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