<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
<!DOCTYPE bugzilla SYSTEM "https://bugs.webkit.org/page.cgi?id=bugzilla.dtd">

<bugzilla version="5.0.4.1"
          urlbase="https://bugs.webkit.org/"
          
          maintainer="admin@webkit.org"
>

    <bug>
          <bug_id>21500</bug_id>
          
          <creation_ts>2008-10-09 02:58:18 -0700</creation_ts>
          <short_desc>Qt-port on Linux should be built with g++ -O3 switch</short_desc>
          <delta_ts>2009-04-16 18:22:41 -0700</delta_ts>
          <reporter_accessible>1</reporter_accessible>
          <cclist_accessible>1</cclist_accessible>
          <classification_id>1</classification_id>
          <classification>Unclassified</classification>
          <product>WebKit</product>
          <component>JavaScriptCore</component>
          <version>528+ (Nightly build)</version>
          <rep_platform>PC</rep_platform>
          <op_sys>Linux</op_sys>
          <bug_status>RESOLVED</bug_status>
          <resolution>FIXED</resolution>
          
          
          <bug_file_loc></bug_file_loc>
          <status_whiteboard></status_whiteboard>
          <keywords></keywords>
          <priority>P2</priority>
          <bug_severity>Normal</bug_severity>
          <target_milestone>---</target_milestone>
          
          
          <everconfirmed>0</everconfirmed>
          <reporter name="Csaba Osztrogonác">ossy</reporter>
          <assigned_to name="Nobody">webkit-unassigned</assigned_to>
          <cc>hausmann</cc>
    
    <cc>laszlo.gombos</cc>
          

      

      

      

          <comment_sort_order>oldest_to_newest</comment_sort_order>  
          <long_desc isprivate="0" >
    <commentid>94731</commentid>
    <comment_count>0</comment_count>
    <who name="Csaba Osztrogonác">ossy</who>
    <bug_when>2008-10-09 02:58:18 -0700</bug_when>
    <thetext>WebKit Qt-port on Linux is built with g++ -O2 switch because of Qt default config (Qt-4.4.1/mkspecs/common/g++.conf) Building with -O3 switch instead of -O2, JSC is faster. (I used gcc version 4.1.2, Qt version 4.4.1 and WebKit rev37424)

runtime and code size of JSC:
- 3.51% speedup on SunSpider (+/-0.8%)
- 1.30% speedup on V8 (+/-0.65%)
- 3.81% speedup on WindScorpion (+/-1.0%)
- Code size of JSC .text section increased by 9.96%

I&apos;ll send patch soon.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>94732</commentid>
    <comment_count>1</comment_count>
    <who name="Csaba Osztrogonác">ossy</who>
    <bug_when>2008-10-09 05:11:36 -0700</bug_when>
    <thetext>Layout tests result - the whole WebKit built with -O3:
489.63s total testing time
7272 test cases (89%) succeeded
836 test cases (10%) had incorrect layout
3 test cases (&lt;1%) crashed

Layout tests result - only JSC built with -O3:
513.98s total testing time
7277 test cases (89%) succeeded
831 test cases (10%) had incorrect layout
3 test cases (&lt;1%) crashed

Layout tests result - original 37424 revision (with default -O2): 
531.12s total testing time
7272 test cases (89%) succeeded
836 test cases (10%) had incorrect layout
3 test cases (&lt;1%) crashed
</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>94733</commentid>
    <comment_count>2</comment_count>
      <attachid>24230</attachid>
    <who name="Csaba Osztrogonác">ossy</who>
    <bug_when>2008-10-09 05:44:13 -0700</bug_when>
    <thetext>Created attachment 24230
proposed patch</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>94744</commentid>
    <comment_count>3</comment_count>
    <who name="Alp Toker">alp</who>
    <bug_when>2008-10-09 09:25:30 -0700</bug_when>
    <thetext>(In reply to comment #2)
&gt; Created an attachment (id=24230) [edit]
&gt; proposed patch
&gt; 

It&apos;s incorrect to build WebCore with -O3 I believe (only JavaScriptCore should be -03).</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>94865</commentid>
    <comment_count>4</comment_count>
    <who name="Csaba Osztrogonác">ossy</who>
    <bug_when>2008-10-10 00:50:06 -0700</bug_when>
    <thetext>(In reply to comment #3)
&gt; It&apos;s incorrect to build WebCore with -O3 I believe (only JavaScriptCore should
&gt; be -03).

Why is it incorrect? Does building WebCore with -O3 cause any problem? It seems to work fine (and faster) here with -O3. Maybe on other platforms have a problem with -O3. A little code size growth can be bad for embedded systems, but not for PC&apos;s.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>96078</commentid>
    <comment_count>5</comment_count>
    <who name="Simon Hausmann">hausmann</who>
    <bug_when>2008-10-21 05:52:47 -0700</bug_when>
    <thetext>I agree with Alp here, only JavaScriptCore should be build with O3. Can you patch JavaScriptCore/JavaScriptCore.pri instead?</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>96728</commentid>
    <comment_count>6</comment_count>
    <who name="Csaba Osztrogonác">ossy</who>
    <bug_when>2008-10-27 02:23:31 -0700</bug_when>
    <thetext>(In reply to comment #5)
I will create patch for ToT and measure its speedup soon.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>96731</commentid>
    <comment_count>7</comment_count>
    <who name="Csaba Osztrogonác">ossy</who>
    <bug_when>2008-10-27 03:32:57 -0700</bug_when>
    <thetext>I&apos;ve created new patch for JavaScriptCore/JavaScriptCore.pri at r37894.

