Loading ...
Sorry, an error occurred while loading the content.

Re: [wmlprogramming] Measuring JavaScript performance

Expand Messages
  • Julio Rabadán González
    Samsung i9000 Galaxy S (with Android 2.3.3 & stock browser) : 5927ms It obtained a similar result with Dolphin browser. ... -- Julio Rabadán González
    Message 1 of 16 , Jun 1, 2011
    View Source
    • 0 Attachment
      Samsung i9000 Galaxy S (with Android 2.3.3 & stock browser) : 5927ms
      It obtained a similar result with Dolphin browser.

      El 01/06/2011 3:54, Luca Passani escribió:
      Thanks guys. We have learned a lot with this. Anyway, we realised that we also 
      made a mistake in the test which invalidates the test to some extent (but not 
      totally. Not need to rush right now, albeit we could use it if someone re-runned 
      the test for our reference).
      
      http://t.wurfl.com
      
      For us, the big question at this point is: how are JS performance and 
      performance of animations related?  of course, we suspect that they are related, 
      but having a clearer indication of what that relation actually is, would help a lot.
      
      Comments?
      
      Luca
      
      
      On 31/05/2011 20.56, Anders Magnus Andersen wrote:
      
      
      I also tested some devices:
      
      *Nokia C7 with Symbian Anna (new browser)*
      *Result:* 49928ms
      
      *Samsung Galaxy S II*
      *Result:* 17724ms
      
      *Samsung Galaxy S*
      *Result:* 22136
      
      *iPod Touch 4 gen iOS 4*
      *Result:* 35584ms
      
      -A
      
      2011/5/31 Yes WAP <yeswap@... <mailto:yeswap@...>>
      
          T-Mobile MyTouch 3G running CyanogenMod 5.0.8
          user-agent: Mozilla/5.0 (Linux; U; Android 2.1-update1; en-us; HTC
          Magic Build/EPE54B) AppleWebKit/530.17 (KHTML, like Gecko) Version/4.0
          Mobile Safari/530.17
          263946ms
      
          Opera Mobile 11 on T-Mobile MyTouch 3G running CyanogenMod 5.0.8
          user-agent: Opera/9.80 (Android 2.1-update1; Linux; Opera
          Mobi/ADR-1105231029; U; en) Presto/2.7.81 Version/11.00
          137256ms
      
          Samsung Wave S8500L
          user-agent: Mozilla/5.0 (SAMSUNG; SAMSUNG-GT-S8500L/1.0; U; Bada/1.0;
          es) AppleWebKit/533.1 (KHTML, like Gecko) Dolfin/2.0 Mobile WVGA
          SMM-MMS/1.2.0 OPN-B
          177269ms
      
          Palm Pixi Plus (Verizon)
          user-agent: Mozilla/5.0 (SAMSUNG; SAMSUNG-GT-S8500L/1.0; U; Bada/1.0;
          es) AppleWebKit/533.1 (KHTML, like Gecko) Dolfin/2.0 Mobile WVGA
          SMM-MMS/1.2.0 OPN-B
          138279ms
      
          Nokia N8 Symbian browser
          user-agent: Mozilla/5.0 (Symbian/3; Series60/5.2 NokiaN8-00/014.002;
          Profile/MIDP-2.1 Configuration/CLDC-1.1 ) AppleWebKit/525 (KHTML, like
          Gecko) Version/3.0 BrowserNG/7.2.8.10 <http://7.2.8.10> 3gpp-gba
          Test did not complete, hung with progress bar at about 8%, waited 20 minutes
      
          Nokia N8 Opera Mobile 11
          user-agent: Opera/9.80 (S60; SymbOS; Opera Mobi/SYB-1103211396; U;
          en-US) Presto/2.7.81 Version/11.00
          51997ms
      
      
      
          On Mon, May 30, 2011 at 2:34 PM, Luca Passani <luca.passani@...
          <mailto:luca.passani%40gmail.com>> wrote:
          >
          > People, with reference to the recent discussions on this and another mailing
          > list about how to decide whether jQuery Mobile should be delivered to a
          device
          > or not (wr actual usability), Steve and I have been doing some thinking
          and have
          > come up with a JavaScript test which will measure actual JS performance.
          >
          > Do not get us wrong, the test "engine" is the SunSpider benchmark from the
          > WebKit project, but we adjusted it slightly to make it more mobile friendly.
          >
          > Can you visit the WURFL test suite:
          >
          > http://t.wurfl.com
          >
          > and click on the link at the top (Javascript Performance Index). Once there,
          > click on the [Start Test] button and report the following:
          >
          > 1) how many milliseconds the test took (the result will be provided on
          the screen)
          >
          > 2) make and model (and UA string if possible)
          >
          > 3) browser
          >
          > No need to go there with legacy browsers/devices, but WebKit based
          browsers are
          > OK, as well as other advanced browsers. At this stage, we are particularly
          > interested in:
          >
          > - low-end Androids (ZTE, Huawei,...)
          >
          > - BlackBerries (RIM OS 5 and 6)
          >
          > - WindowsPhone 7
          >
          > - iPhones of course
          >
          > Of course, if this test proves to be a valid way to assert real JavaScript
          > horsepower, we may turn this into a capability which you guys can use to
          > multiserve your JQM, minimizr or whatever libraries based on *your*
          assessment
          > of what horsepower is good enough.
          >
          > Luca
          >
          >
          >
          > ------------------------------------
      
