| 2018-08-23 Don Olmstead <don.olmstead@sony.com> |
| |
| [CMake] Add HAVE_MALLOC_TRIM definition |
| https://bugs.webkit.org/show_bug.cgi?id=188897 |
| |
| Reviewed by Konstantin Tokarev. |
| |
| MallocBench should only be built on Apple platforms and platforms that |
| have malloc_trim. |
| |
| * CMakeLists.txt: |
| |
| 2018-08-13 Jon Lee <jonlee@apple.com> |
| |
| Update MotionMark version number |
| https://bugs.webkit.org/show_bug.cgi?id=188535 |
| <rdar://problem/43254078> |
| |
| Reviewed by Said Abou-Hallawa. |
| |
| * MotionMark/resources/strings.js: To 1.1. |
| |
| 2018-08-13 Jon Lee <jonlee@apple.com> |
| |
| [MotionMark] Update Multiply test |
| https://bugs.webkit.org/show_bug.cgi?id=188532 |
| <rdar://problem/43252151> |
| |
| Reviewed by Said Abou-Hallawa. |
| |
| Update Multiply test to have a larger upper limit of number of particles. Cycle through three |
| different ways of hiding an element. Add new Multiple suite that isolates those components for |
| debugging purposes. |
| |
| * MotionMark/resources/debug-runner/tests.js: |
| * MotionMark/tests/dom/multiply.html: Added. |
| * MotionMark/tests/dom/resources/multiply.js: Added. |
| * MotionMark/tests/master/resources/multiply.js: |
| |
| 2018-08-13 Jon Lee <jonlee@apple.com> |
| |
| [MotionMark] Update Leaves test |
| https://bugs.webkit.org/show_bug.cgi?id=188530 |
| <rdar://problem/43251862> |
| |
| Reviewed by Said Abou-Hallawa. |
| |
| Update Leaves test to include opacity and scale. Add new Leaves suite that isolate those components |
| for debugging purposes. |
| |
| * MotionMark/resources/debug-runner/tests.js: |
| * MotionMark/tests/dom/leaves.html: |
| * MotionMark/tests/dom/resources/leaves.js: |
| * MotionMark/tests/master/resources/leaves.js: |
| |
| 2018-08-10 Ben Richards <benton_richards@apple.com> |
| |
| Add ability to ignore process prewarming for launch time benchmark |
| https://bugs.webkit.org/show_bug.cgi?id=188462 |
| |
| Reviewed by Ryosuke Niwa. |
| |
| Added a flag to the new tab benchmark that will open a blank tab before the tab to be measured in order to ignore process prewarming. |
| |
| * LaunchTime/launch_time.py: |
| (DefaultLaunchTimeHandler): |
| (DefaultLaunchTimeHandler.on_receive_stop_time): Deleted. |
| (DefaultLaunchTimeHandler.on_receive_stop_signal): |
| (DefaultLaunchTimeHandler.do_HEAD): |
| (DefaultLaunchTimeHandler.do_GET): |
| (DefaultLaunchTimeHandler.do_POST): |
| (LaunchTimeBenchmark): |
| (LaunchTimeBenchmark._standard_deviation): Fixed divide by zero bug when '-n' is set to 1 |
| (LaunchTimeBenchmark.open_tab): Added option to open a blank tab |
| (LaunchTimeBenchmark.run): |
| * LaunchTime/new_tab.py: |
| (NewTabBenchmark.initialize): |
| (NewTabBenchmark.run_iteration): |
| (NewTabBenchmark.will_parse_arguments): |
| (NewTabBenchmark.did_parse_arguments): |
| (NewTabBenchmark.ResponseHandler.Handler.get_test_page): |
| (NewTabBenchmark.ResponseHandler.Handler.on_receive_stop_time): Deleted. |
| (NewTabBenchmark.ResponseHandler.Handler.on_receive_stop_signal): |
| (NewTabBenchmark): |
| * LaunchTime/startup.py: |
| (StartupBenchmark.ResponseHandler.Handler.get_test_page): |
| (StartupBenchmark.ResponseHandler.Handler.on_receive_stop_time): Deleted. |
| (StartupBenchmark.ResponseHandler.Handler.on_receive_stop_signal): |
| (StartupBenchmark): |
| |
| 2018-08-10 Truitt Savell <tsavell@apple.com> |
| |
| Unreviewed, rolling out r234750. |
| |
| Caused 185 perf test failures. |
| |
| Reverted changeset: |
| |
| "Add ability to ignore process prewarming for launch time |
| benchmark" |
| https://bugs.webkit.org/show_bug.cgi?id=188462 |
| https://trac.webkit.org/changeset/234750 |
| |
| 2018-08-09 Ben Richards <benton_richards@apple.com> |
| |
| Add ability to ignore process prewarming for launch time benchmark |
| https://bugs.webkit.org/show_bug.cgi?id=188462 |
| |
| Reviewed by Ryosuke Niwa. |
| |
| Added a flag to the new tab benchmark that will open a blank tab before the tab to be measured in order to ignore process prewarming. |
