ANALYZING BOTTLENECKS IN LARGE RECOMMENDATION SYSTEMS

Zhang, Jialiang

Permalink

https://hdl.handle.net/2142/124900 Copy

Description

Title

ANALYZING BOTTLENECKS IN LARGE RECOMMENDATION SYSTEMS

Author(s)

Zhang, Jialiang

Issue Date

2021-05-01

Keyword(s)

Recommendation Systems; GPU; CUDA; Machine Learning; Embedding

Abstract

Training and inferencing recommendation systems often have a greater need for analysis and computation over a large amount of unstructured user-specific data blobs. One of the state-ofthe-art recommendation models is Deep Learning Recommendation Model (DLRM) by Facebook. DLRM model consumes a large memory for storing embedding features with terabytes in size during training and inference. Aside from the memory cost, the long training time of DLRM is another issue. In this work, we investigated the potential bottlenecks of DLRM and discuss in detail two recent improvements proposed in the literature: pipeDLRM and TTRec. PipeDLRM proposes pipeline parallelism and split the whole model onto several GPUs to address compute time without compromising on accuracy while the TT-Rec proposes a new compression method to save embedding memory consumption at a loss of accuracy to an acceptable range. Our analysis of these two models shows that irrespective of the method of implementation, they still have certain issues to improve. For instance, the embedding memory bottleneck still remains in the lookup operation of the embedding tables in PipeDLRM model. This is because PipeDLRM’s partition only sits on one GPU and impedes the further scaling up process. On the other hand, even though TT-Rec succeeds in reducing the memory complexity of the model, it also requires a significant amount of reuse of the compressed information to retain accuracy. These suggest that there is no right solution to address the memory capacity problem present in the DLRM.

Type of Resource

text

Language

eng

Owning Collections