Node with a bridge network interface device for an AI accelerator system

The transparent bridge configuration with software-assigned task IDs and mirrored BARs in IO streaming devices addresses scalability and latency issues in distributed AI accelerators, facilitating efficient communication and high-performance AI workloads.

DE202025106073U1Active Publication Date: 2026-04-30D-MATRIX CORP
0 Cites 0 Cited by

Patent Information

Application Number
DE202025106073
Authority / Receiving Office
DE · DE
Patent Type
Utility models
Current Assignee / Owner
Filing Date
2025-10-07
Publication Date
2026-04-30
Estimated Expiration
2035-10-31

AI Technical Summary

Technical Problem

Existing network interface solutions for distributed AI accelerators face challenges in scalability, low-latency communication, and complex deployment due to limitations in PCIe fabric topologies and RDMA over Converged Ethernet with InfiniBand, which hinder the performance of modern accelerators.

Method used

A transparent bridge configuration using software-assigned task IDs for data transmission across channels, enabling efficient PCIe peer-to-peer communication between accelerators, and a system with IO streaming devices configured as NICs to expose mirrored BARs for scalable communication between nodes, supporting tensor and pipeline parallelism.

Benefits of technology

Enables fast and efficient communication across multi-node accelerator systems, overcoming scalability and latency limitations, and supporting high-performance AI workloads without explicit software activation.

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

Abstract

AI system comprehensive: a first server device comprehensive: a first host central processing unit; a first PCIe switch coupled with the first host central processing unit; a plurality of first AI accelerator devices coupled with the first PCIe switch; an input network interface configured to receive information using a task ID from a second server device configured with an Ethernet protocol; an intranetwork value interface configured to communicate from one or more of the plurality of first AI accelerator devices to one or more of a plurality of second AI accelerator devices, such that the communication uses one or more task IDs without communicating directly or indirectly from a second host central processing unit; an output network interface configured to transmit information to the second server device using one or more task IDs via the Ethernet protocol; and a second PCIe switch configured to couple the second host central processing unit with the majority of second AI accelerator devices, the output network interface, and the intra-network value interface.
Need to check novelty before this filing date? Find Prior Art