Abstract
The memory consistency model underlying the Unified Parallel C (UPC) language remains a promising but underused feature. We report on our efforts to understand the UPC memory model and assess its potential benefits. We describe problems we have uncovered in the current language specification. These results have inspired an effort in the UPC community to create an alternative memory model definition that avoids these problems. We give experimental results confirming the promise of performance gains afforded by the memory model's relaxed constraints on consistency.