| |
| * LaunchTime/launch_time.py: |
| (DefaultLaunchTimeHandler): |
| (DefaultLaunchTimeHandler.on_receive_stop_time): Deleted. |
| (DefaultLaunchTimeHandler.on_receive_stop_signal): |
| (DefaultLaunchTimeHandler.do_HEAD): |
| (DefaultLaunchTimeHandler.do_GET): |
| (DefaultLaunchTimeHandler.do_POST): |
| (LaunchTimeBenchmark): |
| (LaunchTimeBenchmark._standard_deviation): Fixed divide by zero bug when '-n' is set to 1 |
| (LaunchTimeBenchmark.open_tab): Added option to open a blank tab |
| (LaunchTimeBenchmark.run): |
| * LaunchTime/new_tab.py: |
| (NewTabBenchmark.initialize): |
| (NewTabBenchmark.run_iteration): |
| (NewTabBenchmark.will_parse_arguments): |
| (NewTabBenchmark.did_parse_arguments): |
| (NewTabBenchmark.ResponseHandler.Handler.get_test_page): |
| (NewTabBenchmark.ResponseHandler.Handler.on_receive_stop_time): Deleted. |
| (NewTabBenchmark.ResponseHandler.Handler.on_receive_stop_signal): |
| (NewTabBenchmark): |
| * LaunchTime/startup.py: |
| (StartupBenchmark.ResponseHandler.Handler.get_test_page): |
| (StartupBenchmark.ResponseHandler.Handler.on_receive_stop_time): Deleted. |
| (StartupBenchmark.ResponseHandler.Handler.on_receive_stop_signal): |
| (StartupBenchmark): |
| |
| 2018-07-26 Andy VanWagoner <andy@vanwagoner.family> |
| |
| [INTL] Remove INTL sub-feature compile flags |
| https://bugs.webkit.org/show_bug.cgi?id=188081 |
| |
| Reviewed by Michael Catanzaro. |
| |
| Removed old and unused i18n flag. |
| |
| * StitchMarker/wtf/FeatureDefines.h: |
| |
| 2018-07-20 David Fenton <david_fenton@apple.com> |
| |
| REGRESSION (234006) ERROR: Blocked access to external URL http://code.jquery.com/jquery-1.9.1.min.js |
| https://bugs.webkit.org/show_bug.cgi?id=187858 |
| |
| Unreviewed test gardening, skipped test. |
| |
| * Skipped: LaunchTime |
| |
| 2018-07-19 Ben Richards <benton_richards@apple.com> |
| |
| Add benchmark for WebKit process launch times |
| https://bugs.webkit.org/show_bug.cgi?id=186414 |
| |
| Reviewed by Ryosuke Niwa. |
| |
| Added two benchmarks, one for measuring browser new tab launch time and one for browser startup time. |
| |
| * LaunchTime/.gitignore: Added. |
| * LaunchTime/feedback_client.html: Added. |
| Displays benchmark progress in browser |
| * LaunchTime/feedback_server.py: Added. |
| (FeedbackServer): Sends data to feedback_client via websocket |
| (FeedbackServer.__init__): |
| (FeedbackServer._create_app): |
| (FeedbackServer._start_server): |
| (FeedbackServer._send_all_messages): |
| (FeedbackServer.start): |
| (FeedbackServer.stop): |
| (FeedbackServer.send_message): Send a message to the feedback_client |
| (FeedbackServer.wait_until_client_has_loaded): Wait until the feedback_client has opened a websocket connection to the feedback_server |
| (FeedbackServer.MainHandler): Handler factory to create handler that serves feedback_client.html |
| (FeedbackServer.MainHandler.Handler): |
| (FeedbackServer.MainHandler.Handler.get): |
| (FeedbackServer.WSHandler): Handler factory to create handler that sends data to feedback client |
| (FeedbackServer.WSHandler.Handler): |
| (FeedbackServer.WSHandler.Handler.open): On websocket connection opened |
| (FeedbackServer.WSHandler.Handler.on_close): On websocket connection closed |
| * LaunchTime/launch_time.py: Added. |
| (DefaultLaunchTimeHandler): Abstract HTTP request handler for launch time benchmarks |
| (DefaultLaunchTimeHandler.get_test_page): Default test page to be overridden by benchmarks |
| (DefaultLaunchTimeHandler.get_blank_page): |
| (DefaultLaunchTimeHandler.on_receive_stop_time): |
| (DefaultLaunchTimeHandler.do_HEAD): |
| (DefaultLaunchTimeHandler.do_GET): |
| (DefaultLaunchTimeHandler.do_POST): |
| (DefaultLaunchTimeHandler.log_message): Suppresses HTTP logs from SimpleHTTPRequestHandler |
| (LaunchTimeBenchmark): Abstract class which launch time benchmarks inherit from and override methods desired to customize |
| (LaunchTimeBenchmark.__init__): |
