ASIC implementation of random number generators using SR latches and its evaluation

Torii, Naoya; Kokubo, Hirotaka; Yamamoto, Daisuke; Itoh, Kazuya; Takenaka, Masahiko; Matsumoto, Tsutomu

doi:10.1186/s13635-016-0036-1

Cited by 18 publications

(4 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In our evaluation board, with a single read cycle, to read a full 8KByte page, it takes on average ∼91.2µs. So according to the equation 1, our system level throughput is around ∼0.47Mbps, which is comparable with the performance of many popular hardware-based random numbers [5], [14], [33]. Note that, with our experimental setup, we were only able to read the memory module with an average speed of ∼720Mbps (400MHz system clock frequency), although, the maximum throughput of the memory module is almost double than that.…”

Section: Throughput Analysissupporting

confidence: 51%

Exploiting DRAM Latency Variations for Generating True Random Numbers

Talukder

Kerns

Ray

et al. 2019

2019 IEEE International Conference on Consumer Electronics (ICCE)

View full text Add to dashboard Cite

True random number generator (TRNG) plays a vital role in a variety of security applications and protocols. The security and privacy of an asset rely on the encryption, which solely depends on the quality of random numbers. Memory chips are widely used for generating random numbers because of their prevalence in modern electronic systems. Unfortunately, existing Dynamic Random-access Memory (DRAM)-based TRNGs produce random numbers with either limited entropy or poor throughput. In this paper, we propose a DRAM-latency based TRNG that generates high-quality random numbers. The silicon results from Samsung and Micron DDR3 DRAM modules show that our proposed DRAM-latency based TRNG is robust (against different operating conditions and environmental variations) and acceptably fast.Index Terms-Random number, TRNG, DRAM-based security primitives, DRAM-based TRNG, Memory-based security primitives, Memory-based TRNG, hardware-based security primitives.

show abstract

Section: Throughput Analysissupporting

confidence: 51%

Exploiting DRAM Latency Variations for Generating True Random Numbers

Talukder

Kerns

Ray

et al. 2019

2019 IEEE International Conference on Consumer Electronics (ICCE)

View full text Add to dashboard Cite

show abstract

“…This stochastic binarization is more appealing theoretically (see Section 4) than the sign function, but somewhat harder to implement as it requires the hardware to generate random bits when quantizing (Torii et al, 2016). As a result, we mostly use the deterministic binarization function (i.e., the sign function), with the exception of activations at traintime in some of our experiments.…”

Section: Deterministic Vs Stochastic Binarizationmentioning

confidence: 99%

Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations

Hubara,

Courbariaux,

Soudry

et al. 2016

Preprint

161

View full text Add to dashboard Cite

We introduce a method to train Quantized Neural Networks (QNNs) -neural networks with extremely low precision (e.g., 1-bit) weights and activations, at run-time. At traintime the quantized weights and activations are used for computing the parameter gradients. During the forward pass, QNNs drastically reduce memory size and accesses, and replace most arithmetic operations with bit-wise operations. As a result, power consumption is expected to be drastically reduced. We trained QNNs over the MNIST, CIFAR-10, SVHN and ImageNet datasets. The resulting QNNs achieve prediction accuracy comparable to their 32-bit counterparts. For example, our quantized version of AlexNet with 1-bit weights and 2-bit activations achieves 51% top-1 accuracy. Moreover, we quantize the parameter gradients to 6-bits as well which enables gradients computation using only bit-wise operation. Quantized recurrent neural networks were tested over the Penn Treebank dataset, and achieved comparable accuracy as their 32-bit counterparts using only 4-bits. Last but not least, we programmed a binary matrix multiplication GPU kernel with which it is possible to run our MNIST QNN 7 times faster than with an unoptimized GPU kernel, without suffering any loss in classification accuracy. The QNN code is available online.

show abstract

“…9 by altering the slopes. Further in the forward pass, instead of using stochastic quantization which requires random seeds generating process [43], deterministic quantization are adopted and derived by setting the probability threshold to be p = 0.5,…”

Section: Forward Pass: Threshold Learning Quantizationmentioning

confidence: 99%

Nonuniform-to-Uniform Quantization: Towards Accurate Quantization via Generalized Straight-Through Estimation

Liu¹,

Cheng²,

Huang³

et al. 2021

Preprint

View full text Add to dashboard Cite

The nonuniform quantization strategy for compressing neural networks usually achieves better performance than its counterpart, i.e., uniform strategy, due to its superior representational capacity. However, many nonuniform quantization methods overlook the complicated projection process in implementing the nonuniformly quantized weights/activations, which incurs non-negligible time and space overhead in hardware deployment. In this study, we propose Nonuniform-to-Uniform Quantization (N2UQ), a method that can maintain the strong representation ability of nonuniform methods while being hardware-friendly and efficient as the uniform quantization for model inference. We achieve this through learning the flexible inequidistant input thresholds to better fit the underlying distribution while quantizing these real-valued inputs into equidistant output levels. To train the quantized network with learnable input thresholds, we introduce a generalized straight-through estimator (G-STE) for intractable backward derivative calculation w.r.t. threshold parameters. Additionally, we consider entropy preserving regularization to further reduce information loss in weight quantization. Even under this adverse constraint of imposing uniformly quantized weights and activations, our N2UQ outperforms state-of-the-art nonuniform quantization methods by 0.7∼1.8% on ImageNet, demonstrating the contribution of N2UQ design. Code will be made publicly available.

show abstract

ASIC implementation of random number generators using SR latches and its evaluation

Cited by 18 publications

References 19 publications

Exploiting DRAM Latency Variations for Generating True Random Numbers

Exploiting DRAM Latency Variations for Generating True Random Numbers

Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations

Nonuniform-to-Uniform Quantization: Towards Accurate Quantization via Generalized Straight-Through Estimation

Contact Info

Product

Resources

About