CN114115503A

CN114115503A - System and method for automatically adjusting CPU voltage alarm threshold

Info

Publication number: CN114115503A
Application number: CN202111266961.3A
Authority: CN
Inventors: 李大利
Original assignee: Suzhou Inspur Intelligent Technology Co Ltd
Current assignee: Suzhou Inspur Intelligent Technology Co Ltd
Priority date: 2021-10-28
Filing date: 2021-10-28
Publication date: 2022-03-01
Anticipated expiration: 2041-10-28
Also published as: CN114115503B

Abstract

The invention has proposed a CPU voltage alarm threshold value automatic regulating system and method, this system includes BMC, PCH, CPU and VR; the BMC is in communication connection with the PCH; the PCH is in communication connection with the CPU; the CPU is in communication connection with a voltage output end of the VR; the BMC is also in communication connection with an enabling end of the VR; the VR receives a voltage regulation instruction sent by the BMC to supply power to the CPU; the PCH identifies the type of the CPU through the inherent code of the CPU and feeds back the type information to the BMC; after receiving the model information, the BMC calls the pre-stored information of all CPUs according to the model information and then automatically adjusts the CPU voltage alarm threshold. Based on the system, a method for automatically adjusting the CPU voltage alarm threshold is also provided. The invention avoids the problem that the voltage monitoring data of the BMC is not accurate enough or misinformation, and improves the accuracy of fault diagnosis and the operation reliability of the server.

Description

System and method for automatically adjusting CPU voltage alarm threshold

Technical Field

The invention belongs to the technical field of server voltage monitoring, and particularly relates to a system and a method for automatically adjusting a CPU voltage alarm threshold.

Background

The BMC voltage alarm is an important component of the server system fault diagnosis system. When the server runs, the BMC samples the output voltage of each VR of the mainboard in real time, compares and judges the output voltage with a voltage alarm threshold value prestored in the BMC, judges the voltage state to be abnormal if the voltage value obtained by sampling exceeds the threshold value setting, sends alarm information to the system and records the alarm information in a BMC log. In all BMC voltage alarms, the CPU voltage is the most important item, and it is very important how to ensure that the BMC can accurately determine whether the voltage state of the CPU is normal. The actual voltage value of the CPU is affected by the Load Line value and the maximum current, and the Load Line value and the maximum current of the CPU with different powers are different. On one hand, the current designed CPU voltage alarm threshold is to set a high-low threshold according to the maximum value and the minimum value of VID voltage of the CPU multiplied by a certain margin, and the influence of different Load Line values and maximum currents of CPUs with different powers is not considered, wherein VID: the core of the VID technology is that a plurality of VID pins are added on a CPU, coding signals output by the VID pins control a PWM controller in a Vcore power supply circuit, and different codes represent different Voltage values. On the other hand, the CPU VR is designed to set the Load Line value according to the maximum configuration, the CPU of different models has different requirements for the Load Line value, and the actual voltage value of the CPU is Vcore (VID-I) Load Line value. The actual configuration of the client is various, the CPU power is different, the Load Line value and the maximum current of the CPU with different power are different, and the maximum value and the minimum value of the actual voltage are different. The current design of the fixed voltage alarm threshold and the fixed VR Load Line value can cause the problem of false alarm of BMC.

Fig. 1 is a schematic diagram of a prior art solution. The BMC samples the output voltage of the CPU VR through the PMBUS in real time and compares and judges the output voltage with a voltage alarm threshold value prestored in the BMC. If the voltage value obtained by sampling exceeds the threshold setting, the BMC judges the voltage state to be abnormal, sends out alarm information to the system and records the alarm information into a BMC log. In the current technical scheme, a voltage value threshold is set to be 6 voltage levels, and a high-low value threshold coefficient is as follows: non-severe +/-A, severe +/-B, irreversible +/-C, e.g. non-severe +/-7%, severe +/-10%, irreversible +/-13%, e.g. as specified in the following table:

high non-severe	High severity	High irreversible capacity
			VID MAX*(1+A)	VID MAX*(1+B)	VID MAX*(1+C)

Low non-severity	Low severity	Low irreversible capacity
			VID MIN*(1-A)	VID MIN*(1-B)	VID MIN*(1-C)

In the prior art, the actual voltage value of the CPU is affected by the Load Line value and the maximum current, and the Load Line value and the maximum current of the CPUs with different powers are different. In the current technical scheme, the CPU voltage alarm threshold is set according to the maximum value and the minimum value of VID voltage of the CPU multiplied by a certain margin, the Load Line value and the maximum current of the CPUs with different powers are not considered to be different, and the design of the CPU VR is to set the Load Line value according to the maximum configuration. The CPU voltage alarm threshold should be different according to the CPU model, and in the current technical solution, the design of the fixed voltage alarm threshold may cause the problem of false alarm of the BMC.

