Re: [Openev-discuss] Performance Tuning

[Frank W. <war...@ho...>]

Peter Farris-Manning wrote:
> 
> This message was sent from Geocrawler.com by "Peter Farris-Manning" <pe...@at...>
> 
> Your decisions w.r.t. redraw time and GvTrackTool seem reasonable. The discussion of texture
> cache trashing is interesting. I think our memory requirements are getting a bit high for low end
> systems. The memory requirements you state for expected display modes add up very quickly.
> I think we should attempt your suggested approach of trying to detect when we are in a texture
> thrashing situation, and temporarily bump up the texture cache size for the rest of the session.
> It would also be valuable to use greyscale textures instead of RGBA wherever possible. What
> level of effort would be required to implement both of these suggestions?

Peter,

The alteration to use greyscale textures for all applicable modes of complex
data would likely take on order of 1/2 day; however, since use of complex 
image data is likely to be rare I am not convinced it is a good return on 
time.  Are you sure you want it?

The detecting of thrashing is likely on order of a 1/2 day as well, and I am
not sure how safely I can detect thrashing.  It still might end up being
somewhat unhelpful, but I would like to give it a try. 

By the way, work is going well on numpy integration. 

Best regards,

---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, war...@ho...
light and sound - activate the windows | http://members.home.com/warmerda
and watch the world go round - Rush    | Geospatial Programmer for Rent