Does R memory management slow down vs using Matlab and NVIDIA Fermi Tesla/CUDA/C++?

Does R memory management slow down vs using Matlab and NVIDIA Fermi Tesla/CUDA/C++?