Disclosure of Invention

In order to solve the technical problems, the invention provides a system and a method for automatically adjusting the CPU voltage alarm threshold, so that the system can be self-adapted to different configurations, the problem that BMC voltage monitoring data is not accurate enough or is misreported is avoided, and the accuracy of fault diagnosis and the operation reliability of a server are improved.

In order to achieve the purpose, the invention adopts the following technical scheme:

a system for automatic adjustment of CPU voltage alarm thresholds, comprising: the system comprises a substrate management controller, a platform control center, a central processing unit and a power supply module;

the substrate management controller is in communication connection with the platform control center; the platform control center is in communication connection with the central processing unit; the central processing unit is in communication connection with the voltage output end of the power supply module; the substrate management controller is also in communication connection with an enabling end of the power supply module;

the power supply module receives a voltage regulation instruction sent by the substrate management controller and supplies power to the central processing unit; the platform control center identifies the model of the central processing unit through the inherent code of the central processing unit and feeds back the model information to the substrate management controller; and after receiving the model information, the substrate management controller calls the pre-stored information of all central processing units according to the model information and then automatically adjusts the voltage alarm threshold of the central processing units.

Further, the communication connection between the baseboard management controller and the platform control center specifically includes:

the base plate management controller is in communication connection with the platform control center through an LPC bus and is used for obtaining model information of the central processing unit.

Further, the platform control center is in communication connection with the central processing unit, specifically:

the platform control center is in communication connection with the central processing unit through the DMI bus, and is used for acquiring the model information of the central processing unit and feeding the model information back to the substrate management controller.

Further, the central processing unit is in communication connection with a voltage output end of the power supply module; the substrate management controller is further in communication connection with an enabling end of the power supply module, and specifically comprises the following steps:

the central processing unit is communicated with the power supply module through an SVID bus; the power supply module receives a voltage regulation instruction of the central processing unit through the SVID bus and supplies power to the central processing unit; the baseboard management controller is also in communication connection with an enabling end of the power supply module through a PMBUS.

Further, the information of all the central processing units pre-stored in the substrate management controller includes the model of each central processing unit, the maximum current corresponding to each model of central processing unit, the minimum voltage identification value, the maximum voltage identification value and the load line standard.

Further, after receiving the model information, the baseboard management controller calls pre-stored information of all central processing units according to the model information, and then automatically adjusting the voltage alarm threshold of the central processing unit comprises the following steps:

after receiving the model information of the central processing unit, the substrate management controller writes the standard value of the power supply load line corresponding to the currently configured central processing unit into a standard value register of the power supply load line according to the model information;

the central processing unit adjusts the identification voltage of the power supply module through the SVID bus to obtain the actual output voltage of the power supply module;

and the baseboard management controller reads the standard value of the power supply load line in the register, performs data sampling on the actual output voltage and adds an anti-shake strategy.

Further, the central processing unit adjusts the identification voltage of the power supply module through the SVID bus, and the calculation process of obtaining the actual output voltage of the power supply module is as follows:

the actual output voltage Vcore is VID-I Load Line value; the VID is the identification voltage of the power supply module; the I is the current of the central processing unit; and the Load Line value is a standard value of a power supply Load Line.

Further, the process of sampling the actual output voltage and adding the anti-shake strategy is as follows:

and continuously acquiring the times that the actual output voltage is greater than the set voltage threshold value and exceeds the preset times, alarming, and removing the alarm if the voltage is recovered to be normal.

The invention also provides a method for automatically adjusting the CPU voltage alarm threshold, which is realized based on a system for automatically adjusting the CPU voltage alarm threshold and comprises the following steps:

the platform control center is communicated with the central processing unit, obtains the model information of the central processing unit and feeds back the model information to the substrate management controller;

and the substrate management controller calls the pre-stored information of all central processing units according to the model information, reads the actual output voltage value of the power supply module through the PMBUS and monitors whether the actual voltage value exceeds a voltage alarm threshold value or not.

Further, the method further comprises the step that the substrate management controller records the monitoring voltage exceeding the voltage alarm threshold value to a log file so as to facilitate the next fault diagnosis.

