Intel has collaborated with leading tech giants for forming a new consortium to advance the Compute Express Link (CXL). The aim is to accelerate the performance of data centers.
CXL is an open interconnect technology that optimizes performance and removes the roadblocks in computation-intensive workloads for CPUs and purpose-built accelerators. The technology was developed by Intel, and it has now been donated to consortium.
Co-developed by Alibaba, Cisco, Dell EMC, Facebook, Google, HPE, Huawei, Intel, and Microsoft, the CXL will work as an industry standard group.
“I am proud of the work Intel has done in developing this interconnect technology and the milestone it represents to the technology industry – much like our roles with Universal Serial Bus (USB) and PCI Express – and we look forward to working with the CXL consortium on future versions of the specification,” wrote Navin Shenoy, VP/GM, Data Center Group, Intel, in a blog post.
The explosion of data and continuous innovation in computation-intensive applications like artificial intelligence (AI), encryption and compression are demanding heterogenous computing. The accelerators in these applications work along the general-purpose CPUs and require a high-performance connection to the processor.
CXL technology has the capability to allow memory coherence between accelerator and CPU while offering very high bandwidth. Furthermore, it will enable high-speed and low latency interconnect between the CPU and workload accelerators.
“It’s incredible to see the caliber of companies that came together to form the consortium for CXL. They represent the leaders in data center, communications infrastructure, and cloud computing and services – and we expect more companies to join in the coming future to contribute to the next iteration of this spec,” added Navin.
The Compute Express Link is expected to be available to members of the consortium in the first half of 2019. The products that incorporate CXL technology will start in 2021 with Intel datacenter platforms, including Intel Xeon processors, FPGAs, and SmartNICs.