| (LaunchTimeBenchmark._parse_browser_bundle_path): Parser for bundle path option |
| (LaunchTimeBenchmark._parse_args): |
| (LaunchTimeBenchmark._run_server): Target function for main server thread |
| (LaunchTimeBenchmark._setup_servers): |
| (LaunchTimeBenchmark._clean_up): |
| (LaunchTimeBenchmark._exit_due_to_exception): |
| (LaunchTimeBenchmark._geometric_mean): |
| (LaunchTimeBenchmark._standard_deviation): |
| (LaunchTimeBenchmark._compute_results): Returns mean and std dev of list of results and pretty prints if should_print=True is specified |
| (LaunchTimeBenchmark._wait_times): Mimic numpy.linspace |
| (LaunchTimeBenchmark.open_tab): Open a browser tab with the html given by self.response_handler.get_test_page |
| (LaunchTimeBenchmark.launch_browser): Open a broser to either the feedback client (if option is set) or a blank page |
| (LaunchTimeBenchmark.quit_browser): |
| (LaunchTimeBenchmark.quit_browser.quit_app): |
| (LaunchTimeBenchmark.quit_browser.is_app_closed): |
| (LaunchTimeBenchmark.close_tab): |
| (LaunchTimeBenchmark.wait): |
| (LaunchTimeBenchmark.log): Print to console and send to feedback client if --feedback-in-browser flag is used |
| (LaunchTimeBenchmark.log_verbose): Only logs if --verbose flag is used |
| (LaunchTimeBenchmark.run): |
| (LaunchTimeBenchmark.group_init): Initialization done before each round of iterations |
| (LaunchTimeBenchmark.run_iteration): |
| (LaunchTimeBenchmark.initialize): Convenience method to be overriden by subclasses which is called at the end of __init__ |
| (LaunchTimeBenchmark.will_parse_arguments): Called before argparse.parse_args to let subclasses add new command line arguments |
| (LaunchTimeBenchmark.did_parse_arguments): Called after argparse.parse_args to let subclass initialize based on command line arguments |
| * LaunchTime/new_tab.py: Added |
| (NewTabBenchmark): |
| (NewTabBenchmark._parse_wait_time): Parser for wait time option |
| (NewTabBenchmark.initialize): |
| (NewTabBenchmark.run_iteration): |
| (NewTabBenchmark.group_init): |
| (NewTabBenchmark.will_parse_arguments): |
| (NewTabBenchmark.did_parse_arguments): |
| (NewTabBenchmark.ResponseHandler): |
| (NewTabBenchmark.ResponseHandler.Handler): |
| (NewTabBenchmark.ResponseHandler.Handler.get_test_page): |
| * LaunchTime/startup.py: Added |
| (StartupBenchmark): This benchmark measures browser startup time and initial page load time |
| (StartupBenchmark.initialize): |
| (StartupBenchmark.run_iteration): |
| (StartupBenchmark.ResponseHandler): |
| (StartupBenchmark.ResponseHandler.Handler): |
| (StartupBenchmark.ResponseHandler.Handler.get_test_page): |
| * LaunchTime/thirdparty/__init__.py: Added. |
| (AutoinstallImportHook): Auto installs tornado package for feedback server |
| (AutoinstallImportHook.__init__): |
| (AutoinstallImportHook._ensure_autoinstalled_dir_is_in_sys_path): |
| (AutoinstallImportHook.find_module): |
| (AutoinstallImportHook._install_tornado): |
| (AutoinstallImportHook.greater_than_equal_to_version): |
| (AutoinstallImportHook._install): |
| (AutoinstallImportHook.get_latest_pypi_url): |
| (AutoinstallImportHook.install_binary): |
| (autoinstall_everything): |
| (get_os_info): |
| |
| 2018-06-25 Jon Lee <jonlee@apple.com> |
| |
| [MotionMark] Add support for version numbers |
| https://bugs.webkit.org/show_bug.cgi?id=186479 |
| |
| Reviewed by Said Abou-Hallawa. |
| |
| Add support for displaying the version number as well as including it in the JSON results. |
| |
| When loading the front page, script replaces any element with classname version with the |
| version number of the benchmark, which is stored in Strings.version. |
| |
| The JSON structure for the results includes a new version property: |
| { |
| "version": "1.0", |
| "options": { ... }, |
| "data": [ ... ] |
| } |
| |
| When dragging a results file, the version listed will come from the JSON file. Older |
| results will not have had the version property, in which case it will default to "1.0". |
| |
