Distributed storage system and data processing method

Abstract

This application provides a storage device, a distributed storage system, and a data processing method, and belongs to the field of storage technologies. In this application, an AI apparatus is disposed inside a storage device, so that the storage device has an AI computing capability. In addition, the storage device further includes a processor and a hard disk, and therefore further has a service data storage capability. Therefore, convergence of storage and AI computing power is implemented. An AI parameter and service data are transmitted inside the storage device through a high-speed interconnect network without a need of being forwarded through an external network. Therefore, a path for transmitting the service data and the AI parameter is greatly shortened, and the service data can be loaded nearby, thereby accelerating loading.

What is claimed is: 
     
         1 . A device, comprising:
 a central processing unit (CPU);   a hard disk; and   a graphics processing unit (GPU) coupled to the CPU and the hard disk,   wherein the hard disk is configured to store first service data, and wherein   the GPU is configured to:   retrieve the first service data from the hard disk bypass the GPU, and perform computing on the first service data.   
     
     
         2 . The device according to  claim 1 , wherein the GPU includes a memory, and wherein the GPU is further configured to write the first service data into the memory after retrieving the first service data. 
     
     
         3 . The device according to  claim 2 , wherein the storage device includes a cache for temparily storing data to be processed by the CPU, wherein the cache stores second data, and wherein
 the GPU is further configured to retrieve the second data stored in the cache through the CPU, and store the second data in the memory of the GPU.   
     
     
         4 . The device according to  claim 1 , wherein the GPU communicates with the CPU through peripheral component interconnect express (PCIe). 
     
     
         5 . The device according to  claim 1 , wherein the GPU communicates with the hard disk through peripheral component interconnect express (PCIe). 
     
     
         6 . The device according to  claim 1 , wherein the hard disk includes a solid state disk (SSD). 
     
     
         7 . A method for accessing data implemented by a graphics processing unit (GPU) that is coupled to a hard disk and a central processing unit (CPU), comprising:
 retrieving first service data from the hard disk bypass the GPU; and   performing computing on the first service data.   
     
     
         8 . The method according to  claim 7 , wherein the GPU includes a memory, the method further comprising:
 writing the first service data into the memory after retrieving the first service data.   
     
     
         9 . The method according to  claim 8 , wherein the storage device includes a cache for temparily storing data to be processed by the CPU, wherein the cache stores second data, and the method further comprising:
 retrieving the second data stored in the cache through the CPU, and   storing the second data in the memory of the GPU.   
     
     
         10 . The method according to  claim 7 , further comprising:
 communicating with the CPU through PCIe.   
     
     
         11 . The method according to  claim 7 , further comprising:
 communicating with the disk through PCIe.   
     
     
         12 . The method according to  claim 7 , the hard disk includes a SSD. 
     
     
         13 . A graphics processing unit (GPU), comprising:
 a processor; and   a memory coupled to the processor,   wherein the processor is configured to:   retrieve service data from a solid state disk (SSD);   write the service data into the memory; and   perform computing on the service data stored in the memory.   
     
     
         14 . The GPU according to  claim 13 , wherein the processor is further configured to bypass a central processing unit (CPU) of a server including the GPU during the GPU retrieving the service data.