Combining Static and Dynamic Data Coalescing in Unified Parallel C

Abstract

Abstract-Significant progress has been made in the development of programming languages and tools that are suitable for hybrid computer architectures that group several shared-memory multicores interconnected through a network. This paper addresses important limitations in the code generation for Partitioned Global Address Space (PGAS) languages. These languages allow fine-grained communication and lead to programs that perform many fine-grained accesses to data. When the data is distributed to remote computing nodes, code transformations are required to prevent performance degradation. Until now code transformations to PGAS programs have been restricted to the cases where both the physical mapping of the data or the number of processing nodes are known at compilation time. In this paper, a novel application of the inspector-executor model overcomes these limitations and allows profitable code transformations, which result in fewer and larger messages sent through the network, when neither the data mapping nor the number of processing nodes are known at compilation time. A performance evaluation reports both scaling and absolute performance numbers on up to 32768 cores of a Power 775 supercomputer. This evaluation indicates that the compiler transformation results in speedups between 1.15X and 21X over a baseline and that these automated transformations achieve up to 63% the performance of the MPI versions.

Optimization techniques for fine-grained communication in PGAS environments

“…• It presents coalescing and prefetching techniques that use loop transformations and runtime support to increase communication efficiency and tolerate the network latencies [43,36,38]. It demonstrates that dynamic analysis and coalescing at runtime can significantly improve the performance of fine-grained accesses inside loops, without having to know the physical data mapping or use the upc forall parallel structure.…”

“…During the course of the thesis, we have published a number of articles. First of all, the algorithm of inspector-executor presented in Chapter 4 was submitted as patent [43]. The patent describes the basic design aspects of the transformation.…”