skip to main content


The NSF Public Access Repository (NSF-PAR) system and access will be unavailable from 11:00 PM ET on Friday, September 29 until 11:59 PM ET on Saturday, September 30 due to maintenance. We apologize for the inconvenience.

Title: Runtime Long-Term Reliability Management Using Stochastic Computing in Deep Neural Networks
In this paper, we propose a new dynamic reliability technique using an accuracy-reconfigurable stochastic computing (ARSC) framework for deep learning computing. Unlike the conventional stochastic computing that conducts design time accuracy power/energy trade-off, the new ARSC design can adjust the bit-width of the data in run time. Hence, the ARSC can mitigate the long-term aging effects by slowing the system clock frequency, while maintaining the inference throughput by reducing the data bit-width at a small cost of accuracy. We show how to implement the recently proposed counter-based SC multiplication and bit-width reduction on a layer-wise quantization scheme for CNN networks with dynamic fixed-point data. We validate an ARSC-based five-layer convolutional neural network designs for the MNIST dataset based on Vivado HLS with constraints from Xilinx Zynq-7000 family xc7z045 platform. Experimental results show that new ARSC DNN can sufficiently compensate the NBTI induced aging effects in 10 years with marginal classification accuracy loss while maintaining or even exceeding the pre-aging computing throughput. At the same time, the proposed ARSC computing framework also reduces the active power consumption due to the frequency scaling, which can further improve system reliability due to the reduced temperature.  more » « less
Award ID(s):
1816361 2007135
Author(s) / Creator(s):
; ; ;
Date Published:
Journal Name:
Proc. Int. Symposium. on Quality Electronic Design (ISQED’21)
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. In this paper, we propose a novel accuracy-reconfigurable stochastic computing (ARSC) framework for dynamic reliability and power management. Different than the existing stochastic computing works, where the accuracy versus power/energy trade-off is carried out in the design time, the new ARSC design can change accuracy or bit-width of the data in the run-time so that it can accommodate the long-term aging effects by slowing the system clock frequency at the cost of accuracy while maintaining the throughput of the computing. We validate the ARSC concept on a discrete cosine transformation (DCT) and inverse DCT designs for image compressing/decompressing applications, which are implemented on Xilinx Spartan-6 family XC6SLX45 platform. Experimental results show that the new design can easily mitigate the long-term aging-induced effects by accuracy trade-off while maintaining the throughput of the whole computing process using simple frequency scaling. We further show that one-bit precision loss for the input data, which translated to 3.44dB of the accuracy loss in term of Peak Signal to Noise Ratio (PSNR) for images, we can sufficiently compensate the NBTI induced aging effects in 10 years while maintaining the pre-aging computing throughput of 7.19 frames per second. At the same time, we can save 74\% power consumption by 10.67dB of accuracy loss. The proposed ARSC computing framework also allows much aggressive frequency scaling, which can lead to order of magnitude power savings compared to the traditional dynamic voltage and frequency scaling (DVFS) techniques. 
    more » « less
  2. Wireless networks are being applied in various industrial sectors, and they are posed to support mission-critical industrial IoT applications which require ultra-reliable, low-latency communications (URLLC). Ensuring predictable per-packet communication reliability is a basis of predictable URLLC, and scheduling and power control are two basic enablers. Scheduling and power control, however, are subject to challenges such as harsh environments, dynamic channels, and distributed network settings in industrial IoT. Existing solutions are mostly based on heuristic algorithms or asymptotic analysis of network performance, and there lack field-deployable algorithms for ensuring predictable per-packet reliability. Towards addressing the gap, we examine the cross-layer design of joint scheduling and power control and analyze the associated challenges. We introduce the Perron–Frobenius theorem to demonstrate that scheduling is a must for ensuring predictable communication reliability, and by investigating characteristics of interference matrices, we show that scheduling with close-by links silent effectively constructs a set of links whose required reliability is feasible with proper transmission power control. Given that scheduling alone is unable to ensure predictable communication reliability while ensuring high throughput and addressing fast-varying channel dynamics, we demonstrate how power control can help improve both the reliability at each time instant and throughput in the long-term. Based on the analysis, we propose a candidate framework of joint scheduling and power control, and we demonstrate how this framework behaves in guaranteeing per-packet communication reliability in the presence of wireless channel dynamics of different time scales. Collectively, these findings provide insight into the cross-layer design of joint scheduling and power control for ensuring predictable per-packet reliability in the presence of wireless network dynamics and uncertainties. 