| * MotionMark/index.html: Update title to some other default. Script will update it. |
| Include the version number in the logo title. |
| * MotionMark/developer.html: Ditto. |
| * MotionMark/about.html: Ditto. |
| |
| * MotionMark/resources/runner/motionmark.js: |
| (ResultsDashboard): Update constructor to include version. This is used when serializing |
| results out to JSON, and displaying the results panel in developer mode. |
| (ResultsDashboard._processData): When running the benchmark, include benchmark version string |
| in the results object. |
| (ResultsDashboard.version): |
| (window.benchmarkRunnerClient.willStartFirstIteration): When running the benchmark, pass the |
| benchmark version string to the dashboard, which holds the results. |
| (window.sectionsManager.setSectionVersion): Helper function to update the element in the |
| section with the class name version. |
| (window.benchmarkController.initialize): Populate all DOM elements with class name "version" |
| with the version string. Update the page title. |
| (window.benchmarkController.showResults): When showing results, update the version string |
| based on what is included in the JSON results, which would be the same as the benchmark version. |
| * MotionMark/resources/runner/motionmark.css: Include missing copyright. Wrap the SVG logo |
| in a div and include the version string. |
| * MotionMark/resources/strings.js: Add strings for the page title template, and the version. |
| |
| * MotionMark/resources/debug-runner/motionmark.css: |
| * MotionMark/resources/debug-runner/motionmark.js: |
| (window.benchmarkRunnerClient.willStartFirstIteration): When running the benchmark, pass the |
| benchmark version string to the dashboard, which holds the results. |
| (window.benchmarkController.initialize): Populate all DOM elements with class name "version" |
| with the version string. Update the page title. When dragging in JSON results, look for |
| version to pass to the dashboard. If it doesn't exist, default to "1.0". |
| (window.benchmarkController.showResults): When showing results, update the version string |
| based on what is included in the JSON results, instead of the current benchmark version. |
| * MotionMark/resources/debug-runner/tests.js: Update page title template. |
| |
| 2018-06-08 Jon Lee <jonlee@apple.com> |
| |
| [MotionMark] Rename Suits test files |
| https://bugs.webkit.org/show_bug.cgi?id=186447 |
| |
| Reviewed by Said Abou-Hallawa. |
| |
| * MotionMark/resources/runner/tests.js: |
| * MotionMark/tests/master/resources/suits.js: Renamed from PerformanceTests/MotionMark/tests/master/resources/svg-particles.js. |
| * MotionMark/tests/master/suits.html: Renamed from PerformanceTests/MotionMark/tests/master/svg-particles.html. |
| * MotionMark/tests/svg/suits.html: |
| |
| 2018-06-08 Jon Lee <jonlee@apple.com> |
| |
| Add sub-tests based on Suits |
| https://bugs.webkit.org/show_bug.cgi?id=186260 |
| |
| Reviewed by Said Abou-Hallawa. |
| |
| Add a new developer Suits suite with sub-tests that isolate parts of each particle. |
| The sub-tests are: |
| - Particles using only clip paths |
| - Particles using only shapes |
| - Particles that have no gradients |
| - Particles that have no rotation |
| - Particles that do not move around at all (but all of the physics calculations are |
| still performed) |
| |
| * MotionMark/resources/debug-runner/tests.js: |
| * MotionMark/tests/master/resources/svg-particles.js: Rename the stage, particle, and |
| benchmark to "Suits" from "SVG". Change SuitsParticle so that we guarantee 50% clip and |
| shape paths rather than relying on random chance. Wrap a check around creation of the |
| gradient element, and use a simple fill color when we aren't using a gradient. |
| * MotionMark/tests/svg/suits.html: Added. |
| * MotionMark/tests/svg/suits.js: Added. |
| Look for the query string and set the particle based on what is selected. |
| |
| 2018-06-06 Jon Lee <jonlee@apple.com> |
| |
| Allow the ramp controller to run tests that take less time than the initial ramp-up phase |
