Create our OGL context on the same thread in the OpenGL backend. Same issue with Qualcomm not working with threading correctly.