Replica

Fernando, Vimuth; Franques, Antonio; Abadal, Sergi; Misailovíc, Saša; Torrellas, Josep

doi:10.1145/3297858.3304033

Cited by 31 publications

(7 citation statements)

References 67 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Approximate communication is considered to be one of the most effective approaches to improve network performance when an application can tolerate errors [22,24,25,26,27,28,29,30,31,32]; with a reduced accuracy during communication, the approximation techniques significantly reduce the network latency and power consumption for on-chip communication.…”

Section: Approximate Communication Techniquesmentioning

confidence: 99%

“…Fig. 2 shows an approximate communication NoC [22,24,25,26,27,28,29,30,31,32] implemented in a heterogeneous multi-core system [7,8,9,10,11,12,13] with an L2 shared cache for CNN inference. The data approximation module in the network interface reduces the packet size by truncation or lossy compression according to the approximation information.…”

Section: Approximate Communication Techniquesmentioning

confidence: 99%

“…• Existing approximate communication techniques [22,24,25,26,27,28,29,30,31,32] reduce communication latency and power consumption by utilizing packet approximation in NoCs. However, existing techniques only rely on the relative error for data approximation.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Technique for Approximate Communication in Network-on-Chips for Image Classification

Chen¹,

Liu²,

Lombardi³

et al. 2021

Preprint

View full text Add to dashboard Cite

Approximation is an effective technique for reducing power consumption and latency of on-chip communication in many computing applications. However, existing approximation techniques either achieve modest improvements in these metrics or require retraining after approximation, such when convolutional neural networks (CNNs) are employed. Since classifying many images introduces intensive on-chip communication, reductions in both network latency and power consumption are highly desired. In this paper, we propose an approximate communication technique (ACT) to improve the efficiency of on-chip communications for image classification applications. The proposed technique exploits the error-tolerance of the image classification process to reduce power consumption and latency of on-chip communications, resulting in better overall performance for image classification computation. This is achieved by incorporating novel quality control and data approximation mechanisms that reduce the packet size. In particular, the proposed quality control mechanisms identify the error-resilient variables and automatically adjust the error thresholds of the variables based on the image classification accuracy. The proposed data approximation mechanisms significantly reduce packet size when the variables are transmitted. The proposed technique reduces the number of flits in each data packet as well as the on-chip communication, while maintaining an excellent image classification accuracy. The cycle-accurate simulation results show that ACT achieves 23% in network latency reduction and 24% in dynamic power reduction compared to the existing approximate communication technique with less than 0.99% classification accuracy loss.

show abstract

Section: Approximate Communication Techniquesmentioning

confidence: 99%

Section: Approximate Communication Techniquesmentioning

confidence: 99%

See 1 more Smart Citation

A Technique for Approximate Communication in Network-on-Chips for Image Classification

Chen¹,

Liu²,

Lombardi³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…1 that is integrated with the wireless transceiver to provide wireless communication for each core [3]. Considering RF research, applicable transceivers can be achieved by down-scaling power and area characteristics along with communication frequency [45,46]. Studies show that quadratically reduced passive elements can…”

Section: The Network Modelmentioning

confidence: 99%

“…Authors in [46] proposed a wireless many-core architecture called Replica, for speed-up communication-intensive ordinary data. In Replica, a dynamic MAC protocol has been introduced that can dynamically switch between two different schemes, i.e., carrier sensing and token passing based on the wireless network utilization pattern which could vary across different applications and also within an application.…”

Section: The Network Modelmentioning

confidence: 99%

Traffic-aware performance optimization in Real-time wireless network on chip

Baharloo

Khonsari

Dolati

et al. 2020

Nano Communication Networks

View full text Add to dashboard Cite

Network on Chip (NoC) is a prevailing communication platform for multi-core embedded systems. Wireless network on chip (WNoC) employs wired and wireless technologies simultaneously to improve the performance and power-efficiency of traditional NoCs. In this paper, we propose a deterministic and scalable arbitration mechanism for the medium access control in the wireless plane and present its analytical worst-case delay model in a certain use-case scenario that considers both Real-time (RT) and Non Real-time (NRT) flows with different packet sizes. Furthermore, we design an optimization model to jointly consider the worst-case and the average-case performance parameters of the system. The Optimization technique determines how NRT flows are allowed to use the wireless plane in a way that all RT flows meet their deadlines, and the average case delay of the WNoC is minimized. Results show that our proposed approach decreases the average latency of network flows up to 17.9%, and 11.5% in 5 × 5, and 6 × 6 mesh sizes, respectively.

show abstract