数据处理方法、装置、电子设备及可读存储介质
By introducing a matrix accelerator into the processor to construct the source matrix for data transfer, the problem of low data exchange efficiency in devices without a DMA controller is solved, achieving efficient data transfer and meeting higher performance requirements.
CN118331904BActive Publication Date: 2026-07-17LOONGSON TECH CORP
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- LOONGSON TECH CORP
- Filing Date
- 2024-04-23
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
In the prior art, PCI/PCIE devices without a DMA controller are inefficient during data exchange, consume a lot of processor resources, and cannot effectively utilize DMA technology for fast data exchange.
Method used
By introducing a matrix accelerator into the processor, a source matrix is constructed and used for data transfer, enabling direct transmission of target data from the target device to the system memory, thus avoiding repetitive operations in PIO mode.
Benefits of technology
This improves data exchange efficiency, reduces processor utilization, and enables target devices to meet higher performance requirements.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN118331904B_ABST
Abstract
本发明实施例提供了一种数据处理方法、装置、电子设备及可读存储介质，该方法包括：响应于数据存储指令，从目标设备对应的第一存储空间中确定数据存储指令指示的目标数据以及目标数据对应的目标数据信息；基于目标数据以及目标数据信息，构建源矩阵；通过矩阵加速器基于源矩阵进行数据搬运，以实现将目标数据从第一存储空间传输至目标存储空间；目标存储空间为系统内存中用于存储目标数据的内存空间。这样，在目标设备未配备DMA控制器的情况下，可以基于处理器中的矩阵加速器通过源矩阵以及目标矩阵之间的转换，实现目标数据从目标设备至内存中直接存储空间的数据搬运，降低了处理器的使用率的同时，提高了数据交换的效率。
Need to check novelty before this filing date? Find Prior Art