On Memory System Design for Stochastic Computing

Khatamifard, S. Karen; Najafi, M. Hassan; Ghoreyshi, Ali; Karpuzcu, Ulya R.; Lilja, David J.

doi:10.1109/lca.2018.2804926

Cited by 11 publications

(10 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the digital units to generate unary bitstreams are more expensive compared to the analog counterparts. There exist explorations into the low-cost conversion of analog signals into rate-coded bitstreams for image processing applications [3,17,30,38,63], i.e., analog-to-stochastic conversion. Onizawa et al leverage Magnetic-Tunnel Junction (MTJ) devices [63], whose state switching is probabilistic.…”

Section: Analog-to-stochastic Conversionmentioning

confidence: 99%

uBrain

Pan

et al. 2022

Proceedings of the 49th Annual International Symposium on Computer Architecture

View full text Add to dashboard Cite

Brain computer interfaces (BCIs) have been widely adopted to enhance human perception via brain signals with abundant spatialtemporal dynamics, such as electroencephalogram (EEG). In recent years, BCI algorithms are moving from classical feature engineering to emerging deep neural networks (DNNs), allowing to identify the spatial-temporal dynamics with improved accuracy. However, existing BCI architectures are not leveraging such dynamics for hardware eciency. In this work, we present uBrain, a unary computing BCI architecture for DNN models with cascaded convolutional and recurrent neural networks to achieve high task capability and hardware eciency. uBrain co-designs the algorithm and hardware: the DNN architecture and the hardware architecture are optimized with customized unary operations and immediate signal processing after sensing, respectively. Experiments show that uBrain, with negligible accuracy loss, surpasses the CPU, systolic array and stochastic computing baselines in on-chip power eciency by 9.0⇥, 6.2⇥ and 2.0⇥. CCS CONCEPTS• Hardware ! Emerging interfaces; Application specic integrated circuits; • Computing methodologies ! Specialpurpose algebraic systems; Articial intelligence.

show abstract

Section: Analog-to-stochastic Conversionmentioning

confidence: 99%

uBrain

Pan

et al. 2022

Proceedings of the 49th Annual International Symposium on Computer Architecture

View full text Add to dashboard Cite

show abstract

“…Both of them take up a large proportion in the overall power consumption. Therefore, works on low power converters [12] and new computing shows the conventional structure of SC memory system [6]. Input data from analog image sensor is converted into binary numbers through analog to digital converter (ADC) and then stored in standard SRAM array.…”

Section: A Stochastic Computing (Sc)mentioning

confidence: 99%

“…Ref. [6] proposed an analog memory system called StochMem, which is designed for SC to reduce extra conversion cost. However, analog memory is hard to design, and is incompatible with digital memory.…”

Section: Proposed Memory Structure Designed For Sc-based Mac Enginementioning

confidence: 99%

“…2. shows the conventional structure of SC memory system [6]. Input data from analog image sensor is converted into binary numbers through analog to digital converter (ADC) and then stored in standard SRAM array.…”

Section: B Conventional Structure Of Sc-based Mac Enginementioning

confidence: 99%

“…Ref. [6] proposed a new memory system specifically designed for SC called StochMem. However, StochMem stores data with analog memory [7], which usually costs a large energy and is not compatible with the conventional memory system.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Memory System Designed for Multiply-Accumulate (MAC) Engine Based on Stochastic Computing

Zhang

Wang

Zhang

et al. 2019

2019 International Conference on IC Design and Technology (ICICDT)

View full text Add to dashboard Cite

Convolutional neural network (CNN) achieves excellent performance on fascinating tasks such as image recognition and natural language processing at the cost of high power consumption. Stochastic computing (SC) is an attractive paradigm implemented in low power applications which performs arithmetic operations with simple logic and low hardware cost. However, conventional memory structure designed and optimized for binary computing leads to extra data conversion costs, which significantly decreases the energy efficiency. Therefore, a new memory system designed for SC-based multiply-accumulate (MAC) engine applied in CNN which is compatible with conventional memory system is proposed in this paper. As a result, the overall energy consumption of our new computing structure is 0.91pJ, which is reduced by 82.1% compared with the conventional structure, and the energy efficiency achieves 164.8 TOPS/W. I.

show abstract

Accelerating stochastic‐based reliability estimation for combinational circuits at RTL using GPU parallel computing

Xiao

Shen

et al. 2022

Int J of Intelligent Sys

View full text Add to dashboard Cite

Reliable circuits help prevent artificial intelligence (AI) systems from being corrupted by the soft errors occurred in memories or combinational circuits, which promotes the development of AI security. However, it is a great challenge to measure the reliability of combinational circuits at register transfer level (RTL) rapidly and efficiently. In this paper, a new fast and accurate computational model based on stochastic computation (SC) is presented to meet these objectives.In the proposed approach, the circuit netlists at RTL are parsed to satisfy the requirements of SC on the bitstream structure of the circuits, and then a Sobol sequence-based algorithm for generating uniform non-Bernoulli sequences is built to reduce the random fluctuations occurred in probability calculations. After that, an adaptive algorithm based on a MAX-MIN ant system is constructed using graphics processing unit-based parallel schemes to greatly accelerate the calculation. The experimental results validate our proposed technique, showing that this approach was approximately 51 and 42 times faster than the traditional SC approach and the stochastic computational model (SCM), respectively; its required sequence length was approximately 1.66 times shorter than that

show abstract

On Memory System Design for Stochastic Computing

Cited by 11 publications

References 15 publications

uBrain

uBrain

Memory System Designed for Multiply-Accumulate (MAC) Engine Based on Stochastic Computing

Accelerating stochastic‐based reliability estimation for combinational circuits at RTL using GPU parallel computing

Contact Info

Product

Resources

About