Difference between revisions of "PGO Pref data here"

From CDOT Wiki
Jump to: navigation, search
(Web Page Load Times)
(Web Page Load Times)
Line 1: Line 1:
 
= Web Page Load Times =
 
= Web Page Load Times =
 +
 +
== Browser Comparison ==
 +
 +
<em>Oct 11th, 2008</em><br />
 +
Load Times for opening Google Maps in [http://www.webwait.com Webwait.com]
 +
 +
{|
 +
!style="background:#FFCC00;"|Firefox 3.0.3
 +
!style="background:#FFCC00;"|Minefield PGO (1)
 +
!style="background:#FFCC00;"|Minefield PGO (2)
 +
|-align="right"
 +
|2.10
 +
|1.38
 +
|1.45
 +
|-align="right"
 +
|1.85
 +
|1.24
 +
|1.27
 +
|-align="right"
 +
|2.19
 +
|1.14
 +
|1.30
 +
|-align="right"
 +
|2.19
 +
|1.06
 +
|1.02
 +
|-align="right"
 +
|2.12
 +
|1.10
 +
|0.96
 +
|-align="right"
 +
|2.24
 +
|1.07
 +
|0.97
 +
|-align="right"
 +
|2.32
 +
|1.20
 +
|1.14
 +
|-align="right"
 +
|2.29
 +
|1.15
 +
|1.29
 +
|-align="right"
 +
|2.06
 +
|1.27
 +
|1.16
 +
|-align="right"
 +
|2.13
 +
|1.16
 +
|1.13
 +
|-align="right"
 +
|style="border-top:solid"|2.15 (avg)
 +
|style="border-top:solid"|1.18 (avg)
 +
|style="border-top:solid"|1.17 (avg)
 +
|}
 +
 +
I saved Google Maps to my local server and when loaded from the local server the load times were the same for the 3 browsers.
 +
 +
(1)Profiled using SunSpider JS Test, Google Maps, a bunch of XUL examples and random sites.
 +
(2)Profiled with the opening and closing of the browser.
 +
 +
<br />
 +
 +
== Initial Tests ==
  
 
<strong>Testing Procedure:</strong><br />
 
<strong>Testing Procedure:</strong><br />
Line 7: Line 71:
 
I did the acid3 test site, the SunSpider JS Test, used google maps for a couple of searches, went to cnn.com and the top ten websites from www.web100.com.  I also visted a bunch of xul web applications and examples.
 
I did the acid3 test site, the SunSpider JS Test, used google maps for a couple of searches, went to cnn.com and the top ten websites from www.web100.com.  I also visted a bunch of xul web applications and examples.
  
== Google Maps ==
+
=== Google Maps ===
  
 
<em>Oct 10th, 2008</em>
 
<em>Oct 10th, 2008</em>
Line 25: Line 89:
 
</table>
 
</table>
  
== CNN.com ==
+
=== CNN.com ===
  
 
<em>Oct 10th, 2008</em><br />
 
<em>Oct 10th, 2008</em><br />
Line 46: Line 110:
 
</table>
 
</table>
  
== tomshardware.com ==
+
=== tomshardware.com ===
  
 
<em>Oct 10th, 2008</em><br />
 
<em>Oct 10th, 2008</em><br />

Revision as of 12:26, 11 October 2008

Web Page Load Times

Browser Comparison

Oct 11th, 2008
Load Times for opening Google Maps in Webwait.com

Firefox 3.0.3 Minefield PGO (1) Minefield PGO (2)
2.10 1.38 1.45
1.85 1.24 1.27
2.19 1.14 1.30
2.19 1.06 1.02
2.12 1.10 0.96
2.24 1.07 0.97
2.32 1.20 1.14
2.29 1.15 1.29
2.06 1.27 1.16
2.13 1.16 1.13
2.15 (avg) 1.18 (avg) 1.17 (avg)

I saved Google Maps to my local server and when loaded from the local server the load times were the same for the 3 browsers.

(1)Profiled using SunSpider JS Test, Google Maps, a bunch of XUL examples and random sites. (2)Profiled with the opening and closing of the browser.


Initial Tests

Testing Procedure:
Between each page load I went to a default page like www.google.com and clear private data, check all the boxes. I used the same profile for both Firefox 3.0.3 and Minefield.

Profile Data:
I did the acid3 test site, the SunSpider JS Test, used google maps for a couple of searches, went to cnn.com and the top ten websites from www.web100.com. I also visted a bunch of xul web applications and examples.

Google Maps

Oct 10th, 2008

Firefox 3.0.3Minefield with PGO
2.091.35
2.081.19
2.251.11
2.321.06
2.101.15
2.061.25
2.241.10
2.071.25
2.121.16
2.14 (avg)1.18 (avg)
45% faster

CNN.com

Oct 10th, 2008

This page was served from a local webserver.

Firefox 3.0.3Minefield with PGO
0.750.70
0.890.65
0.750.70
0.740.55
0.680.53
0.780.52
0.760.58
0.690.48
0.690.67
0.750.52
0.748 (avg)0.59 (avg)
21.% faster

tomshardware.com

Oct 10th, 2008
This page was served from a local webserver.

I did not profile this page and strangely enough my load times averaged to be the same. I would assume if I profiled a couple of webpages that would be enough to optimize most webpages.