This paper studies the impact of the architecture features of the Cray-1 and the Cray X-MP and related compiler optimizations on machine performance. We develop a methodology for evaluating the effectiveness of the Cray Fortran compilers in coping with the architecture features and limitations of the Cray-1 and the Cray X-MP. As examples, the effects of vector register reservation and vector index misalignment on the performance of Livermore Fortran Kernels (LFKs) are presented. The causes of the performance differences of two Cray Fortran compilers, CFT1.14 and CFT77.13, on the vectorized LFKs are described and some areas for further improvement are suggested.