Memory thermal trip detecting system and method

Abstract

Disclosed is a memory thermal trip detecting method. The memory thermal trip detecting method is used for detecting thermal trip signals of a plurality of memories in a server and includes a setting step, a detecting step, a send request step, a receiving step, an analyzing step and a recording step, wherein the setting step includes that an identity (ID) number of each memory is set; the detecting step includes that whether thermal trip signals of a main memory of a gate output end are triggered or not is detected; the send request step includes that when the thermal trip signals of the main memory are detected to be triggered, a preset inquiry signal request is sent to inquire a thermal trip signal value of each memory in a complex programmable logic device (CPLD), and each input end of the CPLD is connected with a thermal trip signal output end of each memory; the receiving step includes that serial data of the thermal trip signals of all memories, which are sent by the CPLD according to a preset received signal coded format, are received; the analyzing step includes that the received serial data of the thermal trip signals of all memories are analyzed; the recording step includes that the memories with triggered thermal trip signals are recorded in a BMC journal according to the analyzed data. The invention further provides a memory thermal trip detecting system.

Description

technical field [0001] The invention relates to a system and method for detecting a thermal trip of a memory. Background technique [0002] With the explosive growth of data processing requirements and the continuous maturity of technology, the number of CPUs and memory configured on servers is increasing. For example, the basic Romley EN platform is generally configured with 2 CPUs and 12 memory sticks, and the high-end Romley EP platform can support up to 4 CPUs and 48 memory sticks. When the temperature of the memory is too high, the memory hardware will trigger a thermal trip (thermal trip) signal. An important basic function of the BMC is to monitor the thermal interrupt signal of each memory, and take corresponding measures when the signal occurs (such as recording the system event record SEL, fan acceleration, etc.). But for so many memories, it is obvious that the BMC does not have enough general purpose input / output (GPIO) resources to connect the thermal interrup...

AI Technical Summary

Problems solved by technology

However, this method has several disadvantages: 1) Those supporting the SGPIO protocol are all high-end BMC chips, and the cost is relatively high; 2) More parallel-to-serial conversion chips are required, which not only increases the cost, but also wastes valuable PCB space

In this way, the BMC can detect which memory corresponding to the CPU has a thermal interrupt signal, but the disadvantage is that it is impossible to know which memory triggers the thermal interrupt signal of the memory.

Images