| https://bugs.webkit.org/show_bug.cgi?id=186257 |
| |
| Reviewed by Said Abou-Hallawa. |
| |
| * MotionMark/tests/resources/main.js: Extend the desired end timestamp when ramp-up occurs. |
| Previously the controller assumed that the duration of the test was always longer than the |
| time it took to figure out the order of magnitude of particles to render. When that range |
| is determined, the end timestamp is then extended. If the duration is too short, the controller |
| would throw an exception. The fix is to continue extending the timestamp every time the order of |
| magnitude tiers up. |
| |
| In general testing with really short durations isn't practical, but sometimes it is |
| desired for debugging. |
| |
| To avoid confusion of the private variables, I renamed Benchmark._startTimestamp to |
| Benchmark._benchmarkStartTimestamp, to disambiguate from Controller._startTimestamp. |
| Benchmark._startTimestamp is set once and never changed, after the initial 100ms warm-up. |
| Controller._startTimestamp is initialized to 0 (which means "not initialized"), and then after |
| the warm up is complete, set in Controller.start() to the same wall clock timestamp, and never |
| changed afterwards. |
| |
| 2018-06-06 Jon Lee <jonlee@apple.com> |
| |
| Remove unneeded data processing in MotionMark |
| https://bugs.webkit.org/show_bug.cgi?id=186256 |
| |
| Reviewed by Said Abou-Hallawa. |
| |
| Remove calculation of average frame rate based on multiple samples at a given complexity. |
| It is not used in the calculation of the score, and is just extra work to do at the end |
| of a test. |
| |
| * MotionMark/developer.html: Remove options to show average data and regression based on |
| that data. |
| * MotionMark/resources/debug-runner/graph.js: Remove data. Update the time graph that |
| shows the relationship between the break point in the regression against the complexity |
| graph. |
| * MotionMark/resources/debug-runner/motionmark.css: Make the confidence intervals easier |
| to see. |
| * MotionMark/resources/runner/motionmark.js: Simplify score calculation now that we are |
| only calculating regressions on the raw data. We were calculating it based on the averages, |
| but that was just for analysis, and not used in the actual score. Get rid of the |
| |complexitySamples| local variable. |
| * MotionMark/resources/strings.js: Remove unneeded string constant. |
| * MotionMark/tests/resources/main.js: |
| (processSamples): Refactor so that more of it can be shared between Controller and |
| RampController. It also avoids processing the raw samples (controllerSamples) twice for |
| the ramp controller. |
| (_processComplexitySamples): Remove logic to calculate averages based on complexity. |
| (_processMarks): Split this out from the old processSamples. |
| (_processControllerSamples): Split this out from the old processSamples. |
| |
| 2018-06-06 Jon Lee <jonlee@apple.com> |
| |
| Strip out more characters when creating permalinks |
| https://bugs.webkit.org/show_bug.cgi?id=186259 |
| |
| Reviewed by Said Abou-Hallawa. |
| |
| Tests and suites may have names that make permalinks to the tests less readable when |
| shared. Expand the set of characters that are stripped out of the names. |
| |
| * MotionMark/resources/debug-runner/motionmark.js: Refactor to call |
| stripUnwantedCharactersForURL. For existing links, decode the suite and test names. |
| * MotionMark/resources/extensions.js: |
| (Utilities.stripUnwantedCharactersForURL): Rename from stripNonASCIICharacters. |
| Remove any non-alphanumeric character. |
| |
| 2018-06-04 Jon Lee <jonlee@apple.com> |
| |
| Remove unnecessary MotionMark controllers |
| https://bugs.webkit.org/show_bug.cgi?id=186255 |
| |
| Reviewed by Said Abou-Hallawa. |
| |
| Get rid of the ramp30 and the fixed-with-a-step controllers, which aren't used for testing |
| or debugging. |
| |
| * MotionMark/developer.html: Move the "ramp" option as first, since that is the default |
