HAWX: A Hardware-Aware FrameWork for Fast and Scalable ApproXimation of DNNs
This work addresses the challenge of scalable and fast approximation of DNNs for hardware designers and researchers, offering a framework that scales exponentially with network size, but it is incremental as it builds on existing approximate computing methods.
The authors tackled the problem of efficiently exploring approximate computing configurations for deep neural networks by introducing HAWX, a hardware-aware framework that uses multi-level sensitivity scoring and predictive models to accelerate configuration evaluation, achieving over 23x speedup in layer-level search and more than 3 million times speedup in filter-level search for LeNet-5 while maintaining accuracy comparable to exhaustive search.
This work presents HAWX, a hardware-aware scalable exploration framework that employs multi-level sensitivity scoring at different DNN abstraction levels (operator, filter, layer, and model) to guide selective integration of heterogeneous AxC blocks. Supported by predictive models for accuracy, power, and area, HAWX accelerates the evaluation of candidate configurations, achieving over 23* speedup in a layer-level search with two candidate approximate blocks and more than (3*106)* speedup at the filter-level search only for LeNet-5, while maintaining accuracy comparable to exhaustive search. Experiments across state-of-the-art DNN benchmarks such as VGG-11, ResNet-18, and EfficientNetLite demonstrate that the efficiency benefits of HAWX scale exponentially with network size. The HAWX hardware-aware search algorithm supports both spatial and temporal accelerator architectures, leveraging either off-the-shelf approximate components or customized designs.