Server Temperature Control via Parallel Component Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing server temperature control methods are inefficient and time-consuming, leading to potential overheating and reduced reliability due to the long time required to read temperature data from components, especially graphics processors, which can result in higher fan noise, energy consumption, and shortened fan life.
Innovation Solution
A method involving a baseboard management controller connected to a parallel processing device that determines component types and reads temperature data in parallel, prioritizing graphics processors for timely control, thereby reducing overall reading time and enhancing precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If a BMC polls PCIe cards sequentially via I2C buses to read temperature data, then the system structure remains simple, but the time consumption for temperature data acquisition becomes excessively long
Solution Approach 1:
The patent introduces a parallel processing device that segments the temperature data acquisition task into multiple independent reading operations. Instead of one BMC sequentially polling all PCIe cards, the parallel processing device divides the work by simultaneously reading temperature data from multiple PCIe cards through multiple I2C buses, thereby reducing total acquisition time while maintaining manageable system complexity through modular architecture
Solution Approach 2:
The parallel processing device acts as an intermediary between the BMC and the PCIe cards. It receives polling requests from the BMC, then simultaneously queries multiple PCIe cards through its own I2C interface, and finally returns the collected temperature data to the BMC. This intermediary approach accelerates data acquisition without requiring the BMC itself to become more complex
2Reliability
If the BMC polls all PCIe cards sequentially, then the polling process is simple to implement, but high-temperature-sensitivity components experience delayed temperature control response
Solution Approach 1:
The parallel processing device performs preliminary temperature data collection from all PCIe cards simultaneously before the BMC needs to make cooling decisions. By pre-acquiring all temperature data in parallel and making it available to the BMC at once, the system enables immediate temperature control responses for high-temperature-sensitivity components, improving reliability without implementation complexity
Solution Approach 2:
The system implements periodic temperature monitoring where the parallel processing device continuously collects temperature data from all PCIe cards at regular intervals. This periodic parallel sampling ensures that temperature data is always current and available for timely control actions, enhancing server reliability through consistent monitoring
3Temperature
If sequential polling is used to read temperature data, then the system architecture remains straightforward, but components with high temperature sensitivity experience rapid temperature rises
Solution Approach 1:
The parallel processing device segments temperature monitoring into dedicated functional units, each capable of independently reading temperature data from PCIe cards. This segmentation allows simultaneous monitoring of multiple components with high temperature sensitivity, providing precise temperature control data without requiring complex integration logic in the BMC
Solution Approach 2:
The parallel processing device is designed as a universal platform that can simultaneously handle multiple I2C bus connections and communicate with various types of PCIe cards. It provides multi-functional capability to read temperature data from different components in parallel, enabling precise temperature control across all components while maintaining a single standardized interface to the BMC
4Productivity
If sequential temperature data acquisition is used, then energy consumption is lower per operation, but the overall system efficiency and productivity are reduced
Solution Approach 1:
The parallel processing device performs preliminary bulk data collection from all PCIe cards in a single operational cycle. By gathering all temperature data simultaneously before the BMC processes cooling decisions, the system maximizes monitoring efficiency and minimizes the time the BMC spends on polling operations, thereby improving productivity while managing overall energy consumption
Data Source
AI summary
The present application discloses a method, system and device for controlling temperature of a server, and a storage medium, applied in the technical field of server controlling, which solves the problem in conventional solutions that the temperature controlling over servers is not ideal. The method includes: after the server has been powered on, by a parallel processing device, determining respective component types of all of components of a quantity N and sending the respective component types to a baseboard management controller; at a first stage of each of parameter-reading periods, reading simultaneously respective temperature data of the components of the quantity N by using threads of the quantity N of the parallel processing device in a parallel-reading mode; at a second stage, sending the temperature data of the components the component types of which are a graphics processor to the baseboard management controller; and at a third stage, sending the temperature data of the components the component types of which are not a graphics processor to the baseboard management controller, whereby the baseboard management controller controls the temperature of the server based on the respective temperature data of the components of the quantity N. By applying the solutions of the present application, the temperature controlling over the server may be realized more precisely and effectively.