I've found an inconsistency between the cases where my input texture to
an osgCompute module is a texture with an osg::Image present and where
my texture is an FBO target. I'm using MAP_DEVICE.
In both cases the osgCuda code uses MallocPitch to create the buffer
where the texture data goes (alloc()), but in the Image case it uses
memcpy (not 2D, in setup()) to transfer the data. So, the buffer in the
image case does not honour the memory pitch, while in the FBO case it does.
I ?fixed? it using the modified file (src/osgCuda/Texture.cpp) as
attached and now I get consistent behaviour. Let me know what you think.
This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard.
The full disclaimer details can be found at http://www.csir.co.za/disclaimer.html.
This message has been scanned for viruses and dangerous content by MailScanner,
and is believed to be clean.
Post generated by Mail2Forum