Server resource box FPGA firmware upgrading method and system, terminal and storage medium

Abstract

The invention provides a server resource box FPGA firmware upgrading method and system, a terminal and a storage medium, and the method comprises the steps that a first BMC receives a firmware file sent by a second BMC through an I2C bus and performs integrity verification on the firmware file, the first BMC is a machine head BMC, and the second BMC is a resource box BMC; the first BMC sends the firmware file subjected to integrity verification to a resource box FPGA; and after the firmware file is sent, the first BMC sends a cold restart command to the FPGA, so that the firmware file of the FPGA takes effect after cold restart. Compared with a previous upgrading mode, the method can be operated without powering off the server and a special jig, and is convenient for batch operation during implementation; and meanwhile, a resource box BMC does not need to be accessed to a network during machine deployment, so that IP resources of a client computer room are saved, the wiring complexity is simplified, and the operation and maintenance cost is finally saved.

Description

technical field [0001] The invention relates to the technical field of servers, in particular to a server resource box FPGA firmware upgrade method, system, terminal and storage medium. Background technique [0002] NVIDIA launched the HGX-2Next product in 2020, capable of handling the most computationally intensive workloads and implementing the "world's largest GPU." it consists of 16 Tes A100GPU and NVIDIA NV Switch TM Acceleration, with unprecedented computing power, bandwidth and memory topology, can train massive models, analyze data sets and solve simulation problems faster and more efficiently. Relying on the HGX-2Next computing board, OEMs can develop resource boxes with great computing potential to expand the performance of original servers. FPGA, as an important part of HGX-2Next, manages the interface between HGX-2Next and the outside world, ensures the connectivity and stability of the entire PCIE link, and plays a vital role. [0003] There are two exist...

AI Technical Summary

Problems solved by technology

Disadvantages of offline programming of the HGX-2Next FPGA firmware through the programmer: (1) can only be programmed when the power is off, which affects the normal use of the machine; (2) the machine needs to be disassembled. It is very easy to damage components; (3) only artificial programming can be performed independently, and the efficiency is extremely low

Disadvantages of directly upgrading the firmware through the resource box BMC: additional network cables are required to connect to the resource box BMC, which increases the maintenance cost of the client computer room, and also raises high requirements for the server deployment environment for customers, making it difficult to promote as a general method

Images