Modeling and dynamic management of 3D multicore systems with liquid cooling

Coskun, Ayse K.; Ayala, José L.; Atienza, David; Rosing, Tajana

doi:10.1109/vlsisoc.2009.6041327

Cited by 58 publications

(57 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Removing the heat using fan from a heat sink attached to a package (integrated circuit) and Dynamic Frequency Voltage Scaling (DFVS) are considered as a conventional thermal management [1,3]. The temperature of an element (chip, heat sink, etc.)…”

Section: Introductionmentioning

confidence: 99%

Improvement of microsystem throughput using new cooling system

Samake

Kocanda

Kos

2016

ZN PRz Elektrotechnika

View full text Add to dashboard Cite

This paper presents a new possibility of clock frequency/voltage control in microsystems i.e. high performance processors, exploiting information about cooling efficiency. In this paper, we propose an approach that better exploits the thermal abilities of a chip fixed to cooling system in order to eliminate its energy accumulation. For the purpose of the proposed method, the calculation of so called time shift (TS) is introduced. TS is defined as the duration where the computational system can perform the task at higher frequency without any thermal violation when the chip temperature is close to critical thermal threshold. The analogy between thermal and electrical parameters allows to model RC thermal compact model of structure (chip fixed to the cooling system). Based on this assumption, the authors compute the TS value versus different parameters using RC thermal compact model in Spice environment. The results indicate that TS could fulfil a significant part of die total working time. As an effect the proposed approach may be a means for increasing average clock frequency or voltage supply, consequently enhancing the system's throughput.

show abstract

Section: Introductionmentioning

confidence: 99%

Improvement of microsystem throughput using new cooling system

Samake

Kocanda

Kos

2016

ZN PRz Elektrotechnika

View full text Add to dashboard Cite

show abstract

“…Another class of applications which require runtime thermal simulation is fine grain control: where thermal data is required at a finer granularity than can be provided by thermal sensors. Examples include register assignment optimization [3] and liquid coolant flow control in 3d chips [4].…”

Section: Introductionmentioning

confidence: 99%

Neural network based on-chip thermal simulator

Kumar

Atienza

2010

Proceedings of 2010 IEEE International Symposium on Circuits and Systems

Self Cite

View full text Add to dashboard Cite

Abstract-With increasing power densities, runtime thermal management is becoming a necessity in today's systems, especially so for highly integrated Multi-Processor Systems-on-Chip (MPSoCs). In this paper, we propose a neural network (NN) based approach to implement an on-chip thermal simulator to aid such runtime management for MPSoCs. The proposed method combines the advantage of approximating the thermal properties of the chip as a linear system with the ease of fully parallel analog implementation of NNs. We perform a case study with the Niagara UltraSPARC T1 MPSoC for real-life applications, benchmarking our results with an accurate higher order Runge-Kutta (RK4) solver, that is employed in tools such as HotSpot. Within a few gate delays, the proposed NN design can simulate temperatures of the MPSoC 500 ms into the futurecorresponding to thousands of iterations of the RK4 solver, with a maximum error of 1-2 K.

show abstract

“…3) Thermal Management of Liquid-Cooled 3D MPSoCs Prior liquid cooling work [90] evaluates existing thermal management policies on a 3D MPSoC with a fixedflow rate setting, and also investigates the benefits of variable flow using a policy to increment or decrement the flow rate based on temperature measurements, but without considering pump energy consumption.…”

Section: ) Power and Thermal Management Of Air-cooled 2d And 3d Mpsocsmentioning

confidence: 99%