RESOLVED FIXED Bug 231458
Vectorize EqualPowerPanner::pan()
https://bugs.webkit.org/show_bug.cgi?id=231458
Summary Vectorize EqualPowerPanner::pan()
Chris Dumez
Reported 2021-10-08 15:19:20 PDT
Vectorize EqualPowerPanner::pan() for performance. This is used by PannerNode.
Attachments
Patch (3.35 KB, patch)
2021-10-08 15:20 PDT, Chris Dumez
no flags
Chris Dumez
Comment 1 2021-10-08 15:20:23 PDT
EWS
Comment 2 2021-10-08 17:41:34 PDT
Committed r283855 (242733@main): <https://commits.webkit.org/242733@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 440682 [details].
Radar WebKit Bug Importer
Comment 3 2021-10-08 17:42:18 PDT
Arcady Goldmints-Orlov
Comment 4 2021-11-19 14:01:00 PST
This caused a regression in the imported/w3c/web-platform-tests/webaudio/the-audio-api/the-pannernode-interface/panner-equalpower.html test on GTK and WPE, with the following diff: --- /home/buildbot/worker/gtk-linux-64-release-tests/build/layout-test-results/imported/w3c/web-platform-tests/webaudio/the-audio-api/the-pannernode-interface/panner-equalpower-expected.txt +++ /home/buildbot/worker/gtk-linux-64-release-tests/build/layout-test-results/imported/w3c/web-platform-tests/webaudio/the-audio-api/the-pannernode-interface/panner-equalpower-actual.txt @@ -14,7 +14,13 @@ PASS Mono: Left and right channels is identical to the array [expected array]. PASS < [mono source=listener] All assertions passed. (total 1 assertions) PASS > [stereo source=listener] Source and listener at the same position -PASS Stereo: Left and right channels is identical to the array [expected array]. -PASS < [stereo source=listener] All assertions passed. (total 1 assertions) -PASS # AUDIT TASK RUNNER FINISHED: 3 tasks ran successfully. +FAIL X Stereo: Left and right channels expected to be equal to the array [expected array] but differs in 8017 places: + Index Actual Expected + [129] -9.094678e-1 -9.937367e-1 + [130] -8.316806e-1 -9.996188e-1 + [131] -7.479767e-1 -9.983897e-1 + [132] -6.589519e-1 -9.900582e-1 + ...and 8013 more errors. assert_true: expected true got false +FAIL < [stereo source=listener] 1 out of 1 assertions were failed. assert_true: expected true got false +FAIL # AUDIT TASK RUNNER FINISHED: 1 out of 3 tasks were failed. assert_true: expected true got false
Chris Dumez
Comment 5 2021-11-19 14:32:17 PST
(In reply to Arcady Goldmints-Orlov from comment #4) > This caused a regression in the > imported/w3c/web-platform-tests/webaudio/the-audio-api/the-pannernode- > interface/panner-equalpower.html test on GTK and WPE, with the following > diff: > > --- > /home/buildbot/worker/gtk-linux-64-release-tests/build/layout-test-results/ > imported/w3c/web-platform-tests/webaudio/the-audio-api/the-pannernode- > interface/panner-equalpower-expected.txt > +++ > /home/buildbot/worker/gtk-linux-64-release-tests/build/layout-test-results/ > imported/w3c/web-platform-tests/webaudio/the-audio-api/the-pannernode- > interface/panner-equalpower-actual.txt > @@ -14,7 +14,13 @@ > PASS Mono: Left and right channels is identical to the array [expected > array]. > PASS < [mono source=listener] All assertions passed. (total 1 assertions) > PASS > [stereo source=listener] Source and listener at the same position > -PASS Stereo: Left and right channels is identical to the array [expected > array]. > -PASS < [stereo source=listener] All assertions passed. (total 1 assertions) > -PASS # AUDIT TASK RUNNER FINISHED: 3 tasks ran successfully. > +FAIL X Stereo: Left and right channels expected to be equal to the array > [expected array] but differs in 8017 places: > + Index Actual Expected > + [129] -9.094678e-1 -9.937367e-1 > + [130] -8.316806e-1 -9.996188e-1 > + [131] -7.479767e-1 -9.983897e-1 > + [132] -6.589519e-1 -9.900582e-1 > + ...and 8013 more errors. assert_true: expected true got false > +FAIL < [stereo source=listener] 1 out of 1 assertions were failed. > assert_true: expected true got false > +FAIL # AUDIT TASK RUNNER FINISHED: 1 out of 3 tasks were failed. > assert_true: expected true got false My bet is that the implementation of VectorMath::multiplyByScalarThenAddToVector() or VectorMath::multiplyByScalar() used by Glib ports is wrong somehow. The Apple ports are using a different implementation for these.
Note You need to log in before you can comment on or make changes to this bug.