| controller for the benchmark. |
| * MotionMark/resources/debug-runner/motionmark.js: |
| * MotionMark/resources/runner/motionmark.js: Added missing copyright notice. |
| * MotionMark/tests/resources/main.js: |
| |
| 2018-05-25 Myles C. Maxfield <mmaxfield@apple.com> |
| |
| Improve the performance of Font::canRenderCombiningCharacterSequence() |
| https://bugs.webkit.org/show_bug.cgi?id=185933 |
| |
| Reviewed by Ryosuke Niwa. |
| |
| * Layout/ComplexLongUnique.html: Added. |
| |
| 2018-05-25 Saam Barati <sbarati@apple.com> |
| |
| Have a memory test where we can validate JSCs mini memory mode |
| https://bugs.webkit.org/show_bug.cgi?id=185932 |
| |
| Reviewed by Mark Lam. |
| |
| We add a directory here with the contents of the testmem benchmark. |
| To run it, use `Tools/Scripts/run-testmem`. To add new tests in the future, |
| you just need to add JS files to this directory. |
| |
| * testmem: Added. |
| * testmem/air.js: Added. |
| * testmem/base64.js: Added. |
| * testmem/basic.js: Added. |
| * testmem/box2d.js: Added. |
| * testmem/crypto-md5.js: Added. |
| * testmem/date-format-tofte.js: Added. |
| * testmem/earley-boyer.js: Added. |
| * testmem/hash-map.js: Added. |
| * testmem/regex-dna.js: Added. |
| * testmem/splay.js: Added. |
| * testmem/tagcloud.js: Added. |
| |
| 2018-05-22 Ryan Haddad <ryanhaddad@apple.com> |
| |
| Unreviewed, rolling out r232052. |
| |
| Breaks internal builds. |
| |
| Reverted changeset: |
| |
| "Use more C++17" |
| https://bugs.webkit.org/show_bug.cgi?id=185176 |
| https://trac.webkit.org/changeset/232052 |
| |
| 2018-05-21 Yusuke Suzuki <utatane.tea@gmail.com> |
| |
| Use more C++17 |
| https://bugs.webkit.org/show_bug.cgi?id=185176 |
| |
| Reviewed by JF Bastien. |
| |
| * MallocBench/MallocBench.xcodeproj/project.pbxproj: |
| * MediaTime/Configurations/Base.xcconfig: |
| * MediaTime/MediaTime.xcodeproj/project.pbxproj: |
| |
| 2018-04-30 Myles C. Maxfield <mmaxfield@apple.com> |
| |
| Improve the performance of FontCascadeDescription's effectiveFamilies |
| https://bugs.webkit.org/show_bug.cgi?id=184720 |
| |
| Reviewed by Simon Fraser. |
| |
| This performance test calls SystemFontDatabase::systemFontCascadeList() around 2,000,000 times (before |
| this patch is applied), which is roughly equivalent to the page we found the performance problem on. |
| The calling pattern is roughly equivalent in this test. |
| |
| * Layout/system-ui.html: Added. |
| |
| 2018-03-08 Antti Koivisto <antti@apple.com> |
| |
| Update StyleBench version number in page title to 0.3 |
| |
| * StyleBench/index.html: |
| |
| There has been a bunch of changes, most notably the addition of attribute selector test step. |
| |
| 2018-02-16 Dean Jackson <dino@apple.com> |
| |
| Use OPENGL macros to be more clear about which OpenGL/ES WebGL uses on Cocoa |
| https://bugs.webkit.org/show_bug.cgi?id=182894 |
| |
| Reviewed by Tim Horton. |
| |
| Rename OPENGL_ES_2 to OPENGL_ES. |
| |
| * StitchMarker/wtf/Platform.h: |
| |
| 2018-02-09 Ryosuke Niwa <rniwa@webkit.org> |
| |
| Make run-perf-tests work with Speedometer 2.0 and re-enable the test |
| https://bugs.webkit.org/show_bug.cgi?id=182088 |
| |
| Reviewed by Antti Koivisto. |
| |
| This patch updates the test harness to report new metrics of Speedometer 2.0. |
| |
| Replaced the total time by the geometric mean of the time spent in each suite, and added scores |
| as a metric in addition to time. Also, updated the PerfTestRunner harness and the results template |
| to support "Score" metric with "pt" as its unit. |
| |
| * Skipped: Unskipped the test. |
| * Speedometer/resources/benchmark-report.js: |
| (window.benchmarkClient.willStartFirstIteration..createTest): Added the unit as an optional argument |
| to report the score in addition to time. Also report the geomean of time instead of the total time. |
| (window.benchmarkClient.willStartFirstIteration): Report the geomean of time. |
| (window.benchmarkClient.didRunSuites): Record the entire measuredValues instead of just tests. |
