Title:
GNNBuilder: An Automated Framework for Generic Graph Neural Network Accelerator Generation, Simulation, and Optimization
GNNBuilder: An Automated Framework for Generic Graph Neural Network Accelerator Generation, Simulation, and Optimization
Author(s)
Abi-Karam, Stefan
Advisor(s)
Hao, Cong
Krishna, Tushar
Kim, Hyesoon
Krishna, Tushar
Kim, Hyesoon
Editor(s)
Collections
Supplementary to
Permanent Link
Abstract
There are plenty of graph neural network (GNN) accelerators being proposed. How-
ever, they highly rely on users’ hardware expertise and are usually optimized for one specific GNN model, making them challenging for practical usage. Therefore, in this work, we
propose GNNBuilder, the first automated, generic, end-to-end GNN accelerator generation
framework. It features four advantages: (1) GNNBuilder can automatically generate GNN
accelerators for a wide range of GNN models arbitrarily defined by users; (2) GNNBuilder
takes standard PyTorch programming interface, introducing zero overhead for algorithm
developers; (3) GNNBuilder supports end-to-end code generation, simulation, accelerator
optimization, and hardware deployment, realizing a push-button fashion for GNN accelerator design; (4) GNNBuilder is equipped with accurate performance models of the generated
accelerator, enabling fast and flexible design space exploration (DSE). In the experiments,
we show that our accelerator performance model has errors within 34% for latency prediction and 22% for BRAM count prediction. We also show that our generated accelerators
can outperform CPU by 2.96× and GPU by 2.99×. This framework is open-source, and
the code is available at https://anonymous.4open.science/r/gnn-builder.
Sponsor
Date Issued
2022-12-08
Extent
Resource Type
Text
Resource Subtype
Thesis