          >
          > As of July 14 2005, it's much easier to be banned from WMLProgramming!
          > Please fail to read http://tech.groups.yahoo.com/group/wmlprogramming/
          before you post.Yahoo! Groups Links
          >
          >
          >
          >
      
      
      
      
      -- 
      mvh
      Anders Magnus Andersen
      +46 707 96 1967
      http://www.mobiletech.se
      
      
      
      
      
      
      
      ------------------------------------
      
      As of July 14 2005, it's much easier to be banned from WMLProgramming!
      Please fail to read http://tech.groups.yahoo.com/group/wmlprogramming/ before you post.Yahoo! Groups Links
      
      <*> To visit your group on the web, go to:
          http://groups.yahoo.com/group/wmlprogramming/
      
      <*> Your email settings:
          Individual Email | Traditional
      
      <*> To change settings online go to:
          http://groups.yahoo.com/group/wmlprogramming/join
          (Yahoo! ID required)
      
      <*> To change settings via email:
          wmlprogramming-digest@yahoogroups.com 
          wmlprogramming-fullfeatured@yahoogroups.com
      
      <*> To unsubscribe from this group, send an email to:
          wmlprogramming-unsubscribe@yahoogroups.com
      
      <*> Your use of Yahoo! Groups is subject to:
          http://docs.yahoo.com/info/terms/
      
      
      

      --
      Julio Rabadán González
      Somms.NET
    • bruno.verachten@laposte.net
      Hi, here are the results for the new version: 1) 77725ms 2) HTC Hero: Mozilla/5.0 (Linux; U; Android 2.1- update1; fr-fr; HTC Hero Build/ERE27)
      Message 2 of 16 , Jun 1, 2011
      View Source
      • 0 Attachment
        Hi,

        here are the results for the new version:
        1) 77725ms
        2) HTC Hero: Mozilla/5.0 (Linux; U; Android 2.1- update1; fr-fr; HTC Hero Build/ERE27) AppleWebKit/530.17 (KHTML, like Gecko) Version/4.0 Mobile Safari/530.17
        3) Mobile Safari?+

        Kind regards,

        Bruno Verachten
      • Luca Passani
        ... thank you. We are also observing that the browser seems to account for less than 10% of the result (assuming JS support is powerful enough to complete the
        Message 3 of 16 , Jun 1, 2011
        View Source
        • 0 Attachment
          On 01/06/2011 9.53, Julio Rabadán González wrote:
          > Samsung i9000 Galaxy S (with Android 2.3.3 & stock browser) : 5927ms
          > It obtained a similar result with Dolphin browser.
          >

          thank you. We are also observing that the browser seems to account for less than
          10% of the result (assuming JS support is powerful enough to complete the test,
          that is :).

          Luca
        • belen barros pena
          Mozilla/5.0 (Linux; U; Android 2.3.3; en-gb; HTC_DesireHD_A9191 Build/GRI40) AppleWebKit/533.1 (KHTML, like Gecko) Version/4.0 Mobile Safari/533.1 Completes in
          Message 4 of 16 , Jun 2, 2011
          View Source
          • 0 Attachment
            Mozilla/5.0 (Linux; U; Android 2.3.3; en-gb; HTC_DesireHD_A9191
            Build/GRI40) AppleWebKit/533.1 (KHTML, like Gecko) Version/4.0 Mobile
            Safari/533.1