The effect provided in the summary of the invention is only the effect of the embodiment, not all the effects of the invention, and one of the above technical solutions has the following advantages or beneficial effects:

the invention provides a system and a method for automatically adjusting a CPU voltage alarm threshold, wherein the system comprises a substrate management controller, a platform control center, a central processing unit and a power supply module; the substrate management controller is in communication connection with the platform control center; the platform control center is in communication connection with the central processing unit; the central processing unit is in communication connection with the voltage output end of the power supply module; the substrate management controller is also in communication connection with an enabling end of the power supply module; the power supply module receives a voltage regulation instruction sent by the substrate management controller to supply power to the central processing unit; the platform control center identifies the model of the central processing unit through the inherent code of the central processing unit and feeds back the model information to the substrate management controller; after receiving the model information, the substrate management controller calls the pre-stored information of all central processing units according to the model information and then automatically adjusts the voltage alarm threshold of the central processing units. Based on a system for automatically adjusting the CPU voltage alarm threshold, the method for automatically adjusting the CPU voltage alarm threshold is also provided. The invention detects the code of the central processing unit through the substrate management controller, acquires the model of the central processing unit, automatically adjusts the voltage alarm threshold value of the central processing unit according to the data of the prestored current, the load line value and the like of the substrate management controller, and adjusts the load line value of the power supply module, so that the system can adapt to different configurations, the problem of inaccurate or false alarm of the voltage monitoring data of the substrate management controller is avoided, and the accuracy of fault diagnosis and the operation reliability of the server are improved.

Drawings

Fig. 1 is a schematic diagram illustrating a connection between a baseboard management controller and a power supply module in the prior art;

fig. 2 is a schematic diagram of system connection for automatically adjusting a CPU voltage alarm threshold according to embodiment 1 of the present invention;

fig. 3 is a flowchart of a method for automatically adjusting a CPU voltage alarm threshold according to embodiment 2 of the present invention.

Detailed Description

In order to clearly explain the technical features of the present invention, the following detailed description of the present invention is provided with reference to the accompanying drawings. The following disclosure provides many different embodiments, or examples, for implementing different features of the invention. To simplify the disclosure of the present invention, the components and arrangements of specific examples are described below. Furthermore, the present invention may repeat reference numerals and/or letters in the various examples. This repetition is for the purpose of simplicity and clarity and does not in itself dictate a relationship between the various embodiments and/or configurations discussed. It should be noted that the components illustrated in the figures are not necessarily drawn to scale. Descriptions of well-known components and processing techniques and procedures are omitted so as to not unnecessarily limit the invention.

Example 1

The embodiment 1 of the invention provides a system and a method for automatically adjusting a CPU voltage alarm threshold, which are used for solving the problem of BMC voltage alarm misinformation caused by the fact that the CPU voltage alarm threshold is fixed and the influence of a CPU model and a Load Line value is not considered in the prior art. The problem that in the prior art, the VR Load Line value of the CPU power supply cannot be automatically adjusted according to the CPU model, the voltage alarm threshold is not set accurately enough, and the alarm cannot be given in time when the actual voltage is abnormal is solved.

Fig. 2 is a schematic diagram showing a system connection for automatically adjusting a CPU voltage alarm threshold according to embodiment 1 of the present invention, where the system includes a baseboard management controller BMC, a platform control center PCH, a central processing unit CPU, and a power supply module VR;

wherein, BMC: (Basebard Management Controller Baseboard Management Controller);

VR (Voltage Regulated, power supply module);

a CPU: (Central Processing Unit/Processor); a central processing unit;

PCH: (Platform Controller Hub); a platform control center;

VID: the core of the VID technology is that a plurality of VID pins are added on a CPU, and the encoding signals output by the VID pins control a PWM controller in a Vcore power supply circuit. Different codes represent different voltage values;

DMI: referred to as Direct Media Interface (dmi). DMI is a bus developed by Intel corporation to connect north and south bridges of a motherboard;

CPU Load Line value: the CPU load line standard. The intel's criterion for CPU load is that the voltage is dropped when the CPU is loaded.

The baseboard management controller BMC is in communication connection with the platform control center PCH; the platform control center PCH is in communication connection with the central processing unit CPU; the central processing unit CPU is in communication connection with a voltage output end of the power supply module VR; the baseboard management controller BMC is also in communication connection with an enabling end of the power supply module VR;

the power supply module VR receives a voltage regulation instruction sent by the baseboard management controller BMC to supply power to the central processing unit CPU; the platform control center PCH identifies the type of the central processing unit CPU through the inherent code of the central processing unit CPU, and feeds back the type information to the baseboard management controller BMC; after receiving the model information, the baseboard management controller BMC calls all pre-stored information of the central processing unit CPU according to the model information, and then automatically adjusts the CPU voltage alarm threshold.

