NPU Thermal Management

Amrouch, Hussam; Zervakis, Georgios; Salamin, Sami; Kattan, Hammam; Anagnostopoulos, Iraklis; Henkel, Jörg

doi:10.1109/tcad.2020.3012753

Cited by 39 publications

(30 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Superlattice TEC devices feature a very low internal resistance compared to bulk TEC devices. Hence, compared to our work, the used TEC in [20] has a larger heatsink that prevents Joule heat effects from becoming dominant at relatively low currents as well as it features a smaller internal resistance that leads, in general, to lower Joule heat effects. Consequently, the superlattice TEC in [20] was able to operate up to around 7A.…”

Section: B Imapct Of Our Proposed Tec-based Cooling On the Figures Of Merit Of Dramsmentioning

confidence: 73%

“…The used TEC in [19] was a TEC that has a surface area of 15 × 15mm 2 and a thickness of 3.22mm covering the entire processor chip. In [20], the effectiveness of employing ultra-thin film TEC in managing the temperature of Neural Processing Units (NPUs) was recently presented.…”

Section: A Thermoelectric Cooling (Tec)mentioning

confidence: 99%

“…Consequently, the TEC device in [19] could operate only up to around 1A. On the other hand, the work in [20] employs a superlattice TEC and assumes a much larger heat sink to be attached on top of the chip. Superlattice TEC devices feature a very low internal resistance compared to bulk TEC devices.…”

Section: B Imapct Of Our Proposed Tec-based Cooling On the Figures Of Merit Of Dramsmentioning

confidence: 99%

See 2 more Smart Citations

Longevity of Commodity DRAMs in Harsh Environments Through Thermoelectric Cooling

et al. 2021

Self Cite

View full text Add to dashboard Cite

Today, more and more commodity hardware devices are used in safety-critical applications, such as advanced driver assistance systems in automotive. These applications demand very high reliability of electronic components even in adverse environmental conditions, such as high temperatures. Ensuring the reliability of microelectronic components is a major challenge at these high temperatures. The computing systems of these applications rely on DRAMs as working memory, which are built upon bit cells that store charges in capacitors. These commodity DRAMs are optimized for cost per bit and not for high reliability. Thus, very high temperatures impose an enormous challenge for commodity DRAMs as the data retention time and reliability decrease largely, affecting the data correctness. Data correctness can be ensured up to certain temperatures by increasing the refresh rate to counterbalance the retention time reduction. However, this severely degrades the access latencies and the usable DRAM bandwidth. To overcome these limitations, we present for the first time a Thermoelectric Cooling (TEC) solution for commodity DRAMs in harsh-environments, such as automotive. Our TEC solution enables the use of commodity off-the-shelf DRAMs in safety-critical applications by reducing the temperature conditions to a range where they can operate reliably. This TEC solution is applied a posteriori to the DRAM chips without using high-cost package solutions. Thus, it maintains the low-cost targets of such devices, improves the reliability, and at the same time, counterbalances the adverse effects of increasing the refresh rate. To quantitatively evaluate the benefits of TEC on commodity DRAMs in harsh-environments, we performed system-level evaluations with several applications backed up by the measured data on commodity DRAMs. Our experimental results, using accurate multi-physics simulations that employ finite element method, demonstrate that the TECbased cooling ensures that the maximum temperature of all DRAM chips is always below 85°C despite that the original on-chip temperature (i.e., in the absence of our TEC based cooling) goes beyond 120°C.

show abstract

Section: B Imapct Of Our Proposed Tec-based Cooling On the Figures Of Merit Of Dramsmentioning

confidence: 73%

Section: A Thermoelectric Cooling (Tec)mentioning

confidence: 99%

Section: B Imapct Of Our Proposed Tec-based Cooling On the Figures Of Merit Of Dramsmentioning

confidence: 99%

See 1 more Smart Citation

Longevity of Commodity DRAMs in Harsh Environments Through Thermoelectric Cooling

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…Note that this evaluation targets a specific NN accelerator microarchitecture (i.e., TPU). However, in order to provide a generic evaluation, we focus on investigating the impact of NC-FinFET on the MADD unit that is a core component of any NN accelerator independently of its microarchitecture [4], [5], [17], [19], [30]- [33].…”

Section: A Neural Processing Unit Use Casementioning

confidence: 99%

“…For example, existing NN architectures trade throughput (e.g., [5] combines many MADD units to enable higher computational precision) or speed (e.g., [30] uses 10bit MADD units that are 1.15x slower than the 8-bit ones) to achieve higher inference accuracy. Similarly, [19], [32], [33] apply approximations and trade accuracy to improve the speed and/or energy consumption. As a result, NC-FinFET provides new insights and new directions to future NN accelerators architects as well as NN developers.…”

Section: Neural Network Inference Evaluationmentioning

confidence: 99%

Impact of NCFET on Neural Network Accelerators

et al. 2021

Self Cite

View full text Add to dashboard Cite

This is the first work to investigate the impact that Negative Capacitance Field-Effect Transistor (NCFET) brings on the efficiency and accuracy of future Neural Networks (NN). NCFET is at the forefront of emerging technologies, especially after it has become compatible with the existing fabrication process of CMOS. Neural Network inference accelerators are becoming ubiquitous in modern SoCs and there is an ever-increasing demand for tighter and tighter throughput constraints and lower energy consumption. To explore the benefits that NCFET brings to NN inference regarding frequency, energy, and accuracy, we investigate different configurations of the multiply-add (MADD) circuit, which is the core computational unit in any NN accelerator. We demonstrate that, compared to the baseline 7nm FinFET technology, its negative capacitance counterpart reduces the energy by 55%, without any frequency reduction. In addition, it enables leveraging higher computational precision, which results to a considerable improvement in the inference accuracy. Importantly, the achieved accuracy improvement comes also together with a significant energy reduction and without any loss in frequency.

show abstract

Thermal Management in Neuromorphic Materials, Devices, and Networks

2023

View full text Add to dashboard Cite

Machine learning has experienced unprecedented growth in recent years, often referred to as an “artificial intelligence revolution.” Biological systems inspire the fundamental approach for this new computing paradigm: using neural networks to classify large amounts of data into sorting categories. Current machine‐learning schemes implement simulated neurons and synapses on standard computers based on a von Neumann architecture. This approach is inefficient in energy consumption, and thermal management, motivating the search for hardware‐based systems that imitate the brain. Here, the present state of thermal management of neuromorphic computing technology and the challenges and opportunities of the energy‐efficient implementation of neuromorphic devices are considered. The main features of brain‐inspired computing and quantum materials for implementing neuromorphic devices are briefly described, the brain criticality and resistive switching‐based neuromorphic devices are discussed, the energy and electrical considerations for spiking‐based computation are presented, the fundamental features of the brain's thermal regulation are addressed, the physical mechanisms for thermal management and thermoelectric control of materials and neuromorphic devices are analyzed, and challenges and new avenues for implementing energy‐efficient computing are described.

show abstract

NPU Thermal Management

Cited by 39 publications

References 17 publications

Longevity of Commodity DRAMs in Harsh Environments Through Thermoelectric Cooling

Longevity of Commodity DRAMs in Harsh Environments Through Thermoelectric Cooling

Impact of NCFET on Neural Network Accelerators

Thermal Management in Neuromorphic Materials, Devices, and Networks

Contact Info

Product

Resources

About