Hello,
I wondered if the following vbo useage idea is fast or slow:
1 static vertex buffer
2 dynamic index buffers
feed index buffer
drawcall
swap index buffer & feed it
drawcall
or would one need to have multiple vertex buffers as well ? (I think not)
thanks for your time
Index buffer will work on whatever vertex buffer is bound. So, yes, this would work.
mh i can’t imagine a case in which you have the same vertex positions but different indices… what do you need this for?
the latter is the case, mostly for batching many small objects. so same indices wouldnt be used.
thanks for the info
on a side not, does anyone know if using shorts for normals/texcoords was accelerated on radeon 8500 ? currently I have a fat vertex 64 byte, and a small one with 32 byte (no normals), but I could need normals in the 32byte one too. and our target hardware is older vs1.0 hardware mostly