter-safari/urem.js:
2060         (else.window.performance.window.performance.now.currentTimeInMS):
2061         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2062         (else.else.currentTimeInMS):
2063         (onload.cb):
2064         (onload):
2065         * JSBench/yahoo-chrome-win/urem.js:
2066         (else.window.performance.window.performance.now.currentTimeInMS):
2067         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2068         (else.else.currentTimeInMS):
2069         (onload.cb):
2070         (onload):
2071         * JSBench/yahoo-chrome/urem.js:
2072         (else.window.performance.window.performance.now.currentTimeInMS):
2073         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2074         (else.else.currentTimeInMS):
2075         (onload.cb):
2076         (onload):
2077         * JSBench/yahoo-firefox-win/urem.js:
2078         (else.window.performance.window.performance.now.currentTimeInMS):
2079         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2080         (else.else.currentTimeInMS):
2081         (onload.cb):
2082         (onload):
2083         * JSBench/yahoo-firefox/urem.js:
2084         (else.window.performance.window.performance.now.currentTimeInMS):
2085         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2086         (else.else.currentTimeInMS):
2087         (onload.cb):
2088         (onload):
2089         * JSBench/yahoo-safari/urem.js:
2090         (else.window.performance.window.performance.now.currentTimeInMS):
2091         (else.else.typeof.preciseTime.string_appeared_here.currentTimeInMS):
2092         (else.else.currentTimeInMS):
2093         (onload.cb):
2094         (onload):
2095
2096 2016-05-25  Geoffrey Garen  <ggaren@apple.com>
2097
2098         replaceable own properties seem to ignore replacement after property caching
2099         https://bugs.webkit.org/show_bug.cgi?id=158091
2100
2101         Reviewed by Darin Adler.
2102
2103         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
2104         * MallocBench/MallocBench/Benchmark.cpp:
2105         * MallocBench/MallocBench/Interpreter.cpp:
2106         (Interpreter::doMallocOp):
2107         * MallocBench/MallocBench/Interpreter.h:
2108         * MallocBench/MallocBench/fastMallocLog.63316.ops: Added.
2109         * MallocBench/MallocBench/jetstream.cpp: Added.
2110         (benchmark_jetstream):
2111         * MallocBench/MallocBench/jetstream.h: Added.
2112
2113 2016-05-25  Keith Miller  <keith_miller@apple.com>
2114
2115         Unreviewed, add JSBench to the skipped list for now since it doesn't
2116         work currently.
2117
2118         * Skipped:
2119
2120 2016-05-17  Keith Miller  <keith_miller@apple.com>
2121         We should have JSBench in PerformanceTests
2122         https://bugs.webkit.org/show_bug.cgi?id=157952
2123
2124         Rubber-stamped by Saam Barati.
2125
2126         There are some slight changes to the layout of the test directory
2127         to make it work nicely with run-jsc-benchmarks. Before JSBench
2128         had each of the browser specific sub-tests in a sub-directory.
2129         These have been flattened e.g. amazon/safari/ has become
2130         amazon-safari/.
2131
2132         * JSBench/amazon-chrome-win/urem.html: Added.
2133         * JSBench/amazon-chrome-win/urem.js: Added.
2134         * JSBench/amazon-chrome/urem.html: Added.
2135         * JSBench/amazon-chrome/urem.js: Added.
2136         * JSBench/amazon-firefox-win/urm.html: Added.
2137         * JSBench/amazon-firefox-win/urm.js: Added.
2138         * JSBench/amazon-firefox/urm.html: Added.
2139         * JSBench/amazon-firefox/urm.js: Added.
2140         * JSBench/amazon-safari/urem.html: Added.
2141         * JSBench/amazon-safari/urem.js: Added.
2142         * JSBench/browsercheck.js: Added.
2143         * JSBench/facebook-chrome-win/urem.html: Added.
2144         * JSBench/facebook-chrome-win/urem.js: Added.
2145         * JSBench/facebook-chrome/urem.html: Added.
2146         * JSBench/facebook-chrome/urem.js: Added.
2147         * JSBench/facebook-firefox-win/urem.html: Added.
2148         * JSBench/facebook-firefox-win/urem.js: Added.
2149         * JSBench/facebook-firefox/urem.html: Added.
2150         * JSBench/facebook-firefox/urem.js: Added.
2151         * JSBench/facebook-safari/urem.html: Added.
2152         * JSBench/facebook-safari/urem.js: Added.
2153         * JSBench/google-chrome-win/urem.html: Added.
2154         * JSBench/google-chrome-win/urem.js: Added.
2155         * JSBench/google-chrome/urem.html: Added.
2156         * JSBench/google-chrome/urem.js: Added.
2157         * JSBench/google-firefox-win/urem.html: Added.
2158         * JSBench/google-firefox-win/urem.js: Added.
2159         * JSBench/google-firefox/uem.html: Added.
2160         * JSBench/google-firefox/uem.js: Added.
2161         * JSBench/google-safari/urem.html: Added.
2162         * JSBench/google-safari/urem.js: Added.
2163         * JSBench/harness.html: Added.
2164         * JSBench/harness.js: Added.
2165         * JSBench/harness.py: Added.
2166         * JSBench/index.html: Added.
2167         * JSBench/reload.html: Added.
2168         * JSBench/twitter-chrome-win/rem.html: Added.
2169         * JSBench/twitter-chrome-win/rem.js: Added.
2170         * JSBench/twitter-chrome/urem.html: Added.
2171         * JSBench/twitter-chrome/urem.js: Added.
2172         * JSBench/twitter-firefox-win/urem.html: Added.
2173         * JSBench/twitter-firefox-win/urem.js: Added.
2174         * JSBench/twitter-firefox/urem.html: Added.
2175         * JSBench/twitter-firefox/urem.js: Added.
2176         * JSBench/twitter-safari/urem.html: Added.
2177         * JSBench/twitter-safari/urem.js: Added.
2178         * JSBench/yahoo-chrome-win/urem.html: Added.
2179         * JSBench/yahoo-chrome-win/urem.js: Added.
2180         * JSBench/yahoo-chrome/urem.html: Added.
2181         * JSBench/yahoo-chrome/urem.js: Added.
2182         * JSBench/yahoo-firefox-win/urem.html: Added.
2183         * JSBench/yahoo-firefox-win/urem.js: Added.
2184         * JSBench/yahoo-firefox/urem.html: Added.
2185         * JSBench/yahoo-firefox/urem.js: Added.
2186         * JSBench/yahoo-safari/urem.html: Added.
2187         * JSBench/yahoo-safari/urem.js: Added.
2188
2189 2016-05-20  Ryosuke Niwa  <rniwa@webkit.org>
2190
2191         run-benchmark's results should contain Animometer's debug output
2192         https://bugs.webkit.org/show_bug.cgi?id=157941
2193
2194         Reviewed by Stephanie Lewis.
2195
2196         Made developer.html support the JSON generated by run-benchmark which stores everything under debugOutput.
2197
2198         * Animometer/resources/debug-runner/animometer.js:
2199         (Utilities.initialize): Unwrap debugOutput in the case run-benchmark's result JSON is used.
2200
2201 2016-05-18  Timothy Hatcher  <timothy@apple.com>
2202
2203         Make Animometer work in all browsers
2204         https://bugs.webkit.org/show_bug.cgi?id=157855
2205         rdar://problem/26338521
2206
2207         Reviewed by Jon Lee.
2208
2209         * Animometer/resources/runner/logo.svg: Use a symbol instead
2210         of putting the id on the root svg. This works in all browsers.
2211
2212 2016-05-07  Timothy Hatcher  <timothy@apple.com>
2213
2214         Update Animometer design
2215         https://bugs.webkit.org/show_bug.cgi?id=157449
2216
2217         Reviewed by Darin Adler.
2218
2219         * Animometer/developer.html: Stop including runner/animometer.css.
2220         * Animometer/index.html: Udpated elements to fit new styles.
