[Webkit-unassigned] [Bug 158083] LLInt should support other types of prototype GetById caching.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=158083

--- Comment #33 from Caio Lima <ticaiolima at gmail.com> ---
Checking the benchmark I noticed that the patch does not improve the results. The cases that I paid attention were:

#JSC with the patch
get-by-id-proto-or-self                        11.6561+-2.5266

#JSC without the patch
get-by-id-proto-or-self                        13.2753+-2.3457    
--------------------------------------------------------------
#JSC with the patch
get-by-id-self-or-proto                        13.0930+-1.7589

#JSC without the patch
get-by-id-self-or-proto                        12.1011+-0.7154    
--------------------------------------------------------------
#JSC with the patch
getter-prototype                                5.8231+-0.6807

#JSC without the patch
getter-prototype                                5.5745+-0.2236    
--------------------------------------------------------------
#JSC with the patch
prototype-access-with-mutating-prototype        6.2859+-2.0017

#JSC without the patch
prototype-access-with-mutating-prototype        5.3830+-0.4957
--------------------------------------------------------------

It is important notice the errors presented by the JSC run with the patch are always higher than the run without patch. IMHO it is because of the prototype cache maintenance and setup overhead. Also, It is important consider that this is implementing just the slow_path to access the prototype accessor functions.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20160606/26403522/attachment.html>