Removing the virtual functions - removing the calls - makes measurable performance progression on CSS. We measured the following performance progressions on some very popular websites, for example: msn.com: 33% yandex.ru: 3,3% The number of variables created during the test. m_isPrimitiveValue 261596 The max number of variables in the memory during the test. m_isPrimitiveValue 468
Created attachment 101453 [details] Switch isPrimitiveValue virtual function to inline
Comment on attachment 101453 [details] Switch isPrimitiveValue virtual function to inline Attachment 101453 [details] did not pass chromium-ews (chromium-xvfb): Output: http://queues.webkit.org/results/9200235 New failing tests: css2.1/t1202-counter-16-f.html css2.1/t1202-counters-03-b.html css2.1/t1202-counter-12-b.html css2.1/t1202-counter-08-b.html css2.1/t1202-counter-02-b.html css2.1/t1202-counter-14-b.html css2.1/t1202-counter-00-b.html css2.1/t1202-counter-09-b.html css2.1/t1202-counter-15-b.html css2.1/t1202-counter-05-b.html css2.1/t1202-counter-07-b.html css2.1/t1202-counter-06-b.html css2.1/t1202-counter-01-b.html css2.1/t1202-counter-03-b.html css2.1/t1202-counters-02-b.html css2.1/t1202-counter-04-b.html css2.1/t1202-counters-01-b.html css2.1/t1202-counter-13-b.html css2.1/t1202-counter-11-b.html css2.1/t1202-counters-00-b.html
See my comments in https://bugs.webkit.org/show_bug.cgi?id=64262.
Comment on attachment 101453 [details] Switch isPrimitiveValue virtual function to inline How much of a memory regression is this?
I believe there are some wasted bits that could be put to good use for a perf enhancement without increasing the memory overhead. Please give a chance to upload an RFC patch and some repeatable numbers.
The action requested in this bug was taken in: http://trac.webkit.org/changeset/99468 *** This bug has been marked as a duplicate of bug 71668 ***