2221
2222         * Animometer/resources/debug-runner/animometer.css:
2223         Copy the old runner/animometer.css to here so the debug runner is unaffected.
2224
2225         * Animometer/resources/runner/animometer.css:
2226         New styles for the design. Drop flex box for broad browser support.
2227
2228         * Animometer/resources/runner/animometer.js:
2229         (ResultsTable.clear): Use textContent.
2230         (window.sectionsManager.showSection): Add class to the body.
2231         (window.sectionsManager.setSectionScore): Use textContent, not innerHTML.
2232
2233         * Animometer/resources/runner/crystal.svg: Added.
2234         * Animometer/resources/runner/lines.svg: Added.
2235         * Animometer/resources/runner/logo.svg: Added.
2236
2237         * Animometer/resources/runner/tests.js: Shorten test name and title case them.
2238
2239 2016-05-06  Manuel Rego Casasnovas  <rego@igalia.com>
2240
2241         [css-grid] Unprefix CSS Grid Layout properties
2242         https://bugs.webkit.org/show_bug.cgi?id=157137
2243
2244         Reviewed by Simon Fraser.
2245
2246         Remove "-webkit" prefix from all the grid layout properties,
2247         including the display value.
2248         Update the source code to remove the prefix where it was used too.
2249
2250         * Layout/auto-grid-lots-of-data.html:
2251         * Layout/fixed-grid-lots-of-data.html:
2252         * Layout/fixed-grid-lots-of-stretched-data.html:
2253
2254 2016-04-25  Geoffrey Garen  <ggaren@apple.com>
2255
2256         bmalloc: Misc improvements to MallocBench
2257         https://bugs.webkit.org/show_bug.cgi?id=157004
2258
2259         Reviewed by Darin Adler.
2260
2261         * MallocBench/run-malloc-benchmarks: Added --memory and --memory_warning
2262         modes for focused memory testing.
2263
2264         * MallocBench/MallocBench/Benchmark.cpp:
2265         (Benchmark::printReport): Clarified output.
2266
2267         (Benchmark::currentMemoryBytes): Added compressed memory because top
2268         does the same. (It always happens to zero in the benchmarks we run. But
2269         this is good for sanity.)
2270
2271         * MallocBench/MallocBench/CommandLine.cpp: Moved up to 8 runs to reduce
2272         variance.
2273
2274         * MallocBench/MallocBench/alloc_free.cpp:
2275         (benchmark_alloc_free): Cycle a single allocation in order to stress
2276         the effect of merging on calls to madvise.
2277
2278         * MallocBench/MallocBench/big.cpp:
2279         (benchmark_big): Graduated to 8kB-128kB because medium tests up to 8 and
2280         our large allocator doesn't kick in until 64kB.
2281
2282         * MallocBench/MallocBench/medium.cpp:
2283         (benchmark_medium): Test all the way down to 1kB because our large
2284         allocator used to service 1kB allocations and 1kB is an interesting
2285         middle size where memory is unusually large but allocation throughput
2286         still matters.
2287
2288         * MallocBench/MallocBench/stress.cpp:
2289         (benchmark_stress): Reduced the churn count to match stress_aligned
2290         because this test was taking too long to complete.
2291
2292         * MallocBench/MallocBench/stress_aligned.cpp:
2293         (benchmark_stress_aligned): Our new large allocator can handle even
2294         more absurdly large values.
2295
2296 2016-04-25  Simon Fraser  <simon.fraser@apple.com>
2297
2298         Add a content animation test that uses SVG animation.
2299         https://bugs.webkit.org/show_bug.cgi?id=156827
2300
2301         * Animation/svg-animation.html: Added.
2302         * Skipped:
2303
2304 2016-04-22  Simon Fraser  <simon.fraser@apple.com>
2305
2306         Skip two content animation tests which are only meant for iOS testing.
2307
2308         * Animation/css-animation.html: Added.
2309         * Animation/raf-animation.html: Added.
2310
2311         * Skipped:
2312
2313 2016-04-20  Simon Fraser  <simon.fraser@apple.com>
2314
2315         Add content animation tests to benchmark_runner, and allow the runner to collect device data as part of the results
2316         https://bugs.webkit.org/show_bug.cgi?id=156827
2317
2318         Add two files missing from the previous commit.
2319
2320         * Animation/css-animation.html: Added.
2321         * Animation/raf-animation.html: Added.
2322
2323 2016-04-20  Simon Fraser  <simon.fraser@apple.com>
2324
2325         Add content animation tests to benchmark_runner, and allow the runner to collect device data as part of the results
2326         https://bugs.webkit.org/show_bug.cgi?id=156827
2327
2328         Reviewed by Stephanie Lewis.
2329
2330         Two new files for the content-animation suite.
2331
2332         After patching for use in the benchmark, these tests are a little different to other benchmarks.
2333         The test content itself does not collect performance data; the tests spit out a boilerplate JSON
2334         result with a placeholder for framerate, which is collected in native code.
2335
2336         When run as a benchmark, test completion involves a couple of bounces between the test and MobileSafari.
2337         Test completion is initiated by a setTimeout() in the test, at which point location.hash is set to "#done".
2338         MobileSafari detects that, and asynchronously dumps performance data. Once this is complete, MobileSafari
2339         changes the URL has to "#submit", which triggers the XHR to the test relay that indicates the test is
2340         complete.
2341
2342         * Animation/css-animation.html: Added.
2343         * Animation/raf-animation.html: Added.
2344
2345 2016-04-15  Jon Lee  <jonlee@apple.com>
2346
2347         Animometer test could report a NaN
2348         https://bugs.webkit.org/show_bug.cgi?id=156646
2349
2350         Reviewed by Darin Adler.
2351         Provisionally reviewed by Said Abou-Hallawa.
2352
2353         * Animometer/tests/resources/main.js:
2354         (didFinishInterval): The ramp controller has a first phase where it ramps up the complexity and
2355         reacts based on how well the system handles the load. The assumption was that it would handle at least
2356         1 particle easily. That is not always the case. As a result, an interpolation calculation could
2357         end up setting an upper bound of NaN.
2358
2359         This occurs because we never get out of the first tier, so this._lastTierComplexity is undefined.
2360         Now that we guarantee a minimum complexity of 1, modify the conditional to check for this._lastTierComplexity
2361         before interpolating the upper bound of the first ramp. In the case where the system struggles with
2362         1 particle, set it to currentComplexity.
2363
2364 2016-04-09  Jon Lee  <jonlee@apple.com>
2365
2366         Update Animometer to accommodate different screens
2367         https://bugs.webkit.org/show_bug.cgi?id=156449
2368
2369         Reviewed by Darin Adler.
2370         Provisionally reviewed by Said Abou-Hallawa.
2371
2372         * Animometer/index.html: Wrap button in a container to add padding at the bottom.
2373         * Animometer/resources/debug-runner/animometer.css:
2374         (@media screen and (min-device-width: 1800px)): Deleted.
2375         (@media screen and (min-width: 1800px)): Cannot use min-device-width since it may match incorrectly.
2376         (screen and (max-device-height: 414px) and (orientation: landscape)): Some devices swap device width
2377         and height with orientation change.
2378         * Animometer/resources/runner/animometer.css: Similar.
2379         (screen and (min-device-width: 1024px) and (orientation: landscape)):
2380         (screen and (max-device-height: 414px) and (orientation: landscape)):
2381         (.frame-container): On smaller iPhones, adding 1px prevents the navigation bars from appearing.
2382         (@media screen and (min-device-width: 768px) and (max-device-width: 1024px)): Deleted.
2383         (@media (min-device-height: 768px) and (max-device-height: 1024px)): Target iPad Airs and similar.
2384         (@media screen and (min-device-width: 1024px) and (max-device-width: 1366px)): Deleted.
2385         (@media screen and (max-device-width: 1024px) and (min-device-height: 1366px)): Target iPad Pro.
2386         (#results footer): Add padding below the button for testing again.
2387         * Animometer/tests/master/multiply.html: Remove the center text.
