CONTINUUM DYNAMICS INC — Department of Defense SBIR Phase II: DTRA172-002

Description

A significant performance gap exists between the theoretical number of Floating Point Operations (FLOPS) that an HPC machine is capable of sustaining and the number of FLOPS realized by real-world HPC software. One of the principal reasons for this gap is the parasitic work that computational processes must do to communicate with one another. It has been shown that this communication work can be significantly reduced by using a topology aware mapping of processes to compute nodes. Tools for process mapping have been developed; however, existing tools lack support for dynamic communication patterns, variable network loads, and support legacy applications. This work will overcome these limitations by developing a MPI based library that implements topology aware mapping. Furthermore, a command line utility will be developed that uses application profile data to estimate application performance on new platforms. In Phase I, work towards a prototype MPI based toolkit, named TAPT, along with supporting theory was accomplished. In Phase II a production quality version of TAPT will be developed, along with user and system administration documentation.