Data Transformations for Eliminating

Many cache misses in scientic programs are due to con-icts caused by limited set associativity. We examine two compile-time data-layout transformations for eliminating con-ict misses, concentrating on misses occuring on every loop iteration. Inter-variable padding adjusts variable base addresses , while intra-variable padding modiies array dimension sizes. Two levels of precision are evaluated. PadLite only uses array and column dimension sizes, relying on assumptions about common array reference patterns. Pad analyzes programs, detecting connict misses by linearizing array references and calculating conict distances between uniformly-generated references. The Euclidean algorithm for computing the gcd of two numbers is used to predict conicts between dierent array columns for linear algebra codes. Experiments on a range of programs indicate PadLite can eliminate conicts for benchmarks, but Pad is more eeective over a range of cache and problem sizes. Padding reduces cache miss rates by 16% on average for a 16K direct-mapped cache. Execution times are reduced by 6% on average, with some SPEC95 programs improving up to 15%.