#ComputeExpressLink

In case you missed it: Watch the #CXLConsortium webinar featuring @MicronTech’s Skyler Windh on how #ComputeExpressLink (#CXL) fabric-attached memory enables shared KV Cache across GPUs to improve #AI inference efficiency. View the recording:

At @Supercomputing 2025 #SC25, from Nov. 17 – 20 in St. Louis, MO, the #CXLConsortium will showcase how #ComputeExpressLink (#CXL) memory pooling and sharing address #AI and #HPC challenges. Learn more about our activities by reading our SC25 promo blog:

Read @VIAVISolutions blog on the #CXLConsortium website to learn more about their upcoming demonstration at #CXLDevCon2025, showcasing how #ComputeExpressLink implementations can be enhanced by protocol analyzers:

Join #CXLConsortium tomorrow at 12:45 pm at #AIInfraSummit2026 for “Boost Your AI Performance With CXL,” featuring @Samsung, @AsteraLabs, @Meta, and Primemas to explore how #ComputeExpressLink (#CXL) enables scalable memory and efficient AI infrastructure:

#ComputeExpressLink (#CXL) 4.0 brings major advances for #AI and #HPC workloads, doubling bandwidth to 128 GT/s, enabling bundled ports, and strengthening memory RAS for scalable, reliable next-gen systems. Read the full white paper to learn more:

The #CXLConsortium webinar Introducing the #ComputeExpressLink (#CXL) 4.0 Specification is now on YouTube. Watch to webinar to explore key updates to the specification like 128 GT/s bandwidth, bundled ports, and enhanced memory RAS:

The #ComputeExpressLink (#CXL) 4.0 Specification introduced significant improvements in bandwidth, connectivity, memory maintenance, and security. Download the white paper for a deep dive into the new features in the 4.0 specification:

Watch this brief video to learn how #ComputeExpressLink (#CXL) technology enables a cache coherent architecture and delivers reduced latency and overhead with increased ease of use:

Visit the #ComputeExpressLink (#CXL) Pavilion at #SC26 to see live demos from #CXLConsortium member companies showcasing how CXL is advancing memory expansion, pooling, and sharing to meet the demands of AI and high-performance computing. Learn more:

Learn how Astera Labs is the first to break through the memory wall! Watch the video with @NextGenInfra_io to learn more: @ConvergeDigest @WireRoy @AvidThink #CXL #ComputeExpressLink #AI #cloud #IntelInnovation #datacenter

#ComputeExpressLink (#CXL) 4.0 boosts performance, scalability, and reliability for #AI and #HPC workloads. Read the CXL 4.0 white paper for key advancements, including 128 GT/s bandwidth, bundled ports, and enhanced memory RAS:

#CXLConsortium is pleased to announce the release of the #ComputeExpressLink (#CXL) 3.2 Specification, which enhances security, compliance, and functionality of CXL Memory Devices. Read the press release to learn more:

#CXLConsortium TTF Co-Chair Debendra Das Sharma shared an overview of #ComputeExpressLink (#CXL) and the future of disaggregated and composable architectures @Intel Hot Interconnects 2021. Learn more via @TheNextPlatform:

We are partnering with @SKhynix on interop testing of its DDR5-5600 memory modules with our Leo #CXL Memory Connectivity Platform in our Cloud-Scale Interop Lab! Learn more: #ComputeExpressLink #cloud #AI #datacenter #DDR5 #DDR

Our Leo Memory Connectivity Platform expands memory bandwidth and capacity up to 2TB per #CXL link. Leo is optimized for meeting the complex computational needs of generative #AI workloads at low latency. Learn more: #ComputeExpressLink #cloud #datacenter

#CXLConsortium releases the #ComputeExpressLink (#CXL) 3.0 specification that introduces fabric capabilities and management, improved memory sharing and pooling, enhanced coherency, and peer-to-peer communication. Read the announcement:

Join the #CXLConsortium webinar tomorrow, February 19, at 9:00 am PT, for a live discussion on the new enhancements in the #ComputeExpressLink (#CXL) 3.1:

Intel teams up with @AlibabaGroup, @Cisco, @DellEMC, @facebook, @Google, @HPE, @Huawei and @Microsoft to move data faster. #CXLConsortium #ComputeExpressLink

Compute Express Link technology will improve performance and remove bottlenecks in computation-intensive workloads like #AI and #encryption: #CXL #ComputeExpressLink

Live in 1 hour: Join #CXLConsortium member company representatives from @Giga_IO, @WeAreLiqid, and @ZeroPointTechno during our presentation at #ISC24 to explore use cases and ROI examples enabled with #ComputeExpressLink (#CXL) technology. Learn more:

Read the latest #CXLConsortium blog post to find answers to important questions asked during our “Introducing the CXL 3.1 Specification” webinar, including the performance impacts and the memory impacts for #ComputeExpressLink (#CXL) Type-1 devices: