[mapserver-dev] OpenGL Contribution

[Paul Ramsey]

My main curiosity is: how much faster is it? For certain workloads,
rendering is getting up there are an important part of the draw time,
but not for all workloads. And are there workloads where OpenGL is
worse? Rendering imagery, for example, where the textures would have
to be loaded-and-flushed, loaded-and-flushed from the graphics
hardware, since they never get re-used?

The reason to add a new rendering chain is because it adds something
the others lack. AGG added quality lacking in GD. If the lack isn't
large enough, the rendering chain falls into disrepair (see, PDF,
which while providing better output than render'n'stuff'image
approaches to PDF generation, hasn't received much love). What brought
the original developers to deciding to do an OpenGL chain, and what
metrics for improved speed have they gathered? Inquiring minds want to
know.

P.

On Fri, Dec 12, 2008 at 12:01 PM, Steve Lime <Steve.Lime at dnr.state.mn.us> wrote:
> Hi all: I wanted to bring an offline conversation I've been having to the -dev list for further consideration.
> Toby Rahilly and Jonathan Bourke have developed OpenGL support for MapServer and would like to contribute
> it to the main code base. I've seen output from their work and it looks nice. One big benefit is performance
> under the right circumstances. Seems like a very interesting addition to MapServer.
>
> I've attached a draft RFC they put together. If folks think this is something worthwhile pursuing then I'd
> propose creating a branch in svn for them to work in. Other ideas welcome.
>
> Steve
>
>
> _______________________________________________
> mapserver-dev mailing list
> mapserver-dev at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/mapserver-dev
>
>