The increasing size of input graphs for graph neural networks (GNNs) highlights the demand for using multi-GPU platforms. However, existing multi-GPU GNN systems optimize the computation and communication individually based on the conventional practice of scaling dense DNNs. For irregularly sparse and fine-grained GNN workloads, such solutions miss the opportunity to jointly schedule/optimize the computation and communication operations for high-performance delivery.
To this end, we propose MGG , a novel system design to accelerate full-graph GNNs on multi-GPU platforms. The core of MGG is its novel dynamic software pipeline to facilitate fine-grained computation-communication overlapping within a GPU kernel. Specifically, MGG introduces GNN-tailored pipeline construction and GPU-aware pipeline mapping to facilitate workload balancing and operation overlapping. MGG also incorporates an intelligent runtime design with analytical modeling and optimization heuristics to dynamically improve the execution performance. Extensive evaluation reveals that MGG outperforms state-of-the-art full-graph GNN systems across various settings: on average 4.41×, 4.81×, and 10.83× faster than DGL, MGG-UVM, and ROC, respectively.
more »
« less
IMA-GNN: In-Memory Acceleration of Centralized and Decentralized Graph Neural Networks at the Edge
In this paper, we propose IMA-GNN as an In-Memory Accelerator for centralized and decentralized Graph Neural Network inference, explore its potential in both settings and provide a guideline for the community targeting flexible and efficient edge computation. Leveraging IMA-GNN, we first model the computation and communication latencies of edge devices. We then present practical case studies on GNN-based taxi demand and supply prediction and also adopt four large graph datasets to quantitatively compare and analyze centralized and decentralized settings. Our cross-layer simulation results demonstrate that on average, IMA-GNN in the centralized setting can obtain ~790x communication speed-up compared to the decentralized GNN setting. However, the decentralized setting performs computation ~1400x faster while reducing the power consumption per device. This further underlines the need for a hybrid semi-decentralized GNN approach.
more »
« less
- PAR ID:
- 10426829
- Date Published:
- Journal Name:
- GLSVLSI '23: Proceedings of the Great Lakes Symposium on VLSI 2023
- Page Range / eLocation ID:
- 3 to 8
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Learning safe solutions is an important but challenging problem in multi-agent reinforcement learning (MARL). Shielded reinforcement learning is one approach for preventing agents from choosing unsafe actions. Current shielded reinforcement learning methods for MARL make strong assumptions about communication and full observability. In this work, we extend the formalization of the shielded reinforcement learning problem to a decentralized multi-agent setting. We then present an algorithm for decomposition of a centralized shield, allowing shields to be used in such decentralized, communication-free environments. Our results show that agents equipped with decentralized shields perform comparably to agents with centralized shields in several tasks, allowing shielding to be used in environments with decentralized training and execution for the first time.more » « less
-
Camps-Valls, Gustau ; Ruiz, Francisco J. ; Valera, Isabel (Ed.)Linear contextual bandit is a popular online learning problem. It has been mostly studied in centralized learning settings. With the surging demand of large-scale decentralized model learning, e.g., federated learning, how to retain regret minimization while reducing communication cost becomes an open challenge. In this paper, we study linear contextual bandit in a federated learning setting. We propose a general framework with asynchronous model update and communication for a collection of homogeneous clients and heterogeneous clients, respectively. Rigorous theoretical analysis is provided about the regret and communication cost under this distributed learning framework; and extensive empirical evaluations demonstrate the effectiveness of our solution.more » « less
-
null (Ed.)Web tracking and advertising (WTA) nowadays are ubiquitously performed on the web, continuously compromising users' privacy. Existing defense solutions, such as widely deployed blocking tools based on filter lists and alternative machine learning based solutions proposed in prior research, have limitations in terms of accuracy and effectiveness. In this work, we propose WtaGraph, a web tracking and advertising detection framework based on Graph Neural Networks (GNNs). We first construct an attributed homogenous multi-graph (AHMG) that represents HTTP network traffic, and formulate web tracking and advertising detection as a task of GNN-based edge representation learning and classification in AHMG. We then design four components in WtaGraph so that it can (1) collect HTTP network traffic, DOM, and JavaScript data, (2) construct AHMG and extract corresponding edge and node features, (3) build a GNN model for edge representation learning and WTA detection in the transductive learning setting, and (4) use a pre-trained GNN model for WTA detection in the inductive learning setting. We evaluate WtaGraph on a dataset collected from Alexa Top 10K websites, and show that WtaGraph can effectively detect WTA requests in both transductive and inductive learning settings. Manual verification results indicate that WtaGraph can detect new WTA requests that are missed by filter lists and recognize non-WTA requests that are mistakenly labeled by filter lists. Our ablation analysis, evasion evaluation, and real-time evaluation show that WtaGraph can have a competitive performance with flexible deployment options in practice.more » « less
-
Despite the recent success of Graph Neural Networks (GNNs), training GNNs on large graphs remains challenging. The limited resource capacities of the existing servers, the dependency between nodes in a graph, and the privacy concern due to the centralized storage and model learning have spurred the need to design an effective distributed algorithm for GNN training. However, existing distributed GNN training methods impose either excessive communication costs or large memory overheads that hinders their scalability. To overcome these issues, we propose a communication-efficient distributed GNN training technique named (LLCG). To reduce the communication and memory overhead, each local machine in LLCG first trains a GNN on its local data by ignoring the dependency between nodes among different machines, then sends the locally trained model to the server for periodic model averaging. However, ignoring node dependency could result in significant performance degradation. To solve the performance degradation, we propose to apply on the server to refine the locally learned models. We rigorously analyze the convergence of distributed methods with periodic model averaging for training GNNs and show that naively applying periodic model averaging but ignoring the dependency between nodes will suffer from an irreducible residual error. However, this residual error can be eliminated by utilizing the proposed global corrections to entail fast convergence rate. Extensive experiments on real-world datasets show that LLCG can significantly improve the efficiency without hurting the performance. One-sentence Summary: We propose LLCG a communication efficient distributed algorithm for training GNNs.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1145/3583781.3590248
