637024 - IE9 Speedread Test: 4.0beta13pre nightly runs MUCH slower than IE9 RC and Chrome Canary 11 using i945GM graphics

Status: RESOLVED DUPLICATE of bug 637089

(Core :: Graphics, defect)

Description

[Shaun]

User-Agent:       Mozilla/5.0 (Windows NT 6.1; rv:2.0b13pre) Gecko/20110226 Firefox/4.0b13pre
Build Identifier: Mozilla/5.0 (Windows NT 6.1; rv:2.0b13pre) Gecko/20110226 Firefox/4.0b13pre

The latest nightly gives the following values, which are really slow:

FPS: 3
Average draw duration: 271 ms


In comparison, IE9 RC gives these values:
FPS: 60
Average draw duration: 16 ms

Chrome Canary 11:
FPS: 60
Average draw duration: 9 ms
(It has problems with high CPU usage on display of Speedread page, but that's for their developers to tackle)



Other data:

Total Billboard Draws - 
Fx 4.0beta13pre: 392
IE9 RC: 379
Chrome Canary 11: 390

Window Size - 
Fx 4.0beta13pre: 1400x881
IE9 RC: 1400x906
Chrome Canary 11: 1400x939

Reproducible: Always

Steps to Reproduce:
1. Open Firefox.
2. Copy and paste given URL.
3. Let the Speedreading Test complete.
4. Note down the average draw duration and fps values.
Actual Results:  
Very low fps and very high Average Draw Duration.

Expected Results:  
High fps (60 fps) and very low Average Draw Duration (< 15 ms).

Comment 1

[Shaun]

Failed to add in the initial report that my graphics card is on the blocklist. It is an Intel 945GM with custom drivers (Sherry v1.2 from angelictears.wordpress.com) updated beyond Intel's latest. IE9 and Chrome Canary don't have a blocklist for this slow but evidently operable chipset.

Comment 2

[Shaun]

Sorry for the URL typo, it is: https://angelictears9xxssf.wordpress.com/

Comment 3

[Bas Schouten (:bas.schouten)]

We blacklist significantly more aggressive than the other browsers in an attempt to make sure non of our users have a worse experience than on 3.6, this was a conscious decision but it bites us here. I'm not sure what we can do about it.

Comment 4

[Shaun]

I filed https://bugzilla.mozilla.org/show_bug.cgi?id=637089 to handle the uncertainty. :)

Comment 5

[Leman Bennett [Omega]]

i945GM has a max texture size of 2048x2048 making it a Dx9_2 graphics solution.

This would be blocked per the required texture size limit of 4096x4096 or Dx9_3 graphics solutions.

Comment 6

[Shaun]

The Release Candidate of IE 9 makes do with the limitations of the i945GM chipset. I am requesting parity-IE.

Comment 7

[Joe Drew (not getting mail)]

This bug is at best a duplicate of bug 637089.