Beyond CPU: Considering Memory Power Consumption of Software

Acar, Hayri; Alptekin, Gülfem Işıklar; Gelas, Jean-Patrick; Ghodous, Parisa

doi:10.5220/0005764904170424

Cited by 9 publications

(5 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If we assume that the electrical power consumed by the memory component of our device is 5% of that consumed by the processor (Acar et al, 2016), then the JANET will consume approximately 0.95 × 5 6 + 0.05 × 0.5 = 0.817 of the electrical power consumed by the LSTM. However, this ratio is a theoretical estimation and would be different in practice.…”

Section: Theoretic Computational Benefitsmentioning

confidence: 99%

The unreasonable effectiveness of the forget gate

van der Westhuizen,

Lasenby

2018

Preprint

View full text Add to dashboard Cite

Given the success of the gated recurrent unit, a natural question is whether all the gates of the long short-term memory (LSTM) network are necessary. Previous research has shown that the forget gate is one of the most important gates in the LSTM. Here we show that a forget-gate-only version of the LSTM with chronoinitialized biases, not only provides computational savings but outperforms the standard LSTM on multiple benchmark datasets and competes with some of the best contemporary models. Our proposed network, the JANET, achieves accuracies of 99% and 92.5% on the MNIST and pMNIST datasets, outperforming the standard LSTM which yields accuracies of 98.5% and 91%.

show abstract

Section: Theoretic Computational Benefitsmentioning

confidence: 99%

The unreasonable effectiveness of the forget gate

van der Westhuizen,

Lasenby

2018

Preprint

View full text Add to dashboard Cite

show abstract

“…Using the pyRAPL module, the indirect method is used in this experiment. In systems, CPUs have predominantly been seen to be the most power-hungry component (Acar et al, 2016;Salem, 2017). As a result, the focus was on determining the CPU's energy usage while executing the algorithms.…”

Section: Overhead Performance Measurementmentioning

confidence: 99%

Estimating Overhead Performance of Supervised Machine Learning Algorithms for Intrusion Detection

Baidoo

Yaokumah

Owusu

2023

International Journal of Information Technologies and Systems Approach

View full text Add to dashboard Cite

Estimating the energy and memory consumption of machine learning(ML) models for intrusion detection ensures efficient allocation of system resources. This study investigates the impact of supervised ML algorithms on the energy and memory consumption of intrusion detection systems. Experiments are conducted with seven ML algorithms and a proposed ensemble model, utilizing two intrusion detection datasets. Pearson correlation coefficient(PCC) and Spearman correlation coefficient are employed for the selection of optimum features. Regarding energy consumption, the findings reveal that the PCC with the UNSW-NB15 dataset uses the least amount of DRAM and CPU power. For ML methods, SVM utilizes the highest energy for both feature selection methods and datasets. Concerning memory consumption, the results show that decision tree uses the most current memory with PCC on the UNSW-NB15. The proposed ensemble model demonstrates the highest performance. These findings offer practical guidelines to ML experts when choosing the optimum model with the most efficient utilization of energy and memory.

show abstract

“…In terms of memory capacity of the sub-millimeter node, the current advances allow expecting a memory size of about a few dozens of MB [38]. On the other hand, the memory size impacts on the energy consumption of the nanonode (activation, pre-charging, refreshing, read, write) [39]. The consumption level of each of these tasks depends on the type of memory (DRAM, SRAM, etc.).…”

Section: Memory Capacity At Sub-millimeter Scalementioning

confidence: 99%

Efficient routing protocol for concave unstable terahertz nanonetworks

2020

View full text Add to dashboard Cite

The recent progress in nanotechnologies is giving birth to a novel topology of wireless networks characterized by a high local density and an intensive node instability such as in WBAN and swarm micro-robots systems. In this paper, we show that classical and dedicated ad hoc nanonetwork routing solutions are inefficient in this case and present a low reliability level and add a supplementary delay and control traffic. Majority of these solutions are based on point to point relaying mode, which is not adapted to the instability context. The multirelay to multirelay approaches allow countering the problem of nanonodes uncertainty by using the high number of inter-node connections. However, these approaches perform badly when the nanonetwork deployment presents distortions and concave sides. We propose a new routing protocol called Multirelay to Multirelay Routing Protocol (M2MRPv2), which provides a natural way to manage the residual energy levels on the nanonodes. M2MRPv2 is, to the best of our knowledge, the only approach that proposes a proactive multirelay to multirelay routing mode where the residual energy level of the nanonodes and reliability of the routing paths are taken into account. We study the performances of multirelay to multirelay protocols according to different multi-source to multi-sink communication scenarios. The obtained results show that M2MRPv2 protocol outperforms by far the Sustainable Longevity Routing (SLR) protocol (the reference protocol for Terahertz nanonetworks) in terms of transmission reliability and energy management. This outperformance is accentuated when the Terahertz nanonetwork deployment presents many concavities.

show abstract

Beyond CPU: Considering Memory Power Consumption of Software

Cited by 9 publications

References 14 publications

The unreasonable effectiveness of the forget gate

The unreasonable effectiveness of the forget gate

Estimating Overhead Performance of Supervised Machine Learning Algorithms for Intrusion Detection

Efficient routing protocol for concave unstable terahertz nanonetworks

Contact Info

Product

Resources

About