2388         * Animometer/tests/master/resources/text.js: Update the test so that in every frame the text moves.
2389         * Animometer/tests/master/text.html: Update the text sizing depending on the size of the device.
2390
2391 2016-04-08  Jon Lee  <jonlee@apple.com>
2392
2393         Have Animometer benchmark always start with complexity of 1
2394         https://bugs.webkit.org/show_bug.cgi?id=156432
2395
2396         Reviewed by Ryosuke Niwa.
2397
2398         * Animometer/tests/resources/main.js: Update the default Controller and RampController to
2399         set its minimum complexities to 1 instead of 0.
2400
2401 2016-04-08  Jon Lee  <jonlee@apple.com>
2402
2403         Fix SVG benchmark test
2404         https://bugs.webkit.org/show_bug.cgi?id=156410
2405
2406         Reviewed by Dean Jackson.
2407
2408         * Animometer/resources/extensions.js: Update Point.zero to be a static Point.
2409         * Animometer/tests/simple/resources/tiled-canvas-image.js:
2410         (Stage.call._setupTiles): Refactor.
2411         * Animometer/tests/master/resources/particles.js:
2412         (Particle.prototype.reset): Use Point.center.
2413         (complexity): We are not using a gradient background anymore, so remove the +1.
2414         * Animometer/tests/master/resources/svg-particles.js: Update to use SVG transform
2415         instead of CSS transform.
2416
2417 2016-04-06  Jon Lee  <jonlee@apple.com>
2418
2419         Update master benchmark with SVG test
2420         https://bugs.webkit.org/show_bug.cgi?id=156273
2421
2422         Reviewed by Dean Jackson.
2423         Provisionally reviewed by Said Abou-Hallawa.
2424
2425         Switch masks tests for SVG path test.
2426
2427         * Animometer/resources/debug-runner/tests.js: Move mask test here.
2428         * Animometer/resources/runner/tests.js: Add SVG test here.
2429         * Animometer/tests/dom/particles.html: Renamed from PerformanceTests/Animometer/tests/master/particles.html.
2430         * Animometer/tests/dom/resources/dom-particles.js: Renamed from PerformanceTests/Animometer/tests/master/resources/dom-particles.js.
2431
2432         * Animometer/tests/master/resources/particles.js: Add minPosition for bounds checking. Prevents particle from being
2433         partially obscured.
2434         * Animometer/tests/master/resources/svg-particles.js: Added.
2435         (Particle): The particle is either a path object or a rect using a path as a clip. The
2436         same path is used either way. For each particle create a linear gradient with a random
2437         rotation.
2438         (SVGParticlesStage): Look in #shapes to see how many different kinds of paths are available.
2439         This makes the test more generic in case other shapes need to be tested.
2440         * Animometer/tests/master/svg-particles.html: Added. Have two defs, one that houses each
2441         particle's gradient, and one that holds the shape templates.
2442
2443 2016-03-27  Hunseop Jeong  <hs85.jeong@samsung.com>
2444
2445         [EFL] REGRESSION(r188793): It made 200 layout tests and Bindings/event-target-wrapper.html performance test fail
2446         https://bugs.webkit.org/show_bug.cgi?id=148470 
2447
2448         Reviewed by Darin Adler.
2449
2450         * Skipped: Unskip the Bindings/event-target-wrapper test.
2451
2452 2016-03-23  Jon Lee  <jonlee@apple.com>
2453
2454         Update focus and leaves tests
2455         https://bugs.webkit.org/show_bug.cgi?id=155825
2456
2457         Reviewed by Simon Fraser.
2458
2459         Particles would improperly disappear, before they were fully off stage. Update the calculation.
2460         * Animometer/tests/dom/resources/leaves.js:
2461         * Animometer/tests/master/resources/leaves.js:
2462
2463         * Animometer/tests/master/focus.html: Update the center element styling.
2464         * Animometer/tests/master/resources/focus.js:
2465         (FocusElement.Utilities.createClass): Give enough space in the container for each particle for the blur
2466         to spread. It should be about 3x more space than the blur radius on each side.
2467         (Stage.call.initialize): Ditto for the center element.
2468
2469 2016-03-23  Jon Lee  <jonlee@apple.com>
2470
2471         Update support for other platforms
2472         https://bugs.webkit.org/show_bug.cgi?id=155824
2473
2474         Reviewed by Simon Fraser.
2475
2476         * Animometer/resources/debug-runner/animometer.css: Remove extraneous media query. All of it should
2477         be handled in the release stylesheet.
2478         * Animometer/resources/debug-runner/animometer.js: Add support for browsers that return an HTMLCollection
2479         when retrieving radio elements with a name, instead of the selected value.
2480         * Animometer/resources/runner/animometer.css: Adjust the queries for devices of varying widths.
2481         (.frame-container): Move overflow:hidden to #stage.
2482         * Animometer/tests/master/resources/stage.css: Consolidate into tests/resources/stage.css.
2483         * Animometer/tests/resources/stage.css:
2484         (body): Remove overflow:hidden.
2485         (#stage): Set overflow:hidden here.
2486         (#center-text): Moved from master/resources/stage.css.
2487
2488         Update stylesheet hrefs.
2489         * Animometer/tests/master/canvas-stage.html:
2490         * Animometer/tests/master/focus.html:
2491         * Animometer/tests/master/leaves.html:
2492         * Animometer/tests/master/multiply.html:
2493         * Animometer/tests/master/particles.html:
2494         * Animometer/tests/master/text.html: Adjust text size to fit in the device screen.
2495
2496 2016-03-23  Jon Lee  <jonlee@apple.com>
2497
2498         Make the benchmark require device to be in landscape orientation
2499         https://bugs.webkit.org/show_bug.cgi?id=155822
2500         rdar://problem/25258650
2501
2502         Reviewed by Ryosuke Niwa.
2503
2504         Check that the device is in landscape orientation. If not, disable the button to start it,
2505         and show a warning message.
2506
2507         * Animometer/developer.html: Add a message about how the browser should be set up. Include a
2508         warning message that appears if the orientation is incorrect on mobile devices.
2509         * Animometer/index.html: Ditto.
2510
2511         * Animometer/resources/debug-runner/animometer.css: Migrate the .hidden rule into the release
2512         stylesheet. Update the style to accommodate the new UI.
2513         * Animometer/resources/runner/animometer.css:
2514
2515         * Animometer/resources/runner/animometer.js:
2516         (window.benchmarkController.initialize): Add the orientation listener is needed.
2517         (window.benchmarkController.addOrientationListenerIfNecessary): Only mobile devices need this. Check
2518         to see for support of window.orientation.
2519         (window.benchmarkController._orientationChanged): Toggle the warning based on whether we match the
2520         landscape query. Set a state variable, which is needed for the debug harness. Call updateStartButtonState.
2521         (window.benchmarkController.updateStartButtonState):
2522         * Animometer/resources/debug-runner/animometer.js: Change _updateStartButtonState to return
2523         a boolean about whether at least one test is selected. That will be used in
2524         benchmarkController.updateStartButtonState(). Move the load event listener to the release version.
2525         (window.benchmarkController.updateStartButtonState): Override the release version, and also check
2526         that at least one test is selected.
2527
2528 2016-03-22  Geoffrey Garen  <ggaren@apple.com>
2529
2530         MallocBench: consolidate regression testing for aligned allocation
2531         https://bugs.webkit.org/show_bug.cgi?id=155762
2532
2533         Reviewed by Michael Saboff.
2534
2535         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
2536         * MallocBench/MallocBench/Benchmark.cpp:
2537         * MallocBench/MallocBench/memalign.cpp: Removed.
2538         * MallocBench/MallocBench/memalign.h: Removed. The stress_aligned test
2539         covers this and much more.
2540
2541         * MallocBench/MallocBench/stress_aligned.cpp:
2542         (benchmark_stress_aligned): Include specific tests for extreme sizes
2543         and alignments.