    more » « less
  3. Deep Convolution Neural Network (CNN) has achieved outstanding performance in image recognition over large scale dataset. However, pursuit of higher inference accuracy leads to CNN architecture with deeper layers and denser connections, which inevitably makes its hardware implementation demand more and more memory and computational resources. It can be interpreted as `CNN power and memory wall'. Recent research efforts have significantly reduced both model size and computational complexity by using low bit-width weights, activations and gradients, while keeping reasonably good accuracy. In this work, we present different emerging nonvolatile Magnetic Random Access Memory (MRAM) designs that could be leveraged to implement `bit-wise in-memory convolution engine', which could simultaneously store network parameters and compute low bit-width convolution. Such new computing model leverages the `in-memory computing' concept to accelerate CNN inference and reduce convolution energy consumption due to intrinsic logic-in-memory design and reduction of data communication. 
    more » « less
  4. With the success of deep neural networks (DNN), many recent works have been focusing on developing hardware accelerator for power and resource-limited embedded system via model compression techniques, such as quantization, pruning, low-rank approximation, etc. However, almost all existing DNN structure is fixed after deployment, which lacks runtime adaptive DNN structure to adapt to its dynamic hardware resource, power budget, throughput requirement, as well as dynamic workload. Correspondingly, there is no runtime adaptive hardware platform to support dynamic DNN structure. To address this problem, we first propose a dynamic channel-adaptive deep neural network (CA-DNN) which can adjust the involved convolution channel (i.e. model size, computing load) at run-time (i.e. at inference stage without retraining) to dynamically trade off between power, speed, computing load and accuracy. Further, we utilize knowledge distillation method to optimize the model and quantize the model to 8-bits and 16-bits, respectively, for hardware friendly mapping. We test the proposed model on CIFAR-10 and ImageNet dataset by using ResNet. Comparing with the same model size of individual model, our CA-DNN achieves better accuracy. Moreover, as far as we know, we are the first to propose a Processing-in-Memory accelerator for such adaptive neural networks structure based on Spin Orbit Torque Magnetic Random Access Memory(SOT-MRAM) computational adaptive sub-arrays. Then, we comprehensively analyze the trade-off of the model with different channel-width between the accuracy and the hardware parameters, eg., energy, memory, and area overhead. 
