This summer, PhD student Brandon Neth kicked off the International Conference on Supercomputing (ICS-2021) with a talk on their paper “RAJALC: Inter-loop Optimizations in RAJA.” Authored in collaboration with Professor Michelle Strout (Brandon's advisor) and Drs. Tom Scogland and Bronis de Supinski of Lawrence Livermore National Laboratory, the paper details RAJALC, an extension to the performance portability library RAJA. While RAJA is a powerful tool for writing portable kernels for high-performance computing applications, it leaves performance on the table by only considering kernels in isolation. RAJALC remedies this issue by introducing portable, easy to use optimizations that apply across multiple kernels, like loop fusion. Then, aided by RAJALC’s runtime symbolic evaluation, the safety of the requested optimization is assured. Using RAJALC, developers can achieve nearly the same performance improvements while writing up to 90% less code.