2544
2545 2016-03-21  Jon Lee  <jonlee@apple.com>
2546
2547         Update benchmark tests
2548         https://bugs.webkit.org/show_bug.cgi?id=155723
2549
2550         Reviewed by Darin Adler.
2551         Provisionally reviewed by Said Abou-Hallawa.
2552
2553         Add quadratic and bezier segments to the canvas path test.
2554
2555         * Animometer/resources/runner/tests.js: Some of the query strings are unnecessary.
2556         Rename the test.
2557         * Animometer/tests/master/resources/canvas-stage.js:
2558         (tune): Update to be able to take an array of possible constructors. Choose one randomly.
2559         * Animometer/tests/master/resources/canvas-tests.js:
2560         (CanvasLinePoint.Utilities.createClass): Move the point selection out to a separate
2561         function called randomPoint() for reuse. Move X_LOOPS, Y_LOOPS, and offsets into the
2562         class definition.
2563         (randomPoint): Scale the grid down a little bit so that the lines along the edge of the
2564         canvas are not cut off when the stroke size is thick.
2565         (CanvasQuadraticSegment): Added.
2566         (CanvasBezierSegment): Added.
2567         (SimpleCanvasStage): Pass in an array of the different segment types. Since line segments
2568         are short compared to the curved ones, make it twice as likely to render a line segment.
2569         (SimpleCanvasStage.animate): Update the drawing code so that we render all line segments.
2570
2571         Add a helper method that selects a random element from an array.
2572
2573         * Animometer/tests/resources/main.js:
2574         (Stage.randomElementInArray): Select a random element from the provided array.
2575
2576         * Animometer/tests/bouncing-particles/resources/bouncing-tagged-images.js: Refactor.
2577         * Animometer/tests/dom/resources/leaves.js: Ditto.
2578         * Animometer/tests/master/resources/dom-particles.js: Ditto.
2579         * Animometer/tests/master/resources/image-data.js: Ditto.
2580         * Animometer/tests/master/resources/leaves.js: Ditto.
2581         * Animometer/tests/simple/resources/simple-canvas-paths.js: Ditto.
2582
2583         Add canvas tests that includes all stroke and fill paths. This makes it possible to avoid having to
2584         include the full simple canvas suite for perf testing.
2585
2586         * Animometer/resources/debug-runner/tests.js: Add new tests. Move the canvas test into the Canvas
2587         suite.
2588         * Animometer/tests/simple/resources/simple-canvas-paths.js: Add a CanvasStroke and CanvasFill particle
2589         that random selects an object to render.
2590
2591         * Animometer/resources/debug-runner/tests.js: Move 3D suite before basic canvas suite.
2592         When updating the perf bot script, we will include the suites up to this one, but exclude
2593         the basic canvas suite.
2594
2595         Move compositing transforms test to HTML suite and remove the empty Miscellaneous suite.
2596
2597         * Animometer/resources/debug-runner/tests.js:
2598         * Animometer/tests/dom/compositing-transforms.html: Renamed from PerformanceTests/Animometer/tests/misc/compositing-transforms.html.
2599         * Animometer/tests/dom/resources/compositing-transforms.js: Renamed from PerformanceTests/Animometer/tests/misc/resources/compositing-transforms.js.
2600
2601         Clean up miscellaneous test suite. Add a canvas ellipse test, and remove the other
2602         canvas tests.
2603
2604         * Animometer/resources/debug-runner/tests.js: Add ellipse tests to the simple suite.
2605         * Animometer/tests/misc/canvas-electrons.html: Removed.
2606         * Animometer/tests/misc/canvas-stars.html: Removed.
2607         * Animometer/tests/misc/resources/canvas-electrons.js: Removed.
2608         * Animometer/tests/misc/resources/canvas-stars.js: Removed.
2609         * Animometer/tests/simple/resources/simple-canvas-paths.js: Add ellipse primitives.
2610
2611         Merge text tests together into one.
2612
2613         * Animometer/resources/runner/tests.js: Remove international.html.
2614         * Animometer/tests/master/international.html: Removed.
2615         * Animometer/tests/master/resources/text.js:
2616         (animate): Update styling. Manually calculate gradients for the shadow particles.
2617         Reduce the step size for y direction to avoid cutting text off at the margins. Increase
2618         step for x since there will be a little more room.
2619         * Animometer/tests/master/text.html: Add more translations and lay it out in a table.
2620
2621 2016-03-20  Jon Lee  <jonlee@apple.com>
2622
2623         Add a link to show debug data
2624         https://bugs.webkit.org/show_bug.cgi?id=155724
2625
2626         Reviewed by Simon Fraser.
2627
2628         * Animometer/developer.html: Attach the onclick handler to the score.
2629         * Animometer/index.html: Ditto.
2630
2631 2016-03-18  Dean Jackson  <dino@apple.com>
2632
2633         Add a basic WebGL test to Animometer
2634         https://bugs.webkit.org/show_bug.cgi?id=155475
2635         <rdar://problem/25156860>
2636
2637         Reviewed by Simon Fraser and Jon Lee.
2638
2639         Add a pretty simple WebGL test to Animometer
2640         which tries to draw a number of animated triangles.
2641         The result still needs a bit of tuning.
2642
2643         * Animometer/resources/debug-runner/tests.js: Add a new
2644         Suite for "3D" tests.
2645         * Animometer/tests/3d/resources/webgl.js: Added.
2646         * Animometer/tests/3d/webgl.html: Added.
2647
2648 2016-03-18  Jon Lee  <jonlee@apple.com>
2649
2650         Add support for statically linking to a specific test
2651         https://bugs.webkit.org/show_bug.cgi?id=155631
2652
2653         Rubber-stamped by Darin Adler.
2654
2655         * Animometer/developer.html: Update to call restartBenchmark instead so that it works for
2656         both the preset and the manually set test sessions.
2657         * Animometer/resources/debug-runner/animometer.css:
2658         (.tree .link): Style the "link" UI.
2659         * Animometer/resources/debug-runner/animometer.js: Move updateDisplay from suitesManager
2660         to optionsManager since this is a visual option and has nothing to do with the suites
2661         settings. Update the construction of the test UI by including a "link" after each test.
2662         Clicking on that link brings up a JS prompt with a URL and a query string with the current
2663         parameters of the controller, and selected test. Pasting this URL into the location bar will
2664         automatically start running the selected test.
2665         (suitesManager.suitesFromQueryString): Iterate through the Suites and tests and find the one
2666         that matches the provided parameters. Returns an object similar to the form in
2667         suitesManager.updateLocalStorageFromUI.
2668         (benchmarkController.initialize): After settings up the events and options, try parsing the
2669         query string and running the benchmark immediately. Otherwise, fall back to the form.
2670         (benchmarkController.startBenchmark): Store the options and suites into member variables for
2671         reuse in restartBenchmark.
2672         (benchmarkController.startBenchmarkImmediatelyIfEncoded): Convert the query string to an object.
2673         If that's successful, find the suite and test referenced in the query string. Start the benchmark if
2674         the search for the test succeeded.
2675         * Animometer/resources/extensions.js:
2676         (Utilities.stripNonASCIICharacters): Helper method to convert the name of the suite and test
2677         into a query-string-friendly version.
2678         (Utilities.convertObjectToQueryString): Helper method to convert an object to query string
2679         format.
2680         (Utilities.convertQueryStringToObject): Helper method to convert query string into an object
2681         with properties and values.
2682
2683 2016-03-16  Jon Lee  <jonlee@apple.com>
2684
2685         Add a new benchmark test
2686         https://bugs.webkit.org/show_bug.cgi?id=155570
2687
2688         Reviewed by Simon Fraser.
2689
2690         New Leaves test includes various image sizes and opacity.
2691
2692         * Animometer/resources/debug-runner/tests.js: Add it to the HTML test suite.
2693         * Animometer/tests/dom/leaves.html: Added.
2694         * Animometer/tests/dom/resources/leaves.js: Added. Override the
2695         (Particle.call.reset): Uses a range of sizes, and opacity.
