Cooling of a Many-Core Multiprocessor: Experimental Results for OLTP Workloads

2013 ◽  
Author(s):  
Krishnamachar Sreenivasan
Keyword(s):  
Thermal Conductivity ◽  
Heat Conduction ◽  
Heat Generation ◽  
Stationary Solutions ◽  
Main Memory ◽  
Isotropic Case ◽  
Published Data ◽  
Clock Frequency ◽  
The Core ◽  
Many Core

Stochastic heat conduction differential equation in spite of its complexity allows stationary solutions valid over a certain range of variables characterizing heat flow in multi-processor cores. Heat conduction equation is recast to account for anisotropy of a many core multi-processor in which heat generated at various locations depends on whether it is a cache, processor, bus controller, or memory controller: within the core generated heat depends on the hit rate, processor utilization, cache organization, and the technology used. Thermal conductivity of and heat generation in the core are treated as stochastic variables and influence of workloads, hitherto unrecognized, is explicitly accounted for in determining temperature distribution and its variation with processor clock frequency. Relationships derived from first principles indicate that rise in temperature with processor frequency for OLTP workload is not as catastrophic as predicted by some Industry brochures! A general framework for heat conduction in an orthotropic rectangular slab (representing a many core processor) with stochastic values of thermal conductivity and heat generation is developed; the theoretical trend is validated using published data for OLTP workloads to obtain temperature at the core surface as a function of clock frequency for the deterministic case. Transaction Processing Councils (TPC) openly available data from controlled, closely audited experiments for TPCC workloads during the period 2000–2011 were analyzed to determine the relation between throughput, clock frequency, main memory size, number of cores, and power consumed. Operating systems, compilers, linkers, processor architecture, cache, main memory, and storage sizes have changed drastically during this ten year period, not to mention hyper-threading was unknown in 2000! This analysis yields the following equations for throughput and power consumed, which for a specific case of 64 processors with a main memory of 32 GB and a million users, becomes W = 1075f0.22. For the isotropic case the temperature difference at the surface may be expressed for the case under study as ΔT = 71.1f0.22. This demonstrates that chip temperature for OLTP workloads does not increase to catastrophic values with increase in frequency. This behavior varies for other types of workloads.

scholarly journals Constructal Design of an Arrow-Shaped High Thermal Conductivity Channel in a Square Heat Generation Body

Entropy ◽  
2020 ◽  
Vol 22 (4) ◽  
pp. 475 ◽  
Author(s):  
Fengyin Zhang ◽  
Huijun Feng ◽  
Lingen Chen ◽  
Jiang You ◽  
Zhihui Xie
Keyword(s):  
Thermal Conductivity ◽  
Heat Conduction ◽  
Heat Generation ◽  
Degrees Of Freedom ◽  
High Thermal Conductivity ◽  
Maximum Temperature ◽  
Thermal Conductivity Ratio ◽  
Maximum Temperature Difference ◽  
Conduction Model ◽  
Three Degrees Of Freedom

A heat conduction model with an arrow-shaped high thermal conductivity channel (ASHTCC) in a square heat generation body (SHGB) is established in this paper. By taking the minimum maximum temperature difference (MMTD) as the optimization goal, constructal designs of the ASHTCC are conducted based on single, two, and three degrees of freedom optimizations under the condition of fixed ASHTCC material. The outcomes illustrate that the heat conduction performance (HCP) of the SHGB is better when the structure of the ASHTCC tends to be flat. Increasing the thermal conductivity ratio and area fraction of the ASHTCC material can improve the HCP of the SHGB. In the discussed numerical examples, the MMTD obtained by three degrees of freedom optimization are reduced by 8.42% and 4.40%, respectively, compared with those obtained by single and two degrees of freedom optimizations. Therefore, three degrees of freedom optimization can further improve the HCP of the SHGB. Compared the HCPs of the SHGBs with ASHTCC and the T-shaped one, the MMTD of the former is reduced by 13.0%. Thus, the structure of the ASHTCC is proven to be superior to that of the T-shaped one. The optimization results gained in this paper have reference values for the optimal structure designs for the heat dissipations of various electronic devices.

