Bug 46691

Summary: [EFL] Use fast malloc for WebKit EFL
Product: WebKit Reporter: Gyuyoung Kim <gyuyoung.kim>
Component: WebKit EFLAssignee: Nobody <webkit-unassigned>
Status: RESOLVED FIXED    
Severity: Normal CC: abarth, antognolli+webkit, commit-queue, kenneth, leandro, lucas.de.marchi, tonikitoo
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: PC   
OS: Linux   
Attachments:
Description Flags
Patch
none
Patch none

Gyuyoung Kim
Reported 2010-09-27 19:58:47 PDT
FastMalloc is to allocate memory quickly. I think we need to use fast malloc as well.
Attachments
Patch (2.29 KB, patch)
2010-09-27 22:23 PDT, Gyuyoung Kim
no flags
Patch (2.31 KB, patch)
2010-09-29 00:19 PDT, Gyuyoung Kim
no flags
Gyuyoung Kim
Comment 1 2010-09-27 22:23:00 PDT
Lucas De Marchi
Comment 2 2010-09-28 04:37:44 PDT
Looks good to me. However I wonder how fast it is. Does anyone have some benchmarks? Do other ports use it by default?
Gyuyoung Kim
Comment 3 2010-09-28 18:41:30 PDT
WebKit-GTK uses the fast malloc. However, android port, brewmp port and HAIKU port don't use it. I don't know if mac port, iphone use the fast malloc. According to QT port guys in seged univ, fast malloc based on TCMalloc is about 2 ~ 4% faster than system malloc. http://webkit.sed.hu/node/23 In addition, the fast malloc is good for multi-thread environment. Because, TCMalloc in the fast malloc has thread local cache per thread. http://goog-perftools.sourceforge.net/doc/tcmalloc.html
Zoltan Horvath
Comment 4 2010-09-29 00:05:52 PDT
Comment on attachment 69021 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=69021&action=review Qt, Mac, Gtk, etc. ports use TCmalloc. > cmake/OptionsEfl.cmake:57 > +WEBKIT_FEATURE(ENABLE_FAST_MALLOC "Enable fast memory allocation" DEFAULT ON) It would be better to reflect the switching between system's default allocator and TCmalloc. E.g. Enable TCmalloc instead of system's allocator
Gyuyoung Kim
Comment 5 2010-09-29 00:19:48 PDT
Created attachment 69166 [details] Patch Ok, I change the description according to your opinion. I also think that this description reflects correct information.
Gyuyoung Kim
Comment 6 2010-09-30 18:02:51 PDT
Hello Kenneth and Antonio, If you are not busy, could you please review this bug?
Adam Barth
Comment 7 2010-10-04 20:38:08 PDT
Comment on attachment 69166 [details] Patch Clearing flags on attachment: 69166 Committed r69060: <http://trac.webkit.org/changeset/69060>
Adam Barth
Comment 8 2010-10-04 20:38:14 PDT
All reviewed patches have been landed. Closing bug.
Note You need to log in before you can comment on or make changes to this bug.