When i tried to run on of my lengthy kernel in CUDA it gave me following warning but compiled.
Warning: Olimit was exceeded on function _Z15interpolateP6float3PsPiii; will not perform function-scope optimization.
What is this Olimit? is this due to out of register case? due to bulky kernel code?
what are the ways to over come this or minimize this ?
and what happens when the function-scope optimization is not done, will it hamper the overall performance of my kernel?
i would be very grateful if someone can give me insights about this.