I wanted to ask if there are any performance hits or other flaws with GL_ARB_transpose_matrix I should be worried about? I have a lot of C code written to work with [4][4] style arrays so this extension would be used extensively throughout a lot of code. I’d rather not have my own slow routines to convert matrices between different formats only to feed them to OpenGL.
Just wanted to make sure before I mess around with this. Thanks.
The performance hit of converting the matrices should be not noticable at all. You can easily use this extension, should be no problems. Probably, the driver will convert the matrices anyway, but who cares…
For all you know, the hardware might store it internally in row major order, meaning that a call to the NON-transpose functions is what incurs the shuffling overhead…
In any case, the overhead is going to be negligable
