The goal of the effort is to develop the X-point included Gyrokinetic Code advanced simulation software for utilizing extreme parallelism on leadership computing resources. Magnetic fusion experiments are essential for next-generation burning plasma experiments such as the International Thermonuclear Experimental Reactor. The calculation of the drag and diffusion coefficients of the nonlinear collision operation was highly compute intensive and more expensive than the electron push kernel, taking over 70% of the total compute time. Performance portability on the collision operator has proven to be far more achievable than for the electron push kernel. The algorithm for pushing electrons has a deeply nested call graph and needs access to derived type data structures in modules. Both the optimized push and the multi-species collision operator are currently used in production runs but the pace of development has been rapid enough that they are newer than any large scaling tests.