PCIe 6.4/CXL 3.2 Fabric Switch Sample is Out Now! - Request the Silicon Sample via[email protected]
Learn More
Logo
  • About
  • Product
  • Technology
  • Newsroom
  • Careers
VisionLeadershipHistoryMembership

Hardware

PanSwitchPanRetimer

Silicon IP

LAU IPController IP

Custom Silicon & Solutions

PanEndpointPanFabricTotal AI Solution
Technical ReportsTech BlogPublications
EnglishKorean
CareersPositionsApply
Contact Us
  1. Back to Publications
  2. /
  3. Breaking Barriers: Expanding GPU Memory with Sub-Two Digit Nanosecond Latency CXL Controller

Breaking Barriers: Expanding GPU Memory with Sub-Two Digit Nanosecond Latency CXL Controller

Breaking Barriers: Expanding GPU Memory with Sub-Two Digit Nanosecond Latency CXL Controller
Breaking Barriers: Expanding GPU Memory with Sub-Two Digit Nanosecond Latency CXL Controller

Donghyun Gouk, Seungkwan Kang, Hanyeoreum Bae, Eojin Ryu, Sangwon Lee, Dongpyung Kim, Junhyeok Jang, Myoungsoo Jung

The ACM Workshop on Hot Topics in Storage and File Systems (HotStorage)

2024

Research Areas
Architecture
Coherent Interconnect
Read PaperGoogle Scholar

Abstract

This work introduces a GPU storage expansion solution utilizing CXL, featuring a novel GPU system design with multiple CXL root ports for integrating diverse storage media (DRAMs and/or SSDs). We developed and siliconized a custom CXL controller integrated at the hardware RTL level, achieving two-digit nanosecond roundtrip latency, the first in the field. This study also includes speculative read and deterministic store mechanisms to efficiently manage read and write operations to hide the endpoint's backend media latency variation. Performance evaluations reveal our approach significantly outperforms existing methods, marking a substantial advancement in GPU storage technology.


Related Publications
Featured
MPI-over-CXL: Enhancing Communication Efficiency in Distributed HPC SystemsSPICE • 2025
Coherent Interconnect
Operating Systems
+1 more
Featured
ScalePool: Hybrid XLink-CXL Fabric for Composable Resource Disaggregation in Unified Scale-up DomainsDIMES • 2025
Coherent Interconnect
Architecture
Featured
Compute Can't Handle the Truth: Why Communication Tax Prioritizes Memory and Interconnects in Modern AI InfrastructurearXiv (Technical Report) • 2025
Coherent Interconnect
View All Publications
Logo

Building the future of AI infrastructure with innovative semiconductor solutions.

Privacy Policy© 2025 Panmnesia, Inc.
All rights reserved.
About
VisionLeadershipHistoryMembership
Product

Hardware

PanSwitchPanRetimer

Silicon IP

LAU IPController IP

Custom Silicon & Solutions

PanEndpointPanFabricTotal AI Solution
Technology
Technical ReportsTech BlogPublications
Newsroom
EnglishKorean
Careers
CareersPositionsApply
Logo

Building the future of AI infrastructure

Quick Access
AboutProductsCareersNews
Technical ReportsPublications

About

▼

VisionLeadershipHistoryMembership

Products

▼

PanSwitchPanRetimerLAU IPController IPPanEndpointPanFabricTotal AI Solution

Technology

▼

Technical ReportsTech BlogPublications

Newsroom

▼

EnglishKorean

Careers

▼

CareersPositionsApply
Privacy Policy© 2025 Panmnesia, Inc.