Request processing method and device, equipment and medium

Abstract

The invention discloses a request processing method and device, equipment and a medium, and relates to the field of management software, and the method comprises the steps: obtaining an address of a current to-be-processed request from a local first-in first-out queue, and obtaining the current to-be-processed request from a preset request buffer region based on the address of the current to-be-processed request; analyzing the current to-be-processed request, and obtaining corresponding current to-be-processed source data and a serial address entry used for recording an address of a next to-be-processed request based on an analysis result; processing the current to-be-processed source data to obtain current processed target data, writing the current processed target data into a preset data buffer area, and judging whether a serial address entry is empty or not; and if the serial address entry is null, notifying the host to read all processed target data from the preset data buffer area, and if the serial address entry is non-null, sending an address recorded in the serial address entry to a first-in first-out queue of a next acceleration card for storage. The purposes of reducing the burden of the CPU and improving the request processing efficiency are achieved.

Description

technical field [0001] The invention relates to the field of management software, in particular to a request processing method, device, equipment and medium. Background technique [0002] In recent years, with the development of emerging technologies such as mobile Internet and artificial intelligence, the scale of big data and the scale of Internet applications are growing exponentially. The business that needs to be processed by the data center also increases, and the computing power of the general-purpose processor used in the data center has reached the bottleneck. In order to solve this problem, various PCIe (peripheral component interconnect express, namely high-speed serial computer Expansion bus standard) boards came into being, such as compression and decompression cards, RAID (Redundant Arrays of Independent Disks, that is, redundant array of independent disks) memory cards, and the well-known GPGPU (General-purpose computing) for neural network training and deriv...

AI Technical Summary

Problems solved by technology

However, since different accelerator cards use different driver interfaces, in the prior art, when multiple accelerated calculations are required, each step of calculation processing needs to be scheduled by the host, which undoubtedly increases the CPU (central processing unit, that is, the central processing unit). ) burden and reduce processing efficiency

Images