Rack Liquid Cooling Pump Speed Control for Performance-Per-Watt

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional air-cooling solutions struggle to manage thermal challenges in high power density GPU racks, particularly in data centers, and fail to optimize the interaction between processor temperature and power consumption, leading to suboptimal performance and energy efficiency.

Innovation Solution

A liquid-cooled electronic rack system with a coolant distribution unit and rack management unit that uses control logic to determine an optimal pump speed for the liquid pump, optimizing performance-per-watt by balancing processor power consumption, cooling power consumption, and benchmark performance measurements, while maintaining processor temperatures within a predetermined range.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Temperature

If conventional air-cooling solutions are used for high power density GPU racks, then the cooling system is simpler to implement, but the thermal management performance is insufficient and power consumption is high

Engineering Contradiction:

Improveprocessor temperatureVSAvoidpower consumption

Core Design Contradiction:

TemperatureVSUse of energy by moving object

Solution Approach 1:

The patent transitions from air-cooling to liquid-cooling technology, using coolant circulation through cold plates to remove heat from high power density processors. The liquid cooling system includes pumps, cold plates, and coolant flow management to achieve superior thermal management compared to conventional air-cooling methods.

Inventive Principle:

Principle #29Pneumatics and hydraulics

2Use of energy by stationary object

If cooling water temperature is increased to improve power efficiency, then cooling power consumption is reduced, but processor temperature control becomes less effective

Engineering Contradiction:

Improvecooling power consumptionVSAvoidprocessor temperature

Core Design Contradiction:

Use of energy by stationary objectVSTemperature

Solution Approach 1:

The patent implements dynamic control of cooling parameters including variable speed pumps and adjustable coolant flow rates. The system continuously monitors processor temperatures and cooling efficiency to dynamically optimize the balance between cooling performance and power consumption, rather than using fixed temperature settings.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent incorporates feedback mechanisms that monitor processor temperatures, coolant temperatures, and power consumption metrics. This feedback is used to adjust cooling system operations in real-time, optimizing the trade-off between cooling effectiveness and energy efficiency based on actual system conditions.

Inventive Principle:

Principle #23Feedback

3Productivity

If processor temperature is reduced to improve performance, then computing performance increases, but cooling power consumption increases

Engineering Contradiction:

Improvecomputing performanceVSAvoidcooling power consumption

Core Design Contradiction:

ProductivityVSUse of energy by stationary object

Solution Approach 1:

The patent optimizes multiple parameters including coolant flow rate, coolant temperature, pump speed, and cold plate configuration to achieve the optimal balance between processor temperature control and cooling power consumption. These parameter changes are made dynamically based on workload conditions and performance requirements.

Inventive Principle:

Principle #35Parameter changes

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach enhances energy efficiency by optimizing the performance-per-watt of the liquid-cooled system, reducing power consumption, and improving computing performance in machine learning applications by determining the optimal pump speed for the liquid cooling system.

Implementation Method 1

a coolant distribution unit (CDU) to supply cooling liquid to the processors and to receive the cooling liquid carrying the heat exchanged from the processors

Methodology Applied

Scientific EffectHeat exchange: Heat Exchanger

Implementation Method 2

The CDU includes a liquid pump to pump the cooling liquid

Methodology Applied

Scientific EffectPumping: Pump

Data Source

PatentUS11212946B2Performance-per-watt optimal control logic in liquid cooling solution for high performance machine-learning computing

Publication Date: 2021.12.28 BAIDU USA LLC

AI summary

An electronic rack includes an array of server blades arranged in a stack. Each server blade contains one or more servers and each server includes one or more processors to provide data processing services. The electronic rack further includes a coolant distribution unit (CDU) and a rack management unit (RMU). The CDU is configured to supply cooling liquid to the processors and to receive the cooling liquid carrying the heat exchanged from the processors. The CDU includes a liquid pump to pump the cooling liquid and a pump controller to control a pump speed of the liquid pump. The RMU is configured to manage the operations of the components within the electronic rack such as CDU, etc. The RMU includes control logic to determine an optimal pump speed of the liquid pump by optimizing an objective function based on processor power consumption of the processors, cooling power consumption of the CDU, and a benchmark performance measurement of the data processing services and to control the pump speed of the liquid pump based on the optimal pump speed.