Heat Conduction in Multilayered Rectangular Domains

2007 ◽  
Vol 129 (4) ◽  
pp. 440-451 ◽  
Author(s):  
James Geer ◽  
Anand Desai ◽  
Bahgat Sammakia
Keyword(s):  
Thermal Conductivity ◽  
Heat Conduction ◽  
Heat Generation ◽  
Analytical Study ◽  
Three Dimensional ◽  
Thermal Boundary Conditions ◽  
Wide Range ◽  
Potential Applications ◽  
Steady State Heat ◽  
Spatially Varying

This paper presents the results of an analytical study of steady state heat conduction in multiple rectangular domains. Any finite number of domains that are equally sized (in plane) may be considered in the current analysis. The thermal conductivity and thickness of these domains may be different. The entire geometry composed of these connected domains is considered as adiabatic on the lateral surfaces and can be subjected to a wide range of thermal boundary conditions at the top and bottom. For example, the bottom of the stack may be adiabatic, while the top of the stack may be exposed to a uniform heat transfer coefficient. Spatially varying heat generation rates can be applied in each of the domains. The solutions are found to be in agreement with known solutions for simpler geometries. The analytical solution presented here is very general in that it takes into account the interface resistances between the layers. One application of this analytical study relates to the thermal management of three-dimensional stacks of computer devices and interconnect layers. The devices would have spatially nonuniform power dissipation within them, and the interconnect layers would have a significantly lower thermal conductivity than the devices. Interfacial defects, such as delamination or air voids, between the devices and the interconnect layers may be included in the model. Another possible application is to the study of hot spots in a chip stack with nonuniform heat generation. Many other potential applications may also be simulated.

scholarly journals Experimental Study on the Heat Exchange Mechanism in a Simulated Self-Circulation Wellbore

Energies ◽  
2020 ◽  
Vol 13 (11) ◽  
pp. 2918
Author(s):  
Liang Zhang ◽  
Songhe Geng ◽  
Jun Kang ◽  
Jiahao Chao ◽  
Linchao Yang ◽  
...  
Keyword(s):  
Thermal Conductivity ◽  
Heat Exchange ◽  
Convective Heat Transfer Coefficient ◽  
Injection Rate ◽  
Circulation Rate ◽  
Similarity Analysis ◽  
Published Data ◽  
Outlet Temperature ◽  
Hot Dry Rocks ◽  
Mining Capacity

Self-circulation wellbore is a new technique for geothermal development in hot dry rocks (HDR), which uses a U-shape channel composed of tubing and casing as the heat exchanger. In this study, a self-circulation wellbore in HDR on a laboratory scale was built, and a serial of experiments were conducted to investigate the heat exchange law and the influencing factors on the heat mining rate of the wellbore. A similarity analysis was also made to estimate the heat-mining capacity of the wellbore on a field scale. The experimental results show that the large thermal conductivity and heat capacity of granite with high temperature can contribute to a large heat-mining rate. A high injection rate can cause a high convective heat transfer coefficient in wellbore, while a balance is needed between the heat mining rate and the outlet temperature. An inner tubing with low thermal conductivity can significantly reduce the heat loss to the casing annulus. The similarity analysis indicates that a heat mining rate of 1.25 MW can be reached when using a 2000 m long horizontal well section in a 150 °C HDR reservoir with a circulation rate of 602.8 m3/day. This result is well corresponding to the published data.

Adiabatic Heat Flow in Mercury with a Fe-8.5wt%Si Core

2021 ◽  
Author(s):  
Meryem Berrada ◽  
Richard Secco ◽  
Wenjun Yong
Keyword(s):  
Thermal Conductivity ◽  
Heat Flow ◽  
Internal Structure ◽  
High Pressures ◽  
Inner Core ◽  
Thermal Evolution ◽  
The Core ◽  
Thermally Driven ◽  
Wire Method ◽  
Core Conditions

