Heterogeneous Scheduling of Deep Neural Networks for Low-power Real-time Designs

Abstract

Deep neural networks have become the readiest answer to a range of application challenges including image recognition, stock analysis, natural language processing, and biomedical applications such as seizure detection. All while outperforming prior leading solutions that relied heavily on hand-engineered techniques. However, deployment of these neural networks often requires high-computational and memory-intensive solutions. These requirements make it challenging to deploy Deep Neural Networks (DNNs) in embedded, real-time low-power applications where classic architectures, GPUs and CPUs, still impose significant power burden. Systems-on-Chip (SoC) with Field-programmable Gate Arrays (FPGAs) can be used to improve performance and allow more fine-grain control of resources than CPUs or GPUs, but it is difficult to find the optimal balance between hardware and software to improve DNN efficiency. In the current research literature there have been few proposed solutions to address optimizing hardware and software deployments of DNNs in embedded low-power systems. To address the computation resource restriction and low-power needs for deploying these networks, we describe and implement a domain-specific metric model for optimizing task deployment on differing platforms, hardware and software. Next, we propose a DNN hardware accelerator called Scalable Low-power Accelerator for real-time deep neural Networks (SCALENet) that includes multithreaded software workers. Finally, we propose a heterogeneous aware scheduler that uses the DNN-specific metric models and the SCALENet accelerator to allocate a task to a resource based on solving a numerical cost for a series of domain objectives. To demonstrate the applicability of our contribution, we deploy nine modern deep network architectures, each containing a different number of parameters within the context of two different neural network applications: image processing and biomedical seizure detection. Utilizing the metric modeling techniques integrated into the heterogeneous aware scheduler and the SCALENet accelerator, we demonstrate the ability to meet computational requirements, adapt to multiple architectures, and lower power by providing an optimized task to resource allocation. Our heterogeneous aware scheduler improves power saving by decreasing power consumption by 10% of the total system power, does not affect the accuracy of the networks, and still meets the real-time deadlines. We demonstrate the ability to achieve parity with or exceed the energy efficiency of NVIDIA GPUs when evaluated against Jetson TK1 with embedded GPU SoC and with a 4× power savings in a power envelope of 2.0W. When compared to existing FPGA-based accelerators, SCALENet’s accelerator and heterogeneous aware scheduler achieves a 4× improvement in energy efficiency.

Real-Time Hybrid Flow Shop Scheduling Approach in Smart Manufacturing Environment

“…The hybrid flow shop scheduling in a ubiquitous environment is different from the conventional scheduling. It is equipped with a large number of active intelligent devices, such as RFID tags, RFID readers, wireless networks, and Bluetooth devices [9] . These data collection devices are deployed on shop floors.…”

“…Each job follows rigid modes from the first stage to the last stage and must be processed on a certain machine at each stage. Figure 1 shows a manufacturing shop floor equipped with RFID tags in a ubiquitous environment [9] .…”

“…The real-time HFSP under a smart manufacturing environment can be described as follows [9] : New orders successively arrive at the shop floor with various kinds of information, such as arrival time, due time, and processing time at different stages of jobs. There is a series of parallel machines at each stage.…”

“…The study of Ref. [9] for a specific objective in the existing research. The proposed approach will be compared with these scheduling rules.…”

ALOHA: A Unified Platform-Aware Evaluation Method for CNNs Execution on Heterogeneous Systems at the Edge

“…However, the ongoing research in the field of Edge AI is exploring alternative, more efficient ways to execute CNNs on heterogeneous edge platforms [7]. For example, methods, proposed in papers [39]- [41], enable for better utilization of computational resources, available on the platform. The exploitation of these methods can significantly affect platform-aware CNN metrics, such as CNN throughput and energy consumption.…”