Gluon: a communication-optimizing substrate for distributed heterogeneous graph analytics

Abstract

This paper introduces a new approach to building distributedmemory graph analytics systems that exploits heterogeneity in processor types (CPU and GPU), partitioning policies, and programming models. The key to this approach is Gluon, a communication-optimizing substrate. Programmers write applications in a shared-memory programming system of their choice and interface these applications with Gluon using a lightweight API. Gluon enables these programs to run on heterogeneous clusters and optimizes communication in a novel way by exploiting structural and temporal invariants of graph partitioning policies. To demonstrate Gluon's ability to support different programming models, we interfaced Gluon with the Galois and Ligra shared-memory graph analytics systems to produce distributed-memory versions of these systems named D-Galois and D-Ligra, respectively. To demonstrate Gluon's ability to support heterogeneous processors, we interfaced Gluon with IrGL, a state-of-the-art single-GPU system for * Both authors contributed equally.

Optimizing ordered graph algorithms with GraphIt

“…Approximate Priority Ordering. Some work disregard a strict priority ordering and use an approximate priority ordering [3,4,12,35]. This approach uses several "relaxed" priority queues in parallel to maintain local priority ordering.…”

“…However, it does not synchronize globally among the different priority queues. To the best of our knowledge and from communications with the developers, Galois [12,35] does not currently support strict priority ordering and only supports an approximate ordering. Galois [35] provides an ordered list abstraction, which does not explicitly synchronize after each priority.…”

Custom code generation for a graph DSL

“…Bulk Synchronous Parallel (BSP) Model [23] of execution and asynchronous executions are popular models of executions. Gluon [4] uses Galois and Ligra and generates distributed-memory versions of these systems.…”

DisGCo

“…Here, message aggregation can greatly optimize the number of remote communications. Figure 9 shows how the doubly nested loop (lines [13][14][15][16][17] in Figure 2 is translated to BSP model. It can be seen that the loop headers (lines 5 and 7) in Figure 9 are similar to that in Figure 5 (lines 4 and 6).…”