Accelerating Collective Communication in Data Parallel Training across Deep Learning Frameworks
Offered By: USENIX via YouTube
Course Description
Overview
Explore a 17-minute conference talk from USENIX NSDI '22 that delves into accelerating collective communication in data parallel training across deep learning frameworks. Learn about new techniques developed within Horovod, a generic communication library, to improve the control plane and enhance performance in large-scale distributed training. Discover how the researchers implemented a caching strategy and decentralized orchestration to optimize the coordinator-worker logic, and introduced a feature for users to group collective operations for finer control over communication buffer sizes. Examine the experimental results conducted on the Summit supercomputer, comparing the proposed strategies against Horovod's original design, tf.distribute, torch.DDP, and BytePS. Gain insights into the impressive performance improvements achieved, including a 2x speedup at 6000 GPUs scale and near-linear scaling of 0.93 with 1.54 exaflops sustained performance using 27,600 GPUs on a scientific application (STEMDL).
Syllabus
NSDI '22 - Accelerating Collective Communication in Data Parallel Training across Deep Learning...
Taught by
USENIX
Related Courses
Scaling Memcache at FacebookUSENIX via YouTube Multi-Person Localization via RF Body Reflections
USENIX via YouTube Opaque - An Oblivious and Encrypted Distributed Analytics Platform
USENIX via YouTube Live Video Analytics at Scale with Approximation and Delay-Tolerance
USENIX via YouTube Clipper - A Low-Latency Online Prediction Serving System
USENIX via YouTube