INAM2: InfiniBand network analysis and monitoring with MPI

Abstract

Modern high-end computing is being driven by the tight integration of several hardware and software components. On the hardware front, there are the multi-/many-core architectures (including accelerators and co-processors) and high-end interconnects like InfiniBand that are continually pushing the envelope of raw performance. On the software side, there are several high performance implementations of popular parallel programming models that are designed to take advantage of the high-end features offered by the hardware components and deliver multipetaflop level performance to end applications. Together, these components allow scientists and engineers to tackle grand challenge problems in their respective domains. Understanding and gaining insights into the performance of end applications on these modern systems is a challenging task. Several researchers and hardware manufacturers have attempted to tackle this by designing tools to inspect the network level or MPI level activities. However, all existing tools perform the inspection in a disjoint fashion and are unable to correlate the data generated by profiling the network and MPI. This results in a loss of valuable information that can provide the insights required for understanding the performance of High-End Computing applications. In this paper, we take up this challenge and design InfiniBand Network Analysis and Monitoring with MPI-INAM2. INAM2 allows users to analyze and visualize the communication happening in the network in conjunction with data obtained from the MPI library. Our experimental analysis shows that the INAM2 is able to profile and visualize the communication with very low performance overhead at scale.