    more » « less
  5. Obeid, I. ; Selesnik, I. ; Picone, J. (Ed.)
    The Neuronix high-performance computing cluster allows us to conduct extensive machine learning experiments on big data [1]. This heterogeneous cluster uses innovative scheduling technology, Slurm [2], that manages a network of CPUs and graphics processing units (GPUs). The GPU farm consists of a variety of processors ranging from low-end consumer grade devices such as the Nvidia GTX 970 to higher-end devices such as the GeForce RTX 2080. These GPUs are essential to our research since they allow extremely compute-intensive deep learning tasks to be executed on massive data resources such as the TUH EEG Corpus [2]. We use TensorFlow [3] as the core machine learning library for our deep learning systems, and routinely employ multiple GPUs to accelerate the training process. Reproducible results are essential to machine learning research. Reproducibility in this context means the ability to replicate an existing experiment – performance metrics such as error rates should be identical and floating-point calculations should match closely. Three examples of ways we typically expect an experiment to be replicable are: (1) The same job run on the same processor should produce the same results each time it is run. (2) A job run on a CPU and GPU should produce identical results. (3) A job should produce comparable results if the data is presented in a different order. System optimization requires an ability to directly compare error rates for algorithms evaluated under comparable operating conditions. However, it is a difficult task to exactly reproduce the results for large, complex deep learning systems that often require more than a trillion calculations per experiment [5]. This is a fairly well-known issue and one we will explore in this poster. Researchers must be able to replicate results on a specific data set to establish the integrity of an implementation. They can then use that implementation as a baseline for comparison purposes. A lack of reproducibility makes it very difficult to debug algorithms and validate changes to the system. Equally important, since many results in deep learning research are dependent on the order in which the system is exposed to the data, the specific processors used, and even the order in which those processors are accessed, it becomes a challenging problem to compare two algorithms since each system must be individually optimized for a specific data set or processor. This is extremely time-consuming for algorithm research in which a single run often taxes a computing environment to its limits. Well-known techniques such as cross-validation [5,6] can be used to mitigate these effects, but this is also computationally expensive. These issues are further compounded by the fact that most deep learning algorithms are susceptible to the way computational noise propagates through the system. GPUs are particularly notorious for this because, in a clustered environment, it becomes more difficult to control which processors are used at various points in time. Another equally frustrating issue is that upgrades to the deep learning package, such as the transition from TensorFlow v1.9 to v1.13, can also result in large fluctuations in error rates when re-running the same experiment. Since TensorFlow is constantly updating functions to support GPU use, maintaining an historical archive of experimental results that can be used to calibrate algorithm research is quite a challenge. This makes it very difficult to optimize the system or select the best configurations. The overall impact of all of these issues described above is significant as error rates can fluctuate by as much as 25% due to these types of computational issues. Cross-validation is one technique used to mitigate this, but that is expensive since you need to do multiple runs over the data, which further taxes a computing infrastructure already running at max capacity. GPUs are preferred when training a large network since these systems train at least two orders of magnitude faster than CPUs [7]. Large-scale experiments are simply not feasible without using GPUs. However, there is a tradeoff to gain this performance. Since all our GPUs use the NVIDIA CUDA® Deep Neural Network library (cuDNN) [8], a GPU-accelerated library of primitives for deep neural networks, it adds an element of randomness into the experiment. When a GPU is used to train a network in TensorFlow, it automatically searches for a cuDNN implementation. NVIDIA’s cuDNN implementation provides algorithms that increase the performance and help the model train quicker, but they are non-deterministic algorithms [9,10]. Since our networks have many complex layers, there is no easy way to avoid this randomness. Instead of comparing each epoch, we compare the average performance of the experiment because it gives us a hint of how our model is performing per experiment, and if the changes we make are efficient. In this poster, we will discuss a variety of issues related to reproducibility and introduce ways we mitigate these effects. For example, TensorFlow uses a random number generator (RNG) which is not seeded by default. TensorFlow determines the initialization point and how certain functions execute using the RNG. The solution for this is seeding all the necessary components before training the model. This forces TensorFlow to use the same initialization point and sets how certain layers work (e.g., dropout layers). However, seeding all the RNGs will not guarantee a controlled experiment. Other variables can affect the outcome of the experiment such as training using GPUs, allowing multi-threading on CPUs, using certain layers, etc. To mitigate our problems with reproducibility, we first make sure that the data is processed in the same order during training. Therefore, we save the data from the last experiment and to make sure the newer experiment follows the same order. If we allow the data to be shuffled, it can affect the performance due to how the model was exposed to the data. We also specify the float data type to be 32-bit since Python defaults to 64-bit. We try to avoid using 64-bit precision because the numbers produced by a GPU can vary significantly depending on the GPU architecture [11-13]. Controlling precision somewhat reduces differences due to computational noise even though technically it increases the amount of computational noise. We are currently developing more advanced techniques for preserving the efficiency of our training process while also maintaining the ability to reproduce models. In our poster presentation we will demonstrate these issues using some novel visualization tools, present several examples of the extent to which these issues influence research results on electroencephalography (EEG) and digital pathology experiments and introduce new ways to manage such computational issues. 
    more » « less