Memory pool device fusing extended network and management method

By integrating the super network card chip with the intelligent computing cluster to expand the network, the problem of the lack of a unified architecture in memory expansion solutions is solved, realizing high-bandwidth and low-latency memory sharing, improving data transmission efficiency and memory utilization, and meeting the high-efficiency operation requirements of the MoE model.

CN120929418APending Publication Date: 2025-11-11COMMUNICATION INTERNET (SHENZHEN) TECHNOLOGY CO LTD
0 Cites 1 Cited by

Patent Information

Application Number
CN202511059552.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-07-30
Publication Date
2025-11-11

AI Technical Summary

Technical Problem

In existing computer systems, the lack of a unified architecture for memory expansion schemes and distributed caching leads to high hardware resource consumption for XPU/GPU, and traditional Scale Out networks are unable to meet the high bandwidth and low latency requirements of the MoE model.

Method used

It adopts a super network card chip to connect with the intelligent computing cluster and provides shared memory services. Through remote direct memory access protocol and proprietary protocol, it realizes the integration of horizontal and vertical network expansion. Combined with FPGA and XPU expansion unit, it optimizes data transmission path and provides high bandwidth and low latency memory sharing.

Benefits of technology

It improves data transmission efficiency and memory utilization, reduces latency, and enables flexible scheduling and sharing of memory resources, thus meeting the high-efficiency operation requirements of the MoE model.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a memory pool device fusing an extended network and a management method, and the device comprises a super network card chip which is in communication connection with an intelligent computing cluster network through a network port, and provides a shared memory service for the intelligent computing cluster network based on a remote direct memory access protocol; the intelligent computing cluster network comprises a transverse expansion network and a longitudinal expansion network; the central controller is connected with the super network card chip; the CXL memory unit is connected with the super network card chip, and the FPGA expansion unit and / or the XPU expansion unit are / is connected with the super network card chip. According to the invention, through efficient communication between the super network card chip and the intelligent computing cluster networking, flexible scheduling and sharing of memory resources in the cluster are realized. Wherein the super network card chip is used as a core component, not only is responsible for communication of a transverse expansion network, but also realizes bridging access of a high-bandwidth domain of a longitudinal expansion network, so that the data transmission efficiency and the memory utilization rate can be improved, and high-bandwidth and low-delay data transmission and sharing are realized.
Need to check novelty before this filing date? Find Prior Art