Industry Sounds Off on the Emergence of Compute Express Link™ (CXL)
Updated: Dec 17, 2019
By Scott Knowlton
Compute Express Link™ (CXL) technology made its initial public debut in March 2019, when industry leaders – Alibaba, Cisco, Dell EMC, Facebook, Google, Hewlett Packard Enterprise, Huawei, Intel Corporation and Microsoft – came together to jointly develop and promote this exciting new high performance data center interconnect standard. Since then, the CXL Consortium™ has announced its official incorporation, augmented the Board of Directors and greatly expanded its membership.
The arrival of CXL represents a significant breakthrough in connectivity for the High-Performance Computing (HPC) and Enterprise Cloud industries, which are tackling increasingly complex workloads such as Artificial Intelligence and Machine Learning. To address these data intensive applications, system designers have turned to a growing number of CPU enhancements and workload accelerators, utilizing GPUs, FPGAs and other purpose-built solutions. This is creating a truly heterogeneous computing environment requiring seamless communication while leveraging a wide variety of compute and storage protocols.
This is where CXL enables the heterogeneous computing environment to maintain memory coherency between the CPU and attached devices, allowing optimized system resource sharing, reduced software stack complexity, and resulting in lower overall system cost.
Below is a snapshot of how notable industry observers have reacted to CXL’s emergence:
“CXL will be an important element in bringing processors (including accelerators) closer to memory for reduced power and latency. Working with the PCIe-physical layer and with associations with Open CAPI and GenZ will allow performance boosts and latency improvement both in the box and networked applications.” – Excerpt from CXL Enables Heterogenous Computing, Tom Coughlin, Forbes
“Key players in the data center industry have joined forces to develop a high-speed interconnect specification designed to connect data center CPUs and accelerator chips. The Compute Express Link, or CXL, is aimed at solving the challenges of improving performance and removing bottlenecks between CPUs and other components like GPUs, memory and FPGAs.” – Excerpt from Intel, Google and Others Join Forces for CXL Interconnect, Will Calvert, Data Center Dynamics
“This new interconnect, for which the version 1.0 specification is being launched today, started in the depths of Intel’s R&D Labs over four years ago, however what was made is being launched as an open standard, headed up by a consortium of nine companies. These companies include Alibaba, Cisco, Dell EMC, Facebook, Google, HPE, Huawei, Intel, and Microsoft, which as a collective was described as one of the companies as ‘the biggest group of influencers driving a modern interconnect standard.” – Excerpt from CXL Specification 1.0 Released: New Industry High-Speed Interconnect From Intel, Ian Cutress, AnandTech
“Unleashing the power of high-speed processors and accelerators requires fast and flexible interconnects that can feed compute cores with data for processing. The insatiable desire for more bandwidth is one of the most vexing problems for system architects because data movement consumes time and power, ultimately equating to higher operational costs. The Compute Express Link interconnect is designed to help lessen that burden.” – Excerpt from Intel Donates Compute Express Link, a High-Speed Protocol for PCIe 5.0, Paul Alcorn, Tom’s Hardware.
“…to address the high speed and lower latency requirements between workload accelerators and CPUs, Intel, along with companies including Alibaba, Dell EMC, Cisco, Facebook, Google, HPE and Huawei, have formed an interconnect technology called Compute Express Link (CXL) that will aim to improve performance and remove the bottlenecks in computation-intensive workloads for CPUs and purpose-built accelerators. CXL is focused to create high speed, low latency interconnect between the CPU and workload accelerators, as well as maintain memory coherency between the CPU memory space and memory on attached devices.” – Excerpt from How are Faster Networks Advancing the Next Generation of Data Centres?, Sagar Nangare, Cloud Tech
“…incorporating a group like the CXL Consortium is a fairly big deal, because this typically only happens when an industry standards group gets large enough and gains enough traction that its members are very confident the technology is soon to go into widespread use. This means that the CXL Consortium has been elevated to the same level as the USB-IF, VESA, and other standard groups.” – Excerpt from CXL Consortium Formally Incorporated, Gets New Board Members & CXL 1.1 Specification, Anton Shilov, AnandTech
Compute Express Link™ and CXL Consortium™ are trademarks of the Compute Express Link Consortium.