2696         (Particle.call.animate): Opacity goes up then down. When it hits 0, reset the particle.
2697         (Particle.call.move): Set transform and opacity.
2698         * Animometer/tests/master/resources/leaves.js: Get rid of the closure so that it
2699         can be used in this test. Update the relative path so that it works in both the master
2700         and dom test suite.
2701
2702 2016-03-15  Simon Fraser  <simon.fraser@apple.com>
2703
2704         Add developer Animometer test that bounces P3-tagged images
2705         https://bugs.webkit.org/show_bug.cgi?id=155511
2706
2707         Reviewed by Tim Horton.
2708         
2709         Add a test for rendering performance of tagged images. The 5 images are tagged
2710         with the Display P3 colorspace.
2711
2712         * Animometer/resources/debug-runner/tests.js:
2713         * Animometer/tests/bouncing-particles/bouncing-tagged-images.html: Added.
2714         * Animometer/tests/bouncing-particles/resources/bouncing-tagged-images.js: Added.
2715         * Animometer/tests/bouncing-particles/resources/image1.jpg: Added.
2716         * Animometer/tests/bouncing-particles/resources/image2.jpg: Added.
2717         * Animometer/tests/bouncing-particles/resources/image3.jpg: Added.
2718         * Animometer/tests/bouncing-particles/resources/image4.jpg: Added.
2719         * Animometer/tests/bouncing-particles/resources/image5.jpg: Added.
2720
2721 2016-03-09  Jon Lee  <jonlee@apple.com>
2722
2723         Enhance existing Animometer tests
2724         https://bugs.webkit.org/show_bug.cgi?id=155261
2725
2726         Reviewed by Simon Fraser.
2727
2728         * Animometer/tests/master/resources/canvas-tests.js: Add a gradient to
2729         the filled circles. To expose more of the gradient, add another circle.
2730         (SimpleCanvasStage.call.animate): For each frame, create a gradient with
2731         undulating stop points and colors. Fill the circles twice; once with
2732         the solid color, and once with the gradient.
2733
2734         * Animometer/tests/master/resources/dom-particles.js: Refactor the
2735         emission variables into a separate stage for this test. Add a colorOffset
2736         variable to make the colors of each particle slightly different, since
2737         the ramp controller can add large numbers of particles all at once, which
2738         would otherwise get all the same color.
2739         * Animometer/tests/master/resources/particles.js:
2740         (initialize): Remove the code specific to the SVG mask test.
2741
2742         * Animometer/tests/master/resources/image-data.js:
2743         (initialize): Remove unused local variable.
2744         * Animometer/tests/master/resources/multiply.js:
2745         (initialize): Make the test harder by adding more total particles.
2746
2747 2016-03-09  Jon Lee  <jonlee@apple.com>
2748
2749         Add text tests
2750         https://bugs.webkit.org/show_bug.cgi?id=155257
2751
2752         Reviewed by Simon Fraser.
2753
2754         * Animometer/resources/extensions.js:
2755         (UnitBezier.Utilities.createClass): Add a class that computes Bezier points
2756         assuming that two of the control points are at (0,0) and (1,1). Taken from
2757         WebCore/platform/graphics/UnitBezier.h
2758         * Animometer/resources/runner/tests.js: Add a test for English text and one
2759         for international text.
2760         * Animometer/tests/master/international.html: Added.
2761         * Animometer/tests/master/resources/text.js: Added. The test assumes there is
2762         #template div which it will copy. The copies are placed behind the template, and
2763         are set with different colors each frame. They are moved around with CSS transform.
2764         * Animometer/tests/master/text.html: Added.
2765
2766         Remove the other text tests, since these ones cover the same techniques.
2767         * Animometer/tests/text/layering-text.html: Removed.
2768         * Animometer/tests/text/resources/layering-text.js: Removed.
2769         * Animometer/tests/text/resources/text-boxes.js: Removed.
2770         * Animometer/tests/text/text-boxes.html: Removed.
2771         * Animometer/resources/debug-runner/tests.js:
2772
2773 2016-03-09  Jon Lee  <jonlee@apple.com>
2774
2775         Add a new image test
2776         https://bugs.webkit.org/show_bug.cgi?id=155232
2777
2778         Reviewed by Dean Jackson.
2779         Provisionally reviewed by Said Abou-Hallawa.
2780
2781         The image test renders PNGs and moves them with translate and rotate
2782         transforms. Each particle has a lifetime, and when the lifetime ends or
2783         the particle goes offscreen, it resets itself somewhere on the stage.
2784
2785         * Animometer/resources/debug-runner/tests.js: Remove the CSS bouncing PNG
2786         images test, because this one tests the same technique.
2787         * Animometer/resources/runner/tests.js: Added here as "Leaves".
2788         * Animometer/tests/master/leaves.html: Added.
2789         * Animometer/tests/master/resources/particles.js: Refactor out the parts
2790         specific to the DOM particles test. Consequently make velocity a public
2791         member.
2792         (initialize): The options parameter is never used, so remove it.
2793         (animate): Remove unused local variable.
2794         * Animometer/tests/master/resources/leaves.js: Maintains a focal point that
2795         moves back and forth across the canvas. That point affects the velocity
2796         of the particles. When the focal point is on the leftmost side, it is
2797         between [-6, -2], and set in reset(). When the focal point is on the rightmost
2798         side, it will be from [2, 6].
2799         * Animometer/tests/master/resources/dom-particles.js: Move JS specific to
2800         this test here from particles.js.
2801
2802         New images.
2803         * Animometer/tests/master/resources/compass100.png: Added.
2804         * Animometer/tests/master/resources/console100.png: Added.
2805         * Animometer/tests/master/resources/contribute100.png: Added.
2806         * Animometer/tests/master/resources/debugger100.png: Added.
2807         * Animometer/tests/master/resources/inspector100.png: Added.
2808         * Animometer/tests/master/resources/layout100.png: Added.
2809         * Animometer/tests/master/resources/performance100.png: Added.
2810         * Animometer/tests/master/resources/script100.png: Added.
2811         * Animometer/tests/master/resources/shortcuts100.png: Added.
2812         * Animometer/tests/master/resources/standards100.png: Added.
2813         * Animometer/tests/master/resources/storage100.png: Added.
2814         * Animometer/tests/master/resources/styles100.png: Added.
2815         * Animometer/tests/master/resources/timeline100.png: Added.
2816
2817 2016-03-03  Jon Lee  <jonlee@apple.com>
2818
2819         Add ability to retrieve raw data from release harness
2820         https://bugs.webkit.org/show_bug.cgi?id=155026
2821
2822         Reviewed by Simon Fraser.
2823
2824         * Animometer/developer.html: Remove the special UI in the debug
2825         harness.
2826         * Animometer/resources/runner/animometer.css: Add styles for the
2827         overlay.
2828         * Animometer/resources/runner/animometer.js: Let 'j' show the JSON
2829         results, but only if the overlay doesn't exist. Add 'esc' key to dismiss
2830         the overlay.
2831         (window.benchmarkController.selectResults): Cycle the cases around
2832         so that the first 's' press selects both the benchmark score and
2833         the individual test scores.
2834         * Animometer/resources/debug-runner/animometer.css: Remove
2835         unneeded rules.
2836         * Animometer/resources/debug-runner/animometer.js: Make the same
2837         call to handleKeyPress.
2838
2839 2016-03-03  Jon Lee  <jonlee@apple.com>
2840
2841         Make sure multiply test particles have at least some opacity
2842         https://bugs.webkit.org/show_bug.cgi?id=155027
2843
2844         Reviewed by Simon Fraser.
2845
2846         * Animometer/tests/master/resources/multiply.js:
2847         (initialize): Have the elements spin a little faster to make it
2848         more obvious when the system is being stressed.
2849         (_addTile):
2850         (animate): Make sure each element being animated has at least 1%
2851         opacity.
