compute-express-link-3.0-(cxl)-specifications-released,-up-to-64gt/s-link-rate-–-amd,-intel,-nvidia-&-other-onboard

The CXL Consortium, an business requirements physique, devoted to advancing Compute Express Link (CXL) know-how, introduced the discharge of the CXL 3.0 specification right this moment, which will likely be adopted by AMD, Intel, NVIDIA, and different main server gamers.

CXL Consortium launches Compute Express Link 3.0 specification to increase cloth talents and knowledge administration

The CXL 3.0 specification expands on earlier know-how generations to extend scalability and optimize system stage flows with superior switching and cloth capabilities, environment friendly peer-to-peer communications, and fine-grained useful resource sharing throughout a number of compute domains.

Modern datacenters require heterogenous and composable architectures to help compute intensive workloads for purposes akin to Artificial Intelligence and Machine Learning – and we proceed to evolve Compute Express Link  know-how to fulfill business necessities. Developed by our devoted technical workgroup members, the CXL 3.0 specification will allow new utilization fashions in composable disaggregated infrastructure.

  • Siamak Tavallaei, President, CXL Consortium.

CXL is a transformative know-how that continues to construct momentum throughout the business. The introduction of Compute Express Link 3.0 meets the wants of next-generation knowledge facilities with 64 GT/s signaling and a brand new stage of scalability,” stated Travis Karr, normal supervisor of Interconnect SoCs at Rambus. “At Rambus, we’re extraordinarily proud to be members of the rising CXL ecosystem and dedicated to the event of options that may speed up the adoption of this thrilling new know-how.

  • Travis Karr, normal supervisor of interconnect SoCs at Rambus

Compute Express Link 3.0 (CXL) specifications released, up to 64GT/s link rate 2

Source: CXL Consortium

Key highlights:

  • The Compute Express Link™ 3.0 specification will introduce cloth capabilities and administration, improved reminiscence sharing and pooling, enhanced coherency, and peer-to-peer (P2P) communication.
  • The Compute Express Link 3.0 specification gives two instances the info fee, enhancing knowledge switch to 64 GTs with zero latency over the earlier CXL 2.0. The new specification is at present accessible for public use.
  • CXL Consortium and its partnered firms will convey important insights on Compute Express Link know-how through the Flash Memory Summit (FMS) on August 2nd by means of the 4th.

Compute Express Link 3.0 (CXL) specs launched, as much as 64GT/s hyperlink fee – AMD, Intel, NVIDIA & Different Onboard

Source: CXL Consortium

Highlights of the Compute Express Link 3.0 specification:

  • Fabric capabilities
    • Multi-headed and Fabric Attached Devices
    • Enhanced Fabric Management
    • Composable disaggregated infrastructure
  • Better scalability and improved useful resource utilization
    • Enhanced reminiscence pooling
    • Multi-level switching
    • New enhanced coherency capabilities
    • Improved software program capabilities
  • Doubles the bandwidth to 64GTs
  • Zero added latency over Compute Express Link 2.0
  • Full backward compatibility with 2.0, 1.1, and 1.0

cxl3-3

cxl3-4

2 of 9

The Compute Express Link Consortium will introduce the brand new CXL 3.0 options on the Flash Memory Summit, beginning August 2nd on the Santa Clara Convention Center and ending two days later. You can view the Summit’s agenda for detailed data inside the shared displays.

For readers or customers eager about discovering extra concerning the new CXL 3.0 specs, you will discover out extra right here, in addition to a Consortium member assertion of help. Finally, you possibly can learn the official white paper, describing extra of the method in additional element and a hyperlink to the video above.

News Source: CXL Consortium