In this paper, we pave a novel way towards the concept of bit-wise In-Memory Convolution Engine (IMCE) that could implement the dominant convolution computation of Deep Convolutional Neural Networks (CNN) within memory. IMCE employs parallel computational memory sub-array as a fundamental unit based on our proposed Spin Orbit Torque Magnetic Random Access Memory (SOT-MRAM) design. Then, we propose an accelerator system architecture based on IMCE to efficiently process low bit-width CNNs. This architecture can be leveraged to greatly reduce energy consumption dealing with convolutional layers and also accelerate CNN inference. The device to architecture co-simulation results show that the proposed system architecture can process low bit-width AlexNet on ImageNet data-set favorably with 785.25μJ/img, which consumes ~3× less energy than that of recent RRAM based counterpart. Besides, the chip area is ~4× smaller.
more »
« less
Accelerating Low Bit-Width Deep Convolution Neural Network in MRAM
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
- Award ID(s):
- 1740126
- NSF-PAR ID:
- 10094196
- Date Published:
- Journal Name:
- 2018 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)
- Page Range / eLocation ID:
- 533 to 538
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
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
-
Deep convolutional neural network (DNN) has demonstrated phenomenal success and been widely used in many computer vision tasks. However, its enormous model size and high computing complexity prohibits its wide deployment into resource limited embedded system, such as FPGA and mGPU. As the two most widely adopted model compression techniques, weight pruning and quantization compress DNN model through introducing weight sparsity (i.e., forcing partial weights as zeros) and quantizing weights into limited bit-width values, respectively. Although there are works attempting to combine the weight pruning and quantization, we still observe disharmony between weight pruning and quantization, especially when more aggressive compression schemes (e.g., Structured pruning and low bit-width quantization) are used. In this work, taking FPGA as the test computing platform and Processing Elements (PE) as the basic parallel computing unit, we first propose a PE-wise structured pruning scheme, which introduces weight sparsification with considering of the architecture of PE. In addition, we integrate it with an optimized weight ternarization approach which quantizes weights into ternary values ({-1,0,+1}), thus converting the dominant convolution operations in DNN from multiplication-and-accumulation (MAC) to addition-only, as well as compressing the original model (from 32-bit floating point to 2-bit ternary representation) by at least 16 times. Then, we investigate and solve the coexistence issue between PE-wise Structured pruning and ternarization, through proposing a Weight Penalty Clipping (WPC) technique with self-adapting threshold. Our experiment shows that the fusion of our proposed techniques can achieve the best state-of-the-art ∼21× PE-wise structured compression rate with merely 1.74%/0.94% (top-1/top-5) accuracy degradation of ResNet-18 on ImageNet dataset.more » « less
-
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
-
In this paper, an energy-efficient and high-speed comparator-based processing-in-memory accelerator (CMP-PIM) is proposed to efficiently execute a novel hardware-oriented comparator-based deep neural network called CMPNET. Inspired by local binary pattern feature extraction method combined with depthwise separable convolution, we first modify the existing Convolutional Neural Network (CNN) algorithm by replacing the computationally-intensive multiplications in convolution layers with more efficient and less complex comparison and addition. Then, we propose a CMP-PIM that employs parallel computational memory sub-array as a fundamental processing unit based on SOT-MRAM. We compare CMP-PIM accelerator performance on different data-sets with recent CNN accelerator designs. With the close inference accuracy on SVHN data-set, CMP-PIM can get ∼ 94× and 3× better energy efficiency compared to CNN and Local Binary CNN (LBCNN), respectively. Besides, it achieves 4.3× speed-up compared to CNN-baseline with identical network configuration.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/ISVLSI.2018.00103