2852
2853 2016-03-03  Jon Lee  <jonlee@apple.com>
2854
2855         Update image test
2856         https://bugs.webkit.org/show_bug.cgi?id=154962
2857
2858         Rubber-stamped by Darin Adler.
2859
2860         * Animometer/tests/master/image-data.html: Make each canvas have a compositing layer.
2861         * Animometer/tests/master/resources/image-data.js:
2862         (initialize): Update to load multiple images serially.
2863         (_loadImage):
2864         (tune): Use display instead of visibility. When showing an element anew, move it somewhere else.
2865         (_createTestElement): Set up the element, refactor out placement of element to _refreshElement.
2866         (_refreshElement): Place element in a tile grid.
2867         (animate):
2868         (_getRandomNeighboringPixelIndex): Use a more conservative distribution to make the effect last a little longer.
2869         * Animometer/tests/master/resources/compass.svg: Added.
2870         * Animometer/tests/master/resources/console.svg: Added.
2871         * Animometer/tests/master/resources/contribute.svg: Added.
2872         * Animometer/tests/master/resources/debugger.svg: Added.
2873         * Animometer/tests/master/resources/inspector.svg: Added.
2874         * Animometer/tests/master/resources/layout.svg: Added.
2875         * Animometer/tests/master/resources/performance.svg: Added.
2876         * Animometer/tests/master/resources/script.svg: Added.
2877         * Animometer/tests/master/resources/shortcuts.svg: Added.
2878         * Animometer/tests/master/resources/standards.svg: Added.
2879         * Animometer/tests/master/resources/storage.svg: Added.
2880         * Animometer/tests/master/resources/styles.svg: Added.
2881         * Animometer/tests/master/resources/timeline.svg: Added.
2882
2883 2016-03-02  Jon Lee  <jonlee@apple.com>
2884
2885         Add some new controllers, and refine tests
2886         https://bugs.webkit.org/show_bug.cgi?id=154914
2887
2888         Reviewed by Simon Fraser.
2889
2890         Improve tests.
2891
2892         * Animometer/tests/master/focus.html: Move each particle into a layer. Put the particle
2893         in a container element. Expand the container element by the max blur radius, and clip
2894         overflow. This way, when the blur is applied, it avoids causing layer resizes.
2895         * Animometer/tests/master/resources/focus.js:
2896         (FocusElement.Utilities.createClass): Create a container element, and put the particle
2897         inside.
2898         (hide): Set display:none.
2899         (show): Set display:block.
2900         (animate): Apply filters to the container element.
2901         (FocusStage.call.initialize): Instead of inserting and removing elements from the DOM,
2902         keep them in the stage, but set the display style instead. Use this._offsetIndex to
2903         keep track of which elements are displayed.
2904         (FocusStage.call.animate): Move some calculations around to avoid doing unneeded math.
2905         (FocusStage.call.getBlurValue): Make sure elements are always blurred.
2906         (FocusStage.call.getOpacityValue): Make sure elements have some opacity.
2907         * Animometer/tests/master/resources/multiply.js: Have the particles on the edge of the
2908         stage appear less black when the complexity gets large enough.
2909         (tune): Have this._distanceFactor calculate the factor. Avoid calculating square root each
2910         frame.
2911
2912 2016-03-02  Jon Lee  <jonlee@apple.com>
2913
2914         Add some new controllers, and refine tests
2915         https://bugs.webkit.org/show_bug.cgi?id=154914
2916
2917         Reviewed by Simon Fraser.
2918
2919         Add a controller that centers around 30 fps instead of 60 fps.
2920
2921         * Animometer/developer.html: Add a new option.
2922         * Animometer/resources/debug-runner/animometer.js:
2923         * Animometer/resources/runner/animometer.js:
2924         (this._processData.findRegression): When calculating the complexity-frameLength regression,
2925         check the controller, and use a 30 fps baseline if needed.
2926         * Animometer/resources/statistics.js:
2927         (Regression.Utilities.createClass): Update to allow clients to specify the baseline
2928         frame length.
2929         * Animometer/tests/resources/main.js:
2930         (tune): Override some of the constants in RampController. Move those constants out for
2931         easier reading.
2932
2933         Add a fixed controller, with no step.
2934
2935         * Animometer/developer.html: Add a controller that takes no step.
2936         * Animometer/resources/debug-runner/animometer.js:
2937         * Animometer/tests/resources/main.js:
2938         (Rotater.Utilities.createClass):
2939
2940         Switch to ramp controller as default.
2941
2942         * Animometer/developer.html: Increase the test length to 20 seconds.
2943         * Animometer/resources/debug-runner/animometer.js: Rename the "adjustment" field to "controller"
2944         since that is a more accurate description.
2945         * Animometer/resources/debug-runner/graph.js:
2946         * Animometer/resources/runner/animometer.js: Update preferences for release suite.
2947
2948 2016-02-26  Jon Lee  <jonlee@apple.com>
2949
2950         Address Dean's comments in 154673.
2951
2952         * Animometer/developer.html:
2953         * Animometer/resources/debug-runner/animometer.js:
2954         * Animometer/resources/debug-runner/graph.js:
2955
2956 2016-02-25  Jon Lee  <jonlee@apple.com>
2957
2958         Update animation benchmark and tests
2959         https://bugs.webkit.org/show_bug.cgi?id=154673
2960
2961         Reviewed by Dean Jackson.
2962
2963         Update test visuals.
2964
2965         * Animometer/tests/master/resources/canvas-tests.js:
2966         (CanvasLineSegment.Utilities.createClass): Line segments near the edge of the stage get clipped.
2967         Update circle position and radius to minimize impact.
2968         * Animometer/tests/master/resources/particles.js: Get rid of rotating gradient background, and
2969         have 3 locations for emitting particles.
2970         * Animometer/tests/master/resources/multiply.js: Update distance metric so that fringe tiles get
2971         more color.
2972         (initialize): Fix some of the math for laying out the tiles.
2973         (animate): When a tile isn't used, set visibility: hidden.
2974
2975 2016-02-24  Jon Lee  <jonlee@apple.com>
2976
2977         Update animation benchmark and tests
2978         https://bugs.webkit.org/show_bug.cgi?id=154673
2979
2980         Reviewed by Dean Jackson.
2981
2982         Update the ramp controller.
2983
2984         The controller refines the complexity interval to test across.
2985
2986         * Animometer/resources/statistics.js: Add functions that estimate cumulative distribution function.
2987         (Regression): For the flat regression, force the first segment to be at 60 fps.
2988         (valueAt): Add convenience function to return interpolated value based on the regression used.
2989         (_calculateRegression): Include the number of points included for both segments, and the piecewise
2990         errors.
2991         * Animometer/tests/resources/math.js: Make the Kalman estimator subclass Experiment, and allow it
2992         to be reset.
2993
2994         * Animometer/tests/resources/main.js: Initialize the tier such that it starts at 10^0 = 1.
2995         Increase the number of ramps. Maintain three FPS thresholds-- the frame rate of interest, a limit
2996         on the lowest FPS we care to go for later interpolation, and a minimum FPS threshold we want to
2997         aim for each ramp. Also keep three estimators: a running average of the change point, a minimum
2998         boundary for each ramp, and an estimator for all the frames within an interval. The first two
2999         are used to determine the parameters of the next ramp, and the latter allows us to refine the
3000         parameters.
3001         (update): During the tier phase, it is possible that the highest complexity possible for a test
3002         won't stress the system enough to trigger stopping the tier phase and transitioning to the ramps.
3003         If the complexity doesn't change when going to the next tier, we've maxed the test out, and move
3004         on. When the tier phase completed, turn off Controller.frameLengthEstimator, which estimates the
3005         FPS at each tier.
3006         (tune): At each interval, look at the confidence distribution of being on the 60 FPS side or the
3007         slow side. If the slowest FPS we achieve at the ramp's maximum complexity is not at least
3008         _fpsRampSlowThreshold, then increase the maximum complexity. If we ever achieve 60 FPS, increase
3009         the ramp's minimum complexity to that level. If, at an even lower complexity, a glitch causes the
3010         FPS to drop, we reset the minimum complexity.
3011
3012         Have the bootstrap calculation occur between tests. Clean up harness.
3013
3014         * Animometer/resources/debug-runner/animometer.js: Run bootstrap after a test has
3015         completed to avoid doing all of it at the end before showing the results. Clean up
3016         parameters being passed around.