runtime and code size of JSC:
- 2.02% speedup on SunSpider (+/-0.8%)
- 1.10% speedup on V8 (+/-0.7%)
- 3.45% speedup on WindScorpion (+/-1.1%)
- Code size of JSC .text section increased by 9.70%</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>96732</commentid>
    <comment_count>8</comment_count>
      <attachid>24686</attachid>
    <who name="Csaba Osztrogonác">ossy</who>
    <bug_when>2008-10-27 03:35:52 -0700</bug_when>
    <thetext>Created attachment 24686
proposed patch</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>98610</commentid>
    <comment_count>9</comment_count>
    <who name="Csaba Osztrogonác">ossy</who>
    <bug_when>2008-11-13 05:52:47 -0800</bug_when>
    <thetext>new results for r38367:
- 1.41% speedup on SunSpider (+/-0.5%)
- 0.12% speedup on V8 (+/-0.3%) - not significant
- 6.33% speedup on WindScorpion (+/-0.3%)
- Code size of JSC .text section increased by 9.70%

There isn&apos;t activity a long time. Is anybody interested in landing this patch?
</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>98801</commentid>
    <comment_count>10</comment_count>
      <attachid>24686</attachid>
    <who name="Simon Hausmann">hausmann</who>
    <bug_when>2008-11-14 10:50:31 -0800</bug_when>
    <thetext>Comment on attachment 24686
proposed patch

Looks good. I&apos;ll also add -fstrict-aliasing when landing like we do in WebCore/WebCore.pro, so that we are on the safe side.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>98802</commentid>
    <comment_count>11</comment_count>
    <who name="Simon Hausmann">hausmann</who>
    <bug_when>2008-11-14 10:53:18 -0800</bug_when>
    <thetext>Nevermind the comment about aliasing, that was nonsense :-)

Landed your patch in r38398. Thanks!</thetext>
  </long_desc>
      
          <attachment
              isobsolete="1"
              ispatch="1"
              isprivate="0"
          >
            <attachid>24230</attachid>
            <date>2008-10-09 05:44:13 -0700</date>
            <delta_ts>2008-10-27 03:35:52 -0700</delta_ts>
            <desc>proposed patch</desc>
            <filename>O3.patch</filename>
            <type>text/plain</type>
            <size>949</size>
            <attacher name="Csaba Osztrogonác">ossy</attacher>
            
              <data encoding="base64">SW5kZXg6IENoYW5nZUxvZwo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09
PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBDaGFuZ2VMb2cJKHJldmlzaW9uIDM3NDQ1
KQorKysgQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBAIC0xLDMgKzEsMTYgQEAKKzIwMDgtMTAt
MDkgIENzYWJhIE9zenRyb2dvbmFjIDxvc3ppQGluZi51LXN6ZWdlZC5odT4KKworICAgICAgICBS
ZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBodHRwOi8vYnVncy53ZWJraXQu
b3JnL3Nob3dfYnVnLmNnaT9pZD0yMTUwMAorCQorICAgICAgICAzLjUxJSBzcGVlZHVwIG9uIFN1
blNwaWRlciAoUXQtcG9ydCBvbiBMaW51eCkKKyAgICAgICAgMS4zMCUgc3BlZWR1cCBvbiBWOCAo
UXQtcG9ydCBvbiBMaW51eCkKKworICAgICAgICA3LjgxJSBzcGVlZHVwIG9uIGxheW91dCB0ZXN0
cworCisgICAgICAgICogV2ViS2l0LnByaTogYnVpbGRpbmcgd2l0aCAtTzMgaW5zdGVhZCBvZiAt
TzIgKGdjYykKKwogMjAwOC0xMC0wNyAgQWxwIFRva2VyICA8YWxwQG51YW50aS5jb20+CiAKICAg
ICAgICAgR1RLKyBidWlsZCBmaXggZm9yIG9sZGVyIGF1dG9tYWtlIHZlcnNpb25zICgxLjcpLiBE
aXNjdXNzZWQgaW4gYnVnCkluZGV4OiBXZWJLaXQucHJpCj09PT09PT09PT09PT09PT09PT09PT09
PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFdlYktpdC5w
cmkJKHJldmlzaW9uIDM3NDI0KQorKysgV2ViS2l0LnByaQkod29ya2luZyBjb3B5KQpAQCAtODcs
MyArODcsNSBAQCBkZWZpbmVUZXN0KGFkZEV4dHJhQ29tcGlsZXJXaXRoSGVhZGVyKSB7CiAgICAg
cmV0dXJuKHRydWUpCiB9CiAKKyotZysrKjpRTUFLRV9DWFhGTEFHU19SRUxFQVNFIC09IC1PMgor
Ki1nKysqOlFNQUtFX0NYWEZMQUdTX1JFTEVBU0UgKz0gLU8zCg==
</data>
<flag name="review"
          id="11017"
          type_id="1"
          status="+"
          setter="timothy"
    />
          </attachment>
          <attachment
              isobsolete="0"
              ispatch="1"
              isprivate="0"
          >
            <attachid>24686</attachid>
            <date>2008-10-27 03:35:52 -0700</date>
            <delta_ts>2008-11-14 10:50:31 -0800</delta_ts>
            <desc>proposed patch</desc>
            <filename>O3.patch</filename>
            <type>text/plain</type>
            <size>1133</size>
            <attacher name="Csaba Osztrogonác">ossy</attacher>
            
              <data encoding="base64">SW5kZXg6IEphdmFTY3JpcHRDb3JlL0NoYW5nZUxvZwo9PT09PT09PT09PT09PT09PT09PT09PT09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=
</data>
<flag name="review"
          id="11281"
          type_id="1"
          status="+"
          setter="hausmann"
    />
          </attachment>
      

    </bug>

</bugzilla>