CXL - AshokBhat/notes GitHub Wiki

About

• Compute Express Link (CXL)
• High speed and low latency interconnect - Between CPU and workload accelerators
• Enable heterogenous compute for key workloads - Networking, storage, graphics, machine learning
• Supported by all CPU vendors - Arm, Intel, AMD, IBM
• Alternatives - CCIX, Gen-Z