3017         * Animometer/resources/debug-runner/tests.js:
3018         (text):
3019         * Animometer/resources/runner/animometer.js:
3020         (this._processData.calculateScore): Save the results to the same object holding the data.
3021         (this._processData._processData): In the case where a file is dragged, calculate the score
3022         serially. Grab the results object and move it to the results variable and remove it from
3023         the data object. This avoids serializing the results into the JSON.
3024         (this._processData.findRegression): Include the samples used for bootstrapping. Reduce the
3025         resample size to shorten the wait.
3026         * Animometer/resources/runner/benchmark-runner.js:
3027         * Animometer/resources/statistics.js:
3028         (bootstrap): Update how bootstrapData is sorted. In some regression results the mix of
3029         floats and integers causes an alphabetical sort to occur.
3030         * Animometer/resources/strings.js:
3031
3032         Add meta charset so that encodings between harness and test match.
3033
3034         * Animometer/tests/bouncing-particles/bouncing-canvas-images.html:
3035         * Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html:
3036         * Animometer/tests/bouncing-particles/bouncing-css-images.html:
3037         * Animometer/tests/bouncing-particles/bouncing-css-shapes.html:
3038         * Animometer/tests/bouncing-particles/bouncing-svg-images.html:
3039         * Animometer/tests/bouncing-particles/bouncing-svg-shapes.html:
3040         * Animometer/tests/master/canvas-stage.html:
3041         * Animometer/tests/master/focus.html:
3042         * Animometer/tests/master/image-data.html:
3043         * Animometer/tests/master/multiply.html:
3044         * Animometer/tests/master/particles.html:
3045         * Animometer/tests/misc/canvas-electrons.html:
3046         * Animometer/tests/misc/canvas-stars.html:
3047         * Animometer/tests/misc/compositing-transforms.html:
3048         * Animometer/tests/simple/simple-canvas-paths.html:
3049         * Animometer/tests/simple/tiled-canvas-image.html:
3050         * Animometer/tests/template/template-canvas.html:
3051         * Animometer/tests/template/template-css.html:
3052         * Animometer/tests/template/template-svg.html:
3053         * Animometer/tests/text/layering-text.html:
3054         * Animometer/tests/text/text-boxes.html:
3055
3056         Update test harness reporting.
3057
3058         * Animometer/developer.html: Add missing meta charset.
3059         * Animometer/index.html: Remove unnecessary utf-8 declaration.
3060         * Animometer/resources/debug-runner/animometer.css: Add convenience classes for
3061         formatting the results table.
3062         * Animometer/resources/debug-runner/animometer.js: Adjust which stats are shown.
3063         * Animometer/resources/debug-runner/tests.js: Display bootstrapping statistics.
3064         * Animometer/resources/strings.js: Move strings not used by the release harness.
3065
3066         Switch to a pseudo-random number generator.
3067
3068         * Animometer/resources/statistics.js: Add a Pseudo class, with a simple
3069         pseudo-random number generator.
3070         (_calculateRegression): Reset the generator before running bootstrap.
3071         (bootstrap): Deleted.
3072
3073         Replace Math.random with Pseudo.random.
3074         * Animometer/tests/master/resources/canvas-tests.js:
3075         * Animometer/tests/master/resources/focus.js:
3076         * Animometer/tests/master/resources/particles.js:
3077         * Animometer/tests/resources/main.js:
3078
3079         Use bootstrapping to get confidence interval in the breakpoint.
3080
3081         For the ramp controller, calculate the piecewise regression, and then use
3082         bootstrapping in order to find the 95% confidence interval. Use the raw data.
3083
3084         * Animometer/developer.html: Default to the complexity graph. Add a legend
3085         checkbox to toggle visibility of the bootstrap score and histogram.
3086         * Animometer/resources/debug-runner/animometer.css: Make some more space to show
3087         the old raw and average scores in the legend. Add new styles for the data.
3088         * Animometer/resources/debug-runner/graph.js:
3089         (_addRegressionLine): Allow passing an array for the variance bar tied to the
3090         regression line. Now |stdev| is |range|.
3091         (createComplexityGraph): Add bootstrap median, and overlay a histogram of
3092         the bootstrap samples. Switch raw samples from circles to X's.
3093         (onComplexityGraphOptionsChanged): Allow toggling of the bootstrap data.
3094         (onGraphTypeChanged): Move the regressions for the raw and average samples to the
3095         legend. In the subtitle use the bootstrap median, and include the 95% confidence
3096         interval.
3097         * Animometer/resources/runner/animometer.js:
3098         (this._processData.findRegression): Factor out the code that determines which
3099         samples to include when calculating the piecewise regression. For series that have
3100         many samples, or a wider range of recorded complexities, throw away the 2.5%
3101         lowest and highest samples before calculating the regression. Keep all samples
3102         if the number of samples to regress is small or the range of complexities is
3103         narrow.
3104         (this._processData._calculateScore): Factor out regression calculation to
3105         findRegression(). Bootstrap the change point of the regression. The score is the
3106         median.
3107         * Animometer/resources/statistics.js:
3108         (_calculateRegression): Correct an issue in the calculation of the regression, where
3109         the denominator can be 0.
3110         (bootstrap): Template for bootstrapping. Create a bootstrap sample array, Create
3111         re-samples by random selection with replacement. Return the 95% confidence samples,
3112         the bootstrap median, mean, and the data itself.
3113         * Animometer/resources/strings.js: Add bootstrap.
3114         * Animometer/tests/resources/main.js:
3115         (processSamples): Don't prematurely cut the sample data.
3116
3117         Fix graph drawing.
3118
3119         * Animometer/resources/debug-runner/animometer.js: Add spacing in the JSON output.
3120         Multiple tests output a lot of JSON and can hang when selecting JSON with no whitespace.
3121         * Animometer/resources/debug-runner/animometer.css:
3122         (#complexity-graph .series.raw circle): Update the color.
3123         * Animometer/resources/debug-runner/graph.js: Use the FPS axis instead of the
3124         complexity axis, which can vary in domain. For determining the complexity domain,
3125         only use samples after samplingTimeOffset.
3126
3127         Allow dropping results JSON.
3128
3129         * Animometer/developer.html: Add a button.
3130         * Animometer/resources/debug-runner/animometer.css:
3131         * Animometer/resources/debug-runner/animometer.js: Read the data and go straight
3132         to the dashboard. With JSON output, write out only the options and the raw data.
3133
3134         Teach the harness to evaluate the samples and determine the test score.
3135
3136         This will allow us to update how the score is calculated separately from the samples recorded.
3137         This also prepares the harness to be able to accept JSON of prior runs.
3138
3139         * Animometer/resources/strings.js: Clean up and remove unneeded strings and reduce some of the
3140         hierarchy.
3141         * Animometer/resources/debug-runner/tests.js: Update to use the new strings.
3142
3143         * Animometer/tests/resources/main.js: Allow all controllers to show a complexity-FPS graph.
3144         (_processComplexitySamples): Factor out some of the sample processing done in the ramp
3145         controller for the benefit of the other controllers. |complexitySamples| contains a list of
3146         samples. Sort the samples by complexity. Optionally remove the top x% of samples.
3147         Group them, and calculate distribution of samples within the same complexity, and add those as
3148         new entries into |complexityAverageSamples|.
3149         (Controller.processSamples): Move the code responsible for determining the complexity and FPS
3150         scores out to ResultsDashboard. The structure of the data returned by the controller is:
3151
3152         {
3153             controller: [time-regression, time-regression, ...], // optional, data specific to controller
3154             marks: [...],
3155             samples: {                    // all of the sample data
3156                 controller: [...],
3157                 complexity: [...],        // processed from controller samples
3158                 complexityAverage: [...], // processed from complexity samples
3159             }
3160         }
3161
3162         (AdaptiveController.processSamples): Adding the target frame length is no longer necessary; we
3163         now pass the test options to the graph.