The BMC carries out data communication with the PCH through the LPC bus to acquire CPU model information; the VR output voltage Vcore is read by the PMBUS, and the voltage value is monitored for exceeding a voltage alarm threshold.

The PCH carries out data communication with the CPU through the DMI to acquire CPU model information and sends the CPU model information to the BMC through the LPC, and the BMC also acquires the CPU model information.

VR in this application provides Vcore voltage for CPU, supports PMBUS protocol and SVID protocol. The VID Min and VID Max voltage values of the VR are the minimum and maximum VID voltages sent to the VR by the CPU through the SVID bus. The requirement of the CPU on the function of the Load Line value of the power supply chip is supported, and the Load Line value can be set through a register. And the BMC reads VR voltage through the PMBUS and edits and stores the Load Line value register. The CPU communicates with the VR through the SVID, sends the SVID command, and adjusts the VID voltage.

In the application, different CPU models have different codes, the codes are inherent attributes of the CPU and are confirmed when the CPU leaves a factory, and the different codes represent different CPU models. According to the encoding attribute, parameter values such as CPU power consumption, maximum current, Load Line value and the like can be confirmed.

After receiving the model information, the substrate management controller calls the pre-stored information of all central processing units according to the model information, and then the process of automatically adjusting the voltage alarm threshold of the central processing unit comprises the following steps:

The actual output voltage Vcore is VID-I Load Line value; VID is the identification voltage of the power supply module; the I is the current of the central processing unit; the Load Line value is a power supply Load Line standard value.

The process of sampling data of the actual output voltage and adding an anti-shake strategy comprises the following steps: and continuously acquiring the times that the actual output voltage is greater than the set voltage threshold value and exceeds the preset times, alarming, and removing the alarm if the voltage is recovered to be normal.

According to the system for automatically adjusting the CPU voltage alarm threshold, which is provided by the embodiment 1 of the invention, the CPU code is detected through the BMC, the CPU model is obtained, the CPU voltage alarm threshold is automatically adjusted according to the data such as the BMC pre-stored current and the Load Line value, and the VR Load Line value is adjusted, so that the system can adapt to different configurations, the problem that the BMC voltage monitoring data is not accurate enough or is misreported is avoided, and the fault diagnosis accuracy and the server operation reliability are improved.

Example 2

Based on the system for automatically adjusting the CPU voltage alarm threshold value provided by the embodiment 1 of the invention, the embodiment 2 of the invention also provides a method for automatically adjusting the CPU voltage alarm threshold value. Fig. 3 is a flowchart of a method for automatically adjusting a CPU voltage alarm threshold according to embodiment 2 of the present invention.

In step S201: the platform control center is communicated with the central processing unit, obtains the model information of the central processing unit and feeds back the model information to the substrate management controller;

in step S202: and the substrate management controller calls the pre-stored information of all central processing units according to the model information, reads the actual output voltage value of the power supply module through the PMBUS and monitors whether the actual voltage value exceeds a voltage alarm threshold value or not.

The method further comprises the step that the baseboard management controller records the monitoring voltage exceeding the voltage alarm threshold value to a log file so as to be used for fault diagnosis in the next step.

The detailed process of the method for automatically adjusting the CPU voltage alarm threshold value comprises the following steps:

setting the connection relation of BMC, PCH, CPU and VR, where BMC communicates with PCH via LPC bus, PCH communicates with CPU via DMI bus, CPU communicates with VR via SVID bus, and VR receives the voltage regulation instruction of CPU and provides Vcore voltage to CPU.

And storing specific information of each type of CPU into BMC, wherein the specific information comprises different types of CPUs and maximum currents In, (VID Min) n, (VID Max) n and Load Line values Ln, wherein n is 1,2,3 … n, and n represents different CPU types.

After the server is started every time, the CPU configuration may be changed, the PCH performs data communication with the CPU through the DMI, identifies the CPU model according to the inherent code of the CPU, and sends the information to the BMC through the LPC.