            Completes in 5939 ms

            On Wed, Jun 1, 2011 at 2:54 AM, Luca Passani <luca.passani@...> wrote:
            >
            > Thanks guys. We have learned a lot with this. Anyway, we realised that we also
            > made a mistake in the test which invalidates the test to some extent (but not
            > totally. Not need to rush right now, albeit we could use it if someone re-runned
            > the test for our reference).
            >
            > http://t.wurfl.com
            >
            > For us, the big question at this point is: how are JS performance and
            > performance of animations related?  of course, we suspect that they are related,
            > but having a clearer indication of what that relation actually is, would help a lot.
            >
            > Comments?
            >
            > Luca
            >
            >
            > On 31/05/2011 20.56, Anders Magnus Andersen wrote:
            >>
            >>
            >> I also tested some devices:
            >>
            >> *Nokia C7 with Symbian Anna (new browser)*
            >> *Result:* 49928ms
            >>
            >> *Samsung Galaxy S II*
            >> *Result:* 17724ms
            >>
            >> *Samsung Galaxy S*
            >> *Result:* 22136
            >>
            >> *iPod Touch 4 gen iOS 4*
            >> *Result:* 35584ms
            >>
            >> -A
            >>
            >> 2011/5/31 Yes WAP <yeswap@... <mailto:yeswap@...>>
            >>
            >>     T-Mobile MyTouch 3G running CyanogenMod 5.0.8
            >>     user-agent: Mozilla/5.0 (Linux; U; Android 2.1-update1; en-us; HTC
            >>     Magic Build/EPE54B) AppleWebKit/530.17 (KHTML, like Gecko) Version/4.0
            >>     Mobile Safari/530.17
            >>     263946ms
            >>
            >>     Opera Mobile 11 on T-Mobile MyTouch 3G running CyanogenMod 5.0.8
            >>     user-agent: Opera/9.80 (Android 2.1-update1; Linux; Opera
            >>     Mobi/ADR-1105231029; U; en) Presto/2.7.81 Version/11.00
            >>     137256ms
            >>
            >>     Samsung Wave S8500L
            >>     user-agent: Mozilla/5.0 (SAMSUNG; SAMSUNG-GT-S8500L/1.0; U; Bada/1.0;
            >>     es) AppleWebKit/533.1 (KHTML, like Gecko) Dolfin/2.0 Mobile WVGA
            >>     SMM-MMS/1.2.0 OPN-B
            >>     177269ms
            >>
            >>     Palm Pixi Plus (Verizon)
            >>     user-agent: Mozilla/5.0 (SAMSUNG; SAMSUNG-GT-S8500L/1.0; U; Bada/1.0;
            >>     es) AppleWebKit/533.1 (KHTML, like Gecko) Dolfin/2.0 Mobile WVGA
            >>     SMM-MMS/1.2.0 OPN-B
            >>     138279ms
            >>
            >>     Nokia N8 Symbian browser
            >>     user-agent: Mozilla/5.0 (Symbian/3; Series60/5.2 NokiaN8-00/014.002;
            >>     Profile/MIDP-2.1 Configuration/CLDC-1.1 ) AppleWebKit/525 (KHTML, like
            >>     Gecko) Version/3.0 BrowserNG/7.2.8.10 <http://7.2.8.10> 3gpp-gba
            >>     Test did not complete, hung with progress bar at about 8%, waited 20 minutes
            >>
            >>     Nokia N8 Opera Mobile 11
            >>     user-agent: Opera/9.80 (S60; SymbOS; Opera Mobi/SYB-1103211396; U;
            >>     en-US) Presto/2.7.81 Version/11.00
            >>     51997ms
            >>
            >>
            >>
            >>     On Mon, May 30, 2011 at 2:34 PM, Luca Passani <luca.passani@...
            >>     <mailto:luca.passani%40gmail.com>> wrote:
            >>     >
            >>     > People, with reference to the recent discussions on this and another mailing
            >>     > list about how to decide whether jQuery Mobile should be delivered to a
            >>     device
            >>     > or not (wr actual usability), Steve and I have been doing some thinking
            >>     and have
            >>     > come up with a JavaScript test which will measure actual JS performance.
            >>     >
            >>     > Do not get us wrong, the test "engine" is the SunSpider benchmark from the
            >>     > WebKit project, but we adjusted it slightly to make it more mobile friendly.
            >>     >
            >>     > Can you visit the WURFL test suite:
            >>     >
            >>     > http://t.wurfl.com
            >>     >
            >>     > and click on the link at the top (Javascript Performance Index). Once there,
            >>     > click on the [Start Test] button and report the following:
            >>     >
            >>     > 1) how many milliseconds the test took (the result will be provided on
            >>     the screen)
            >>     >
            >>     > 2) make and model (and UA string if possible)
            >>     >
            >>     > 3) browser
            >>     >
            >>     > No need to go there with legacy browsers/devices, but WebKit based
            >>     browsers are
            >>     > OK, as well as other advanced browsers. At this stage, we are particularly
            >>     > interested in:
            >>     >
            >>     > - low-end Androids (ZTE, Huawei,...)
            >>     >
            >>     > - BlackBerries (RIM OS 5 and 6)
            >>     >
            >>     > - WindowsPhone 7
            >>     >
            >>     > - iPhones of course
            >>     >
            >>     > Of course, if this test proves to be a valid way to assert real JavaScript
            >>     > horsepower, we may turn this into a capability which you guys can use to
            >>     > multiserve your JQM, minimizr or whatever libraries based on *your*
            >>     assessment
            >>     > of what horsepower is good enough.
            >>     >
            >>     > Luca
            >>     >
            >>     >
            >>     >
            >>     > ------------------------------------
            >>
            >>     >
            >>     > As of July 14 2005, it's much easier to be banned from WMLProgramming!
            >>     > Please fail to read http://tech.groups.yahoo.com/group/wmlprogramming/
            >>     before you post.Yahoo! Groups Links
            >>     >
            >>     >
            >>     >
            >>     >
            >>
            >>
            >>
            >>
            >> --
            >> mvh
            >> Anders Magnus Andersen
            >> +46 707 96 1967
            >> http://www.mobiletech.se
            >>
            >>
            >>
            >>
            >
            >
            >
            > ------------------------------------
            >
            > As of July 14 2005, it's much easier to be banned from WMLProgramming!
            > Please fail to read http://tech.groups.yahoo.com/group/wmlprogramming/ before you post.Yahoo! Groups Links
            >
            >
            >
            >
          Your message has been successfully submitted and would be delivered to recipients shortly.