| (window.benchmarkClient.didFinishLastIteration): Report the scores. |
| |
| 2018-02-09 Ross Kirsling <ross.kirsling@sony.com> |
| |
| Use REQUEST_ANIMATION_FRAME_DISPLAY_MONITOR on any non-Windows port. |
| https://bugs.webkit.org/show_bug.cgi?id=182623 |
| |
| Reviewed by Alex Christensen. |
| |
| * StitchMarker/wtf/Platform.h: |
| Simplify #if. |
| |
| 2018-02-01 Antti Koivisto <antti@apple.com> |
| |
| StyleBench: Attribute selectors and other improvements |
| https://bugs.webkit.org/show_bug.cgi?id=182387 |
| |
| Reviewed by Joseph Pecoraro. |
| |
| - Add some attributes to elements in all tests |
| - Add some attribute selectors to stylesheets in all tests |
| - Also add some * selectors to all stylesheets. |
| - Add attribute mutation step to all suites |
| - Make test steps do more mutations (25->100) and reduce the number of steps to keep testing time in check. |
| Too fast steps were running into timer resolution limits. |
| |
| * StyleBench/resources/style-bench.js: |
| (defaultConfiguration): |
| (prototype.randomAttributeName): |
| (prototype.randomAttributeValue): |
| (prototype.randomAttributeSelector): |
| (prototype.makeCompoundSelector): |
| (prototype.makeElement): |
| (prototype.addClasses): |
| (prototype.removeClasses): |
| (prototype.mutateAttributes): |
| (prototype.async.runForever): |
| * StyleBench/resources/tests.js: |
| (makeSteps): |
| (makeSuite): |
| |
| 2018-02-01 Geoffrey Garen <ggaren@apple.com> |
| |
| Make MallocBench easier for non-WebKit engineers to run |
| https://bugs.webkit.org/show_bug.cgi?id=182415 |
| |
| Reviewed by Saam Barati. |
| |
| * MallocBench/MallocBench.xcodeproj/project.pbxproj: Use c++14 so we |
| can make_unique. |
| |
| Specify that we support all Darwin platforms so you can test them. |
| |
| * MallocBench/run-malloc-benchmarks: Specify the path to MallocBench |
| and libmbmalloc explicitly, rather than computing them implicitly |
| using webkitdirs. Non-WebKit folks don't have the directory structure |
| required by webkitdirs. |
| |
| Remove Linux-specific and cmake-specific behaviors because we only |
| needed them in the world of implicit path computation. |
| |
| 2018-01-31 Ryosuke Niwa <rniwa@webkit.org> |
| |
| Make run-perf-tests work with StyleBench and re-enable it |
| https://bugs.webkit.org/show_bug.cgi?id=182322 |
| |
| Reviewed by Antti Koivisto. |
| |
| This patch updates the test harness to report new metrics of StyleBench. |
| |
| Replaced the total time by the geometric mean of the time spent in each suite, and added scores |
| as a metric in addition to time. Also, updated the PerfTestRunner harness and the results template |
| to support "Score" metric with "pt" as its unit. |
| |
| * Skipped: Unskipped the test. |
| * StyleBench/resources/benchmark-report.js: |
| (window.benchmarkClient.willStartFirstIteration..createTest): Added the unit as an optional argument |
| to report the score in addition to time. Also report the geomean of time instead of the total time. |
| (window.benchmarkClient.willStartFirstIteration): Report the geomean of time. |
| (window.benchmarkClient.didRunSuites): Record the entire measuredValues instead of just tests. |
| (window.benchmarkClient.didFinishLastIteration): Report the scores. |
| * resources/results-template.html: Added "Score" as a metric and "pt" as its unit. |
| * resources/runner.js: |
| (PerfTestRunner..finish): Ditto. |
| |
| 2018-01-30 Ryosuke Niwa <rniwa@webkit.org> |
| |
| StyleBench: Generate more mutation steps |
| https://bugs.webkit.org/show_bug.cgi?id=182291 |
| |
| Rubber-stamped by Antti Koivisto |
| |
| Use zero-based index instead of one-based index. |
| |
| * StyleBench/resources/tests.js: |
| (makeSteps): |
| |
| 2018-01-30 Antti Koivisto <antti@apple.com> |
| |
| StyleBench: Generate more mutation steps |
| https://bugs.webkit.org/show_bug.cgi?id=182291 |
| |
| Reviewed by Ryosuke Niwa. |
| |
