[PyQt] Performance deterioration when using integer spin boxes PS ...

[Preisig, Heinz A]

OK, problem has been removed, though I do not know what the problem was. 
The issue was that MatPlotLib did not go together with the rest and 
simply got the whole program to a stand-still. I have not had the time 
to check what really was the problem, but it seemed obvious that a long 
queue of events is somewhere in the background generated.

Exchanging the plotting with a PyQwt implementation that is a little 
less luxurious, the problem disappeared. Performance was actually 
significantly improved also when not using any other event-driven parts 
of the interface. Thus looks like PyQwt is recommendable.

Heinz Preisig

On 09/22/2010 05:06 PM, Preisig, Heinz A wrote:
> PS: playing more it is not only the spin box. Other actions taken on 
> the interface slow it down too. Thus it is not a problem with the spin 
> boxes, but must be more generic. Only noticed the problem there first.
>
> On 09/22/2010 04:48 PM, Preisig, Heinz A wrote:
>> Certainly:
>>
>> I am using a timer to generate time events. On each time event I am 
>> doing a computation, in this case a data acquisition (DA converter) 
>> and then a computation the result of which is sent to a DA converter. 
>> The interface has a couple of spin boxes and sliders and buttons to 
>> manipulate what is happening. In the application I am using now, I 
>> replaced the process interface, which actually goes and receives and 
>> sends data to a serial port, by a simple simulation, which 
>> essentially does nothing.
>>
>> What I observe is that when the task executes nicely in the time grid 
>> (1 s). Once I am changing things in a spin box, the task does not 
>> make it to catch the next time event and it keeps on slowing down the 
>> more I am manipulating. I presume there are lots of events being 
>> generated in the background or passed on to other levels.
>>
>> I should mention that this is not the first application of this type. 
>> We wrote one last year and it worked just fine. We used it 
>> extensively at the time on very slow machines. The one I am working 
>> on now is a much faster machine and it runs the newer version of 
>> python and pyqt. On the slow one it gets bugged down tremendously.
>>
>> Not sure what the problem is....
>>
>> Heinz
>>
>> On 09/22/2010 04:04 PM, Hans-Peter Jansen wrote:
>>> On Wednesday 22 September 2010, 15:45:15 Preisig, Heinz A wrote:
>>>    
>>>> I observed that when using integer spin boxes in a timed loop the
>>>> performance drops significantly after changing a value.
>>>>      
>>> Heinz, could you elaborate on "using integer spin boxes in a timed loop" a
>>> bit, please? I don't grasp the concept of it ATM.
>>>
>>> Pete
>>>
>>>    
>>>> After dxchanging
>>>> them by double spin boxes no deterioration was observed.
>>>>
>>>> Heinz Preisig
>>>> _______________________________________________
>>>> PyQt mailing listPyQt at riverbankcomputing.com
>>>> http://www.riverbankcomputing.com/mailman/listinfo/pyqt
>>>>      
>>> _______________________________________________
>>> PyQt mailing listPyQt at riverbankcomputing.com
>>> http://www.riverbankcomputing.com/mailman/listinfo/pyqt
>>>    
>>
>> -- 
>>
>>
>> Heinz A Preisig
>> Professor of Process Systems Engineering 	Private:
>> Øvre Bakklandet 62 B, 7013 Trondheim, Norway
>> Department of Chemical Engineering
>> Norwegian University of Science and Technology
>> N -- 7491 Trondheim, Norway 	Tel direct: +47 735 92807
>> Tel mob: +47 9754 1334
>> e-mail: Heinz.Preisig at chemeng.ntnu.no 
>> <mailto:Heinz.Preisig at chemeng.ntnu.no>
>> web: www.chemeng.ntnu.no\~preisig 
>> <http://www.chemeng.ntnu.no/%7Epreisig>
>>
>>
>>
>> _______________________________________________
>> PyQt mailing listPyQt at riverbankcomputing.com
>> http://www.riverbankcomputing.com/mailman/listinfo/pyqt
>
> -- 
>
>
> Heinz A Preisig
> Professor of Process Systems Engineering 	Private:
> Øvre Bakklandet 62 B, 7013 Trondheim, Norway
> Department of Chemical Engineering
> Norwegian University of Science and Technology
> N -- 7491 Trondheim, Norway 	Tel direct: +47 735 92807
> Tel mob: +47 9754 1334
> e-mail: Heinz.Preisig at chemeng.ntnu.no 
> <mailto:Heinz.Preisig at chemeng.ntnu.no>
> web: www.chemeng.ntnu.no\~preisig <http://www.chemeng.ntnu.no/%7Epreisig>
>
>
>
> _______________________________________________
> PyQt mailing list    PyQt at riverbankcomputing.com
> http://www.riverbankcomputing.com/mailman/listinfo/pyqt

-- 


Heinz A Preisig
Professor of Process Systems Engineering 	Private:
Øvre Bakklandet 62 B, 7013 Trondheim, Norway
Department of Chemical Engineering
Norwegian University of Science and Technology
N -- 7491 Trondheim, Norway 	Tel direct: +47 735 92807
Tel mob: +47 9754 1334
e-mail: Heinz.Preisig at chemeng.ntnu.no <mailto:Heinz.Preisig at chemeng.ntnu.no>
web: www.chemeng.ntnu.no\~preisig <http://www.chemeng.ntnu.no/%7Epreisig>


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.riverbankcomputing.com/pipermail/pyqt/attachments/20100924/3a37a467/attachment-0001.html>