I have been looking into it more and I think the way to get maximum texture upload performance (and do it async) is to use pixel buffer objects.
See: http://www.opengl.org/registry/specs/ARB/pixel_buffer_object.txt http://developer.download.nvidia.com/SDK/9.5/Samples/samples.html#TexturePerformancePBO I have seen some references online claiming multi-GB/sec upload rates with PBOs. The spec does a great job of explaining it, but at the core, it lets you: - Allocate a DMA memory buffer from the driver - Start that data uploading to the GPU - Later check (fences) if the buffer has been loaded - Bind a texture to the buffer and have it happen immediately on the GPU So the question is, how could we use something like this in OpenSG? In the near term, I would be satisfied with a way to hack it in, but I think a long term solution would also be very good. Maybe something that reuses the same buffers as a way to shuttle data to the GPU from memory. Any ideas? -Allen ------------------------------------------------------------------------------ Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA -OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise -Strategies to boost innovation and cut costs with open source participation -Receive a $600 discount off the registration fee with the source code: SFAD http://p.sf.net/sfu/XcvMzF8H _______________________________________________ Opensg-users mailing list Opensg-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/opensg-users