<p>Recent theoretical studies have tried to constrain Mercury’s internal structure and composition using thermal evolution models. The presence of a thermally stratified layer of Fe-S at the top of an Fe-Si core has been suggested, which implies a sub-adiabatic heat flow on the core side of the CMB. In this work, the adiabatic heat flow at the top of the core was estimated using the electronic component of thermal conductivity (k<sub>el</sub>), a lower bound for thermal conductivity. Direct measurements of electrical resistivity (ρ) of Fe-8.5wt%Si at core conditions can be related to k<sub>el</sub> using the Wiedemann-Franz law. Measurements were carried out in a 3000 ton multi-anvil press using a 4-wire method. The integrity of the samples at high pressures and temperatures was confirmed with electron-microprobe analysis of quenched samples at various conditions. Unexpected behaviour at low temperatures between 6-8 GPa may indicate an undocumented phase transition. Measurements of ρ at melting seem to remain constant at 127 µΩ·cm from 10-24 GPa, on both the solid and liquid side of the melting boundary. The adiabatic heat flow at the core side of Mercury’s core-mantle boundary is estimated between 21.8-29.5 mWm<sup>-2</sup>, considerably higher than most models of an Fe-S or Fe-Si core yet similar to models of an Fe core. Comparing these results with thermal evolution models suggests that Mercury’s dynamo remained thermally driven up to 0.08-0.22 Gyr, at which point the core became sub-adiabatic and stimulated a change from dominant thermal convection to dominant chemical convection arising from the growth of an inner core. Simply considering the internal structure of Mercury, these results support the capture of Mercury into a 3:2 resonance orbit during the thermally driven era of the dynamo.</p>

scholarly journals Redesigning Systems Thinking

2017 ◽  
Vol 10 (1) ◽  
Author(s):  
Birger Sevaldson
Keyword(s):  
Strategic Management ◽  
Systems Thinking ◽  
Design Thinking ◽  
Design Strategies ◽  
Working Paper ◽  
The Core ◽  
Core Competences ◽  
Core Concepts ◽  
Many Core ◽  
Systemic Design

The resent movement of Systemic Design seeks for new synergies between Design and Systems. While the usefulness of systems approaches in design has been fairly obvious, this paper argues that many core concepts in design are beneficial in systems thinking. This seems reasonable when it comes to the concept of Design Thinking. However, as this paper argues, the more practical core concepts of design are equally important. Designerly skills have been regarded as belonging mainly in the realm of traditional commercial design, whereas design thinking has been regarded as useful in strategic management settings. This paper argues against the idea of separating design thinking from design action. The skills and competences of design, such as the composition of the shape and form that are obvious in product design, are central to Systems Oriented Design (SOD). SOD is a version in the emerging pluralistic field of Systemic Design. The Systemic Design movement should recognise the core values of design and integrate them in systems thinking. This integration would contribute to innovation in both Systemic Design and systems thinking. Among the core competences of design discussed in the paper are composition, choreography, orchestration, the notion of the Gesamtkunstwerk and open-ended multi-scalar design strategies that allow for both structural and organic development. The paper provides examples to support its proposal for the use of concrete aesthetic principles to guide Systemic Design processes. This paper expands the working paper entitled “Holistic and dynamic concepts in design: What design brings to systems thinking”, which was presented at the RSD3 symposium (2014). 

Thermal and Electrical Conductivities of Sodium from 40 to 360 K

1972 ◽  
Vol 50 (12) ◽  
pp. 1386-1401 ◽  
Author(s):  
J. G. Cook ◽  
M. P. Van der Meer ◽  
M. J. Laubitz
Keyword(s):  
Thermal Conductivity ◽  
Electrical Resistivity ◽  
Low Temperatures ◽  
Characteristic Temperature ◽  
Published Data ◽  
Inelastic Electron ◽  
Conductivity Data ◽  
Electron Collisions ◽  
Electrical Conductivities ◽  
Flow Apparatus

We present data on the electrical and thermal resistivities and the thermopower of three pure Na specimens from 40 to 360 K. The measurements were made using a guarded longitudinal heat flow apparatus that had previously been calibrated with Au and Al. The specimens were placed in a vacuum environment using no solid inert liner.The electrical resistivity data indicate ΘR = 194 K. The thermal conductivity data show a 4% minimum near 70 K and an ice point value of 1.420 W/cm K. The reduced Lorenz function L/L0 agrees with published data at low temperatures but above 300 K levels off at approximately 0.91. On the basis of published data for liquid Na, L/L0 does not change by more than 3% at the melting point.The minimum in the thermal conductivity and a part of the high temperature deviations of L from L0 are tentatively ascribed to inelastic electron–phonon collisions having a characteristic temperature near that of longitudinal phonons. The possibility that electron–electron collisions further depress L at high temperatures is critically examined.

Sign in / Sign up

Export Citation Format

Share Document