History | Log In     View a printable version of the current page.  
Issue Details (XML | Word | Printable)

Key: LPP-714
Type: Bug Bug
Status: Verified Verified
Resolution: Fixed
Priority: P0 P0
Assignee: Max Carlson
Reporter: Amy Muntz
Votes: 0
Watchers: 0
Operations

If you were logged in you would be able to see more operations.
OpenLaszlo

Increased CPU usage

Created: 21/Sep/05 10:27 AM   Updated: 26/Jun/06 09:36 PM
Component/s: Components - all
Affects Version/s: 3.1 (aka Denver)
Fix Version/s: 3.1 (aka Denver)

Time Tracking:
Not Specified

Severity: Blocker
Runtime: N/A
Release Note Text: CPU usage is decreased during idle.


 Description  « Hide
Henry: Going back to changeset #32357, the window_example.lzx app is using much less CPU (1% or less), so I'm going to try to track down which changeset caused the jump in idle loop consumption of CPU.

On 9/19/05, Henry wrote:

    I believe I am seeing high CPU usage, around 40% from the Windows task manager, with Emerald and firefox. Not 80% though.

    The example
    http://localhost:8080/lps-dev/examples/components/window_example.lzx?lzt=html
    seems to hover around 10-12% CPU usage.

    I'm going to try going back a couple of weeks in lps-dev and see if this amount of
    CPU usage has gone up.

pablo writes:
I've seen CPU usage go up to 60-80% with emerald. I was
running Flash 7.


 All   Comments   Work Log   Change History      Sort Order: Ascending order - Click to sort in descending order
Amy Muntz - 21/Sep/05 01:58 PM
From email bcummins, 9/21/05:
I am seeing the CPU spike in today's QA build of Laszlo Mail (http://emerald.laszlosystems.com/lzmailqa-latest/). Once the login dialog is completely loaded I am seeing CPU usage of 33%+. After login I am seeing CPU usage of 50%+.
 
Running Windows XP, SP2. P4 3.4 GHz with 1 GB of RAM.
 
This application is built in solo mode, with debug set to false.

Henry Minsky - 22/Sep/05 08:18 AM
I tracked this down to changeset 32502

The new view tracker code in LzView.prototype.checkPlayStatus2 = function (){

seems to get called for *every* view, and keeps getting called every frame for
every view!

Henry Minsky - 22/Sep/05 08:20 AM
Shouldn't the play status code only get called for views which have swf movie resources? I'm not sure but it seems like it might be getting called for even views without resources. Maybe not. But checkPlayStatus2 is getting called every frame for every view, from what I can see with a print statement.

Max Carlson - 22/Sep/05 02:01 PM
so I isolated it down to views with resources in modaldialog that don't have stop frames in them, so they're always tracking. it seems our tracking mechanism is too good!

I think it's doing the 'correct' thing, but the 'correct' thing is expensive. Perhaps we should only turn play tracking on for runtime loaded reosurces?

Max Carlson - 23/Sep/05 11:06 AM
Fixed with change 32854