Pith

open record

sign in
Browse

arxiv: 2204.11224 · v1 · pith:UBEMCSCV · submitted 2022-04-24 · cs.DC

Optimizing Task Placement and Online Scheduling for Distributed GNN Training Acceleration

Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 reserved pith:UBEMCSCVrecord.jsonopen to challenge →

classification cs.DC
keywords trainingtaskdistributedplacementdataexecutiongraphscheduling
0
0 comments X
read the original abstract

Training Graph Neural Networks (GNN) on large graphs is resource-intensive and time-consuming, mainly due to the large graph data that cannot be fit into the memory of a single machine, but have to be fetched from distributed graph storage and processed on the go. Unlike distributed deep neural network (DNN) training, the bottleneck in distributed GNN training lies largely in large graph data transmission for constructing mini-batches of training samples. Existing solutions often advocate data-computation colocation, and do not work well with limited resources where the colocation is infeasible. The potentials of strategical task placement and optimal scheduling of data transmission and task execution have not been well explored. This paper designs an efficient algorithm framework for task placement and execution scheduling of distributed GNN training, to better resource utilization, improve execution pipelining, and expediting training completion. Our framework consists of two modules: (i) an online scheduling algorithm that schedules the execution of training tasks, and the data transmission plan; and (ii) an exploratory task placement scheme that decides the placement of each training task. We conduct thorough theoretical analysis, testbed experiments and simulation studies, and observe up to 67% training speed-up with our algorithm as compared to representative baselines.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.