公开(公告)号：US20240264957A1

公开(公告)日：2024-08-08

申请号：US18425079

申请日：2024-01-29

Applicant: KOREA ADVANCED INSTITUTE OF SCIENCE AND TECHNOLOGY

Inventor： Myoungsoo JUNG ,  Donghyun GOUK ,  Miryeong KWON

IPC: G06F13/16 ,  G06F12/02 ,  G06F13/42

CPC classification number: G06F13/1673 ,  G06F12/0246 ,  G06F13/4221

Abstract: A compute express link (CXL) computing system includes a host device including a CPU that supports CXL, and a CXL storage connected to a CXL root port of the CPU based on the CXL interconnect and including a flash memory-based memory module.

公开(公告)号：US20240012684A1

公开(公告)日：2024-01-11

申请号：US18161169

申请日：2023-01-30

Applicant: KOREA ADVANCED INSTITUTE OF SCIENCE AND TECHNOLOGY

Inventor： Myoungsoo JUNG ,  Donghyun GOUK

IPC: G06F9/50 ,  G06F3/06 ,  G06F13/42

CPC classification number: G06F9/5016 ,  G06F3/0613 ,  G06F3/0659 ,  G06F3/067 ,  G06F13/4221

Abstract: Disclosed is a memory disaggregation computing system including a host server and a memory device connected through a compute express link (CXL) network, in which a computing complex of the host server is connected to a memory resource of the memory device through a CXL packet transmitted through the CXL network, and executes an application program by using the memory resource.

公开(公告)号：US20230418673A1

公开(公告)日：2023-12-28

申请号：US18166685

申请日：2023-02-09

Applicant: KOREA ADVANCED INSTITUTE OF SCIENCE AND TECHNOLOGY

Inventor： Myoungsoo JUNG ,  Junhyeok JANG ,  Miryeong KWON ,  Donghyun GOUK ,  Hanyeoreum BAE

IPC: G06F9/50 ,  G06F9/48 ,  G06F9/54

CPC classification number: G06F9/5027 ,  G06F9/54 ,  G06F9/4881

Abstract: Provided is an apparatus for accelerating a graph neural network for efficient parallel processing of massive graph datasets, including a streaming multiprocess (SM) scheduler and a computation unit, wherein the SM scheduler obtains a subgraph and an embedding table per layer, determines a number of SMs to be allocated for processing embeddings of a destination-vertex based on a feature dimension and a maximum number of threads in each of the SMs, and allocates the determined number of SMs to each of all destination-vertices included in the subgraph, and the computation unit obtains, by each of the SMs, embeddings of a destination-vertex allocated to each SM, obtains, by each SM, embeddings of at least one or more neighbor-vertices of the destination-vertex using the subgraph, and performs, by each SM, a user-designated operation using the embeddings of the destination-vertex and the embeddings of the neighbor-vertices.