3164         (Regression): Move to statistics.js.
3165         * Animometer/resources/statistics.js: Move Regression to here. Add a check if the sampling range
3166         only contains one sample, since we cannot calculate a regression from one sample point.
3167
3168         Teach the test harness to evaluate the data.
3169         * Animometer/resources/runner/animometer.js:
3170         (ResultsDashboard): Store the options used to run the test and the computed results/score separately
3171         from the data. The results are stored as:
3172
3173         {
3174             score: /* geomean of iteration score */,
3175             iterationsResults: [
3176                 {
3177                     score: /* geomean of tests */,
3178                     testsResults: {
3179                         suiteName: {
3180                             testName: {
3181                                 controller: {
3182                                     average:
3183                                     concern:
3184                                     stdev:
3185                                     percent:
3186                                 },
3187                                 frameLength: { ... },
3188                                 complexity: {
3189                                     complexity:
3190                                     stdev:
3191                                     segment1:
3192                                     segment2:
3193                                 },
3194                                 complexityAverage: { ... }
3195                             },
3196                             testName: { ... },
3197                         },
3198                         ... next suite ...
3199                     }
3200                 },
3201                 { ...next iteration... }
3202             ]
3203         }
3204
3205         * Animometer/resources/debug-runner/animometer.js: Pass options around instead of relying
3206         on what was selected in the form. This will later allow for dropping previous results, and
3207         using those runs' options when calculating scores.
3208         (ResultsTable._addGraphButton): Simplify button action by using attached test data.
3209         * Animometer/resources/debug-runner/graph.js: Refactor to use the data.
3210
3211         Consolidate JS files, and move statistics out to a separate JS.
3212
3213         Preparation for having the Controller only handle recording and storage of the samples,
3214         and leave the evaluation of the test score out to the harness. Move Experiment to
3215         a new statistics.js, where Regression will also eventually go. Get rid of algorithm.js
3216         and move it to utilities.js since the Heap is used only for Experiments.
3217
3218         * Animometer/tests/resources/algorithm.js: Removed. Heap is in utilities.js.
3219         * Animometer/tests/resources/sampler.js: Removed. Experiment is in statistics.js,
3220         Sampler in main.js.
3221         * Animometer/tests/resources/main.js: Move Sampler here.
3222         * Animometer/resources/statistics.js: Added. Move Statistics and Experiment here.
3223         * Animometer/resources/extensions.js: Move Heap here. Attach static method to create
3224         a max or min heap to Heap, instead of a new Algorithm object.
3225
3226         Update JS files.
3227         * Animometer/developer.html:
3228         * Animometer/index.html:
3229         * Animometer/tests/bouncing-particles/bouncing-canvas-images.html:
3230         * Animometer/tests/bouncing-particles/bouncing-canvas-shapes.html:
3231         * Animometer/tests/bouncing-particles/bouncing-css-images.html:
3232         * Animometer/tests/bouncing-particles/bouncing-css-shapes.html:
3233         * Animometer/tests/bouncing-particles/bouncing-svg-images.html:
3234         * Animometer/tests/bouncing-particles/bouncing-svg-shapes.html:
3235         * Animometer/tests/master/canvas-stage.html:
3236         * Animometer/tests/master/focus.html:
3237         * Animometer/tests/master/image-data.html:
3238         * Animometer/tests/master/multiply.html:
3239         * Animometer/tests/master/particles.html:
3240         * Animometer/tests/misc/canvas-electrons.html:
3241         * Animometer/tests/misc/canvas-stars.html:
3242         * Animometer/tests/misc/compositing-transforms.html:
3243         * Animometer/tests/simple/simple-canvas-paths.html:
3244         * Animometer/tests/simple/tiled-canvas-image.html:
3245         * Animometer/tests/template/template-canvas.html:
3246         * Animometer/tests/template/template-css.html:
3247         * Animometer/tests/template/template-svg.html:
3248         * Animometer/tests/text/layering-text.html:
3249         * Animometer/tests/text/text-boxes.html:
3250
3251         Fix the cursor in the graph analysis when the min
3252         complexity is not 0.
3253
3254         * Animometer/resources/debug-runner/graph.js:
3255         (_addRegression):
3256         (createComplexityGraph):
3257
3258 2016-02-23  Geoffrey Garen  <ggaren@apple.com>
3259
3260         Fix some issues in MallocBench
3261         https://bugs.webkit.org/show_bug.cgi?id=154600
3262
3263         Reviewed by Oliver Hunt.
3264
3265         * MallocBench/MallocBench.xcodeproj/project.pbxproj: Make nimlang.ops
3266         a part of the build so that it copies to the right place for execution.
3267
3268         * MallocBench/MallocBench/Interpreter.cpp:
3269         (Interpreter::Interpreter): Don't try to open .ops files for writing
3270         because we're only going to read and we might need extra permissions to
3271         write.
3272
3273         (Interpreter::~Interpreter): Give more context when opening a file fails
3274         to help with debugging.
3275
3276         * MallocBench/MallocBench/stress.cpp:
3277         (benchmark_stress): Reduce iterations to complete in less than 10 seconds.
3278
3279 2016-02-23  Michael Saboff  <msaboff@apple.com>
3280
3281         Unreviewed change to revert extraneous changes made part of change set 196955.
3282
3283         * MallocBench/MallocBench/Interpreter.cpp:
3284         (Interpreter::doMallocOp):
3285         (Interpreter::Thread::switchTo):
3286         (writeData): Deleted.
3287
3288 2016-02-19  Michael Saboff  <msaboff@apple.com>
3289
3290         MallocBench: Added recording for nimlang website, new recording details and added new options
3291         https://bugs.webkit.org/show_bug.cgi?id=154485
3292
3293         Reviewed by Geoff Garen.
3294
3295         Added new capabilities to MallocBench.  These include:
3296             Added a recording of http://nim-lang.org/docs/lib.html.
3297             Added thread id to the recording and the ability to playback switching threads in MallocBench
3298             Added aligned allocations to recordings and the ability to playback
3299             Added --use-thread-id option to honor recorded thread ids
3300             Added --detailed-report to output remaining allocations by size after playback
3301             Added --no-warmup to not run the warm up iteration
3302
3303         Changed the way that options are passed down to the benchmarks.  Instead of passing individual
3304         boolean or numeric option values, just pass a reference the CommandLine itself.  Each benchmark
3305         can access the options that are appropriate.  The Benchmark class also uses the options for
3306         is parallel, run counts and warm up.
3307
3308         Added thread id and aligned malloc to the Op by noticing that structure padding and Opcode allowed
3309         for another 32 bits of data.  Breaking that unused 32 bits into a 16 bit thread id value and a
3310         16 bit log base 2 of the alignment for aligned malloc allowed for existing recordings to playback
3311         without any incompatibilities.
3312
3313         Threaded operation is simulated by creating threads as needed.  As long as the next Op's thread id
3314         is the same as the last, operation continues as normal.  When the next Op has a different thread id,
3315         we switch to that thread using the shared Op stream to continue playing back.  There is a mutex to
3316         assure that only one thread is really running at a time and a condition variable used to wait
3317         that the current thread id matches each block thread's thread id.  This doesn't simulate true
3318         concurrent threading, but is instead plays back Ops recorded for multiple thread faithfully.
3319
3320         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
3321         * MallocBench/MallocBench/Benchmark.cpp:
3322         (deallocateHeap):
3323         (Benchmark::Benchmark):
3324         (Benchmark::runOnce):
3325         (Benchmark::run):
3326         * MallocBench/MallocBench/Benchmark.h:
3327         (Benchmark::isValid):
3328         * MallocBench/MallocBench/CommandLine.cpp:
3329         (CommandLine::printUsage):
3330         * MallocBench/MallocBench/CommandLine.h:
3331         (CommandLine::isValid):