Re: [Jmol-developers] sluggishness & CPU usage

[Nicolas V. <nve...@cl...>]

Miguel wrote:
>> Miguel wrote:
>>     
>>> I will see how difficult it will be for me to revert back to shorts in
>>> order to understand what is going on ... not that going back to shorts is
>>> an option.
>>>       
>
> This is more on Nico's problem than Bob's problem ...
>
> I just did a very unscientific test ...
>
> On my linux box
> using current trunk r5050
> using the Jmol application
> using caffeine.xyz
> using a window size of about 400x400
> using 'spin on'
> using 'top' to measure CPU consumption
>
> CPU percentage seemed to be about 37%
>
> I rolled the zbuffer back from 'int' to 'short'
>
> The CPU percentage seemed to be about 35%
>
> I am satisfied that the switch short->int did not introduce a noticable
> performance problem ... certainly nothing that would account for a 3 fold
> CPU consumption increase.
>
> I am going to check in the code and then immediately revert it ... so that
> we will have the 'diff' to reapply it again in the future if we need to.
>
> Nico, if you want to try this on Windows you can ... the Windows video
> system is generally faster ... I suppose it is possible that there could
> be some difference, but I don't think so.
>   
Ok, I will try to get some time this evening to make a few tests with 
the application.
> Nico, we should see if we can reproduce this problem in the Jmol
> application ... comparing 10.00.xx with 10.2.0 and the current trunk.
>   
Yes, I will try with extracting a few SVN revisions.

I have to go to work now, will try this evening.

Nico