| The test should be doing multiple rounds of mutations in a single document to test what is supposed to test (like it did before r227756). |
| This also makes test results more stable and less dependent on possible first-time costs. |
| |
| * StyleBench/resources/benchmark-runner.js: |
| |
| Adjust correctionFactor to match the new steps count. With 1.5 Firefox is ~72rpm. |
| |
| * StyleBench/resources/tests.js: |
| (makeSteps): |
| |
| Generate 10 rounds of class/element mutation steps. |
| Give each generated step unique name so reporting works. |
| Make individual steps do fewer mutations. |
| |
| 2018-01-29 Antti Koivisto <antti@apple.com> |
| |
| StyleBench: Remove : from a test name |
| https://bugs.webkit.org/show_bug.cgi?id=182258 |
| |
| Reviewed by Ryosuke Niwa. |
| |
| It breaks reporting. |
| |
| * StyleBench/resources/style-bench.js: |
| (nthPseudoClassConfiguration): |
| |
| 2018-01-29 Ryosuke Niwa <rniwa@webkit.org> |
| |
| Temporarily skip StyleBench until webkit.org/b/182088 is fixed. |
| |
| * Skipped: |
| |
| 2018-01-29 Ryosuke Niwa <rniwa@webkit.org> |
| |
| Make StyleBench compatible with run-benchmark and run-perf-tests |
| https://bugs.webkit.org/show_bug.cgi?id=182262 |
| |
| Reviewed by Antti Koivisto. |
| |
| Copied resource files referenced from Speedometer directory since run-benchmark needs to be able |
| to checkout each benchmark separately. |
| |
| Removed the code to create tests of the same name five times in makeSteps since this |
| won't be compatible with either run-benchmark or run-perf-tests. |
| |
| * StyleBench/index.html: Removed the code to show warnings for local files since run-benchmark |
| doesn't use HTTP server in WebDriver mode. |
| * StyleBench/resources/benchmark-report.js: Copied from resources/benchmark-report.js. |
| * StyleBench/resources/benchmark-runner.js: Copied from resources/benchmark-runner.js. |
| (BenchmarkRunner.prototype._finalize): Use the correction factor of 8 instead of 5 in StyleBench. |
| * StyleBench/resources/gauge.png: Copied from resources/gauge.png. |
| * StyleBench/resources/gauge@2x.png: Copied from resources/gauge@2x.png. |
| * StyleBench/resources/main.css: Copied from resources/main.css. |
| * StyleBench/resources/main.js: Copied from resources/main.js. |
| * StyleBench/resources/tests.js: |
| (makeSteps): Only make each test once. |
| |
| 2018-01-26 Chris Nardi <cnardi@chromium.org> |
| |
| Addressing post-review comments after r226614 |
| https://bugs.webkit.org/show_bug.cgi?id=182151 |
| |
| Reviewed by Myles C. Maxfield. |
| |
| * StitchMarker/wtf/text/StringImpl.h: |
| (WTF::isSpaceOrNewline): |
| * StitchMarker/wtf/text/TextBreakIterator.cpp: |
| (WTF::numCodeUnitsInGraphemeClusters): |
| * StitchMarker/wtf/text/TextBreakIterator.h: |
| |
| 2018-01-22 Antti Koivisto <antti@apple.com> |
| |
| StyleBench: Separate test for :nth pseudo classes |
| https://bugs.webkit.org/show_bug.cgi?id=181932 |
| |
| Reviewed by Sam Weinig. |
| |
| Optimization strategies for nth are different from other structural pseudo classes (:first-child etc). Move them to a separate test. |
| |
| * StyleBench/resources/style-bench.js: |
| (structuralPseudoClassConfiguration): |
| (nthPseudoClassConfiguration): |
| |
| Add a new configuration. |
| |
| (predefinedConfigurations): |
| (pseudoClassConfiguration): Deleted. |
| |
| 2018-01-17 Ryosuke Niwa <rniwa@webkit.org> |
| |
| Temporarily skip Speedometer until we fix the metrics. |
| |
| * Skipped: |
| |
| 2018-01-16 Alex Christensen <achristensen@webkit.org> |
| |
| Fix Mac CMake build after r222900 |
| |
| * MallocBench/MallocBench/CMakeLists.txt: |
| The mbmalloc target doesn't exist if we're not in developer mode, so only use it if we are. |
| |
| 2018-01-09 Ryosuke Niwa <rniwa@webkit.org> |
| |
| Speedometer 2.0: Update the About page. |
| https://bugs.webkit.org/show_bug.cgi?id=175918 |
| |
| Reviewed by Keith Miller. |
| |
| Updated the about page for Speedometer 2.0. |
| |
| * Speedometer/index.html: |
| |
| == Rolled over to ChangeLog-2018-01-01 == |