After receiving the CPU model information sent by the PCH, the BMC calls prestored data according to the CPU model, automatically adjusts alarm threshold setting, sets 6 voltage levels for voltage value threshold, and the high-low value coefficients are respectively: non-severe +/-A, severe +/-B, irreversible +/-C. The threshold coefficient is calculated as follows:

high irreversible capacity	[VID Max]n*(1+C)
		High severity	[VID Max]n*(1+B)
High non-severe	[VID Max]n*(1+A)
		Low non-severity	([VID Min]n-ILn)(1-A)
Low severity	([VID Min]n-ILn)(1-B)
		Low irreversible capacity	([VID Min]n-ILn)(1-C)

After receiving the CPU model information sent by the PCH, the BMC calls prestored data according to the CPU model, writes the VR Load Line value corresponding to the actually configured CPU model into the VR Load Line value register and stores the VR Load Line value. And then, according to the actual time sequence requirement of the system, the BMC re-enables the VR to enable the changed register to take effect.

The CPU adjusts VR VID voltage through the SVID, the VID voltage and the current value I of the CPU are constantly changed in the actual service operation process, and the actual output Vcore is VID-I Load Line value.

The BMC reads the VR voltage value register through the PMBUS, performs data acquisition on the Vcore voltage, and adds an anti-shaking strategy, so that misjudgment caused by signal interference is avoided, for example, if the voltage acquired for 5 times exceeds a set threshold value, an alarm is given, and if the voltage returns to normal, the alarm is released. The BMC records the monitoring voltage exceeding the corresponding threshold value to a log file, so that fault diagnosis is facilitated.

According to the method for automatically adjusting the CPU voltage alarm threshold, which is provided by the embodiment 2 of the invention, the CPU code is detected through the BMC, the CPU model is obtained, the CPU voltage alarm threshold is automatically adjusted according to the data such as the BMC pre-stored current and the Load Line value, and the VR Load Line value is adjusted, so that the system can adapt to different configurations, the problem that the BMC voltage monitoring data is not accurate enough or is misreported is avoided, and the fault diagnosis accuracy and the server operation reliability are improved.

It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Furthermore, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include elements inherent in the list. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element. In addition, parts of the above technical solutions provided in the embodiments of the present application, which are consistent with the implementation principles of corresponding technical solutions in the prior art, are not described in detail so as to avoid redundant description.

Although the embodiments of the present invention have been described with reference to the accompanying drawings, the scope of the present invention is not limited thereto. Various modifications and alterations will occur to those skilled in the art based on the foregoing description. And are neither required nor exhaustive of all embodiments. On the basis of the technical scheme of the invention, various modifications or changes which can be made by a person skilled in the art without creative efforts are still within the protection scope of the invention.

Claims

1. A system for automatic adjustment of a CPU voltage alarm threshold, comprising: the system comprises a substrate management controller, a platform control center, a central processing unit and a power supply module;

2. The system for automatically adjusting the CPU voltage alarm threshold according to claim 1, wherein the communication connection between the baseboard management controller and the platform control center is specifically:

3. The system for automatically adjusting the CPU voltage alarm threshold according to claim 2, wherein the platform control center is in communication connection with the central processing unit, specifically:

4. The system for automatically adjusting the CPU voltage alarm threshold according to claim 3, wherein the CPU is communicatively connected to the voltage output terminal of the power supply module; the substrate management controller is further in communication connection with an enabling end of the power supply module, and specifically comprises the following steps:

5. The system according to claim 4, wherein the information of all the central processing units pre-stored in the baseboard management controller includes the type of each central processing unit, the maximum current corresponding to each type of central processing unit, the minimum voltage identification value, the maximum voltage identification value, and the load line standard.

6. The system according to claim 5, wherein the baseboard management controller calls pre-stored information of all central processing units according to the model information after receiving the model information, and then automatically adjusting the voltage alarm threshold of the central processing unit comprises:

7. The system for automatically adjusting the CPU voltage alarm threshold according to claim 6, wherein the CPU adjusts the identification voltage of the power supply module through the SVID bus, and the calculation process for obtaining the actual output voltage of the power supply module is:

8. The system for automatically adjusting the CPU voltage alarm threshold according to claim 6, wherein the process of performing data sampling on the actual output voltage and adding an anti-jitter strategy is as follows:

9. A method for automatically adjusting a CPU voltage alarm threshold, which is implemented based on the system for automatically adjusting a CPU voltage alarm threshold according to any one of claims 1 to 8, and is characterized by comprising the following steps:

10. The method of claim 9, further comprising the baseboard management controller recording the monitored voltage exceeding the voltage alarm threshold to a log file for further fault diagnosis.