On the Reproducibility of MPI Reduction Operations

TitleOn the Reproducibility of MPI Reduction Operations
Publication TypeConference Paper
Year of Publication2013
AuthorsBalaji, P, Kimpe, D
Conference NameHigh Performance computing and Communications (HPCC 2013)
Conference LocationZhangjiajie, China
Other NumbersANL/MCS-P4093-0713
AbstractMany scientific applications go through a thorough validation and verification ("V&V") process to demonstrate that the computer simulation does, in fact, mirror what can be analyzed through physical experimentation. Given the complexity of and the time required for the V&V process, applications that have been validated and verified are typically conservative with respect to changes that might impact the reproducibility of their results. In the extreme case, some applications require bitwise reproducibility for their simulations. Thus, any change made to the application, the hardware, or the software on the system needs to preserve the bitwise reproducibility of the application. Such a constraint, however, can drastically affect the performance efficiency of the system in many ways. In this paper, we analyze the impact of such bitwise reproducibility on the performance efficiency of MPI reduction operations.
PDFhttp://www.mcs.anl.gov/papers/P4093-0713_1.pdf