THE GLAF SOFTWARE DEVELOPED AT VIRGINIA TECH PROVIDES AN INTUITIVE VISUAL PROGRAMMING INTERFACE WITH ANAUTO-PARALLELIZING AND AUTO-TUNING FRAMEWORK THAT AUTOMATICALLY GENERATES EFFICIENT SERIAL AND PARALLELFORTRAN AND C CODE. THIS APPROACH PROVIDES A BALANCE BETWEEN PERFORMANCE AND PROGRAMMABILITY. THE CURRENT VERSION OF GLAF TARGETS SHARED-MEMORY PARALLEL COMPUTERS NAMELY SINGLE MULTI-CORE NODES SUCH AS INTEL XEONS (E.G. BROADWELL HASWELL ETC.) XEON PHI CO-PROCESSORS AND NVIDIA GPUS (ACCELERATORS). THIS EFFORT EVALUATES GLAF VISUAL PROGRAMMING INTERFACE AND ITS AUTO-TUNING FRAMEWORK FOR SUITABILITY AND APPLICABILITY AS A PARALLEL PROGRAMMING APPROACH FOR APPLICATIONS IN BOTH THE CLIMATE SCIENCE BRANCH (E302)AND FROM THE RESEARCH DIRECTORATE (RD). NAMELY LARC WILL IDENTIFY REASONABLY SIZED SINGLE-NODEAPPLICATIONS OR STANDALONE CODE SEGMENTS (I.E. COMPUTATIONAL KERNELS OR MINI-APPLICATIONS) AND VATECH ANDLARC WILL DEMONSTRATE AND CREATE SIMILAR GLAF REPRESENTATIONS. THIS WILL EXERCISE AND DEMONSTRATE THE USEOF THE GUI TO REPRESENT THE PROBLEM DOMAIN (IN SERIAL AND PARALLEL) DEMONSTRATE THE CODE GENERATION AND COMPILATION AND EVALUATE THE PERFORMANCE BETWEEN THE ORIGINAL AND THE GLAF GENERATED CODE ON CPUS XEON PHIS AND NVIDIA GPUS. GLAF = GRID-BASED LANGUAGE AND AUTO-PARALLELIZING AUTO-TUNING FRAMEWORK