Partners Demo CXL Memory Sharing Innovation to Supercharge AI Performance
SANTA CLARA, Calif., Oct. 15, 2024 /PRNewswire/ -- XConn Technologies (XConn), the innovation leader in next-generation interconnect technology for the future of high-performance computing and AI applications, and MemVerge®, a leader in AI-first Big Memory Software, today announced a joint demonstration of the industry's first scalable Compute Express Link® (CXL®) memory sharing solution. Specifically designed to supercharge AI workloads, this demo will highlight how CXL technology can dramatically enhance performance, scalability and efficiency of memory, while reducing total cost of ownership (TCO) for AI applications, in-memory databases and large-scale data analytics workloads.
The 2024 OCP Global Summit, held from October 15-17, 2024, at the San Jose Convention Center, serves as the premier global event for innovation in open hardware and software. At this event, XConn and MemVerge will demonstrate how their cutting-edge CXL memory sharing solution can unlock unprecedented performance gains for AI-driven workloads, positioning it as a game-changing development for data centers and high-performance computing environments.
During the event, XConn's Jianping Jiang, Senior Vice President of Business and Product, will also detail the benefits of composable memory for AI workloads during his presentation, "Enabling Composable Scalable Memory for AI Inference with CXL Switch," to be presented during OCP on Thursday, October 17 at 9:45 a.m.
"AI workloads are among the most resource-intensive applications, and our scalable CXL memory sharing solution addresses the growing need for increased memory bandwidth and capacity," said Gerry Fan, CEO of XConn Technologies. "At the OCP Global Summit, we will show how our collaboration with MemVerge not only improves AI model inference performance but also enables more efficient memory usage, reducing infrastructure costs while maximizing compute power."
Charles Fan, CEO and co-founder of MemVerge, added, "The rise of AI has transformed the demands on memory architectures. Our partnership with XConn brings a new class of solutions featuring fabric-attached CXL memory with sharing enabled by our Memory Machine X software. This innovation unlocks the performance potential of AI workloads by enabling seamless scalability, reducing latency, and dramatically improving system throughput."
The focal point of the demonstration will be the XConn "Apollo" CXL 2.0 switch, a first-of-its-kind hybrid solution supporting both CXL 2.0 and PCIe Gen 5 in a single design. AI and high-performance computing (HPC) environments stand to benefit significantly from this technology, as it eliminates memory bottlenecks, boosts memory bandwidth, and enhances system efficiency. The Apollo switch is engineered to enable AI models to process larger datasets faster and more efficiently, driving improvements in applications such as deep learning, AI model inference, real-time data analytics, and in-memory databases.
In tandem, MemVerge's Memory Machine™ X software will be highlighted during the demonstration as the software's Global IO-free Shared Memory Objects (GISMO) capability allows multiple nodes to have direct access to the same CXL memory region while providing cache coherence. To learn more about shared fabric-attached memory, visit the Memory Machine X web page, or attend the Memory Fabric Forum at OCP Global Summit on October 16.
Demos of Composable Memory Systems powered by XConn Apollo switch will be shown in the exhibit floor at Innovation Village location FN-125 and FN-126. Separately a full rack of computing system with two dual socket servers and up to 12 TB CXL Memory running MemVerge's GISMO will be shown at booth B31 Innovation Village hosted by Credo Technologies. For customer samples and/or Apollo reference boards, contact XConn at xconn-tech.com.
If you're ready to test drive a shared memory system, PoCs are available from MemVerge upon request.
For more details about the OCP Global Summit and to register, visit: https://www.opencompute.org/summit/global-summit.
About XConn Technologies
XConn Technologies Holdings, Inc. (XConn) is the innovation leader in next-generation interconnect technology for high-performance computing and AI applications. The company is the industry's first to deliver a hybrid switch supporting both CXL 2.0 and PCIe Gen5 on a single chip. Privately funded, XConn is setting the benchmark for data center interconnect with scalability, flexibility and performance. For more information visit: xconn-tech.com.
About MemVerge
MemVerge is a leading AI infrastructure software company specializing in optimizing big memory GPU-centric computing in the era of Generative AI. Leveraging its unique expertise in memory software, MemVerge solutions empower organizations across various industries, including financial services, scientific computing, and cloud service providers, to accelerate their AI and scientific computing workloads. For more information, please visit http://www.memverge.com
Media Contacts:
Steve Sturgeon
Avista PR
ssturgeon@avistapr.com
858.472.5669
SOURCE XConn Technologies