Item Spotlights
NVIDIA/Mellanox C8180 (900-9X81E-00EX-ST0) ConnectX-8 InfiniBand & Ethernet SuperNIC, Socket-Direct XDR 800Gb/s PAM4 200G/Lane, Single-Port OSFP, PCIe 6.0 x 16, Tall Bracket
ConnectX-8 SuperNIC is optimized to supercharge hyperscale AI computing workloads. With support for both InfiniBand and Ethernet networking at up to 800 gigabits per second (Gb/s), ConnectX-8 SuperNIC delivers high-speed, efficient network connectivity, significantly enhancing system performance for AI factories and cloud data center environments.
The ConnectX-8 is not just a high-speed network adapter—it integrates native PCIe Gen6 support with an on-board PCIe switching fabric, eliminating the need for external PCIe switch chips. This design simplifies server architecture while improving overall system bandwidth, scalability, and multi-host connectivity.
ConnectX-8 integrates a Data Path Accelerator (DPA), a RISC-V–based event processor, and a multi-threaded PSA (Packet Stream Processor / Packet Switching Architecture) to enable hardware offload for packet parsing, classification, flow control, and SDN/NFV/QoS functions at the NIC level. In AI and HPC workloads, it accelerates collective operations such as All-Reduce and All-to-All communication, significantly improving distributed training efficiency across multi-GPU and multi-node environments.
ConnectX-8 supports advanced in-network computing capabilities such as SHARP (Scalable Hierarchical Aggregation and Reduction Protocol), which accelerates aggregation and reduction operations in large-scale HPC and AI clusters. The solution also provides telemetry-based routing and congestion control mechanisms, enabling more efficient utilization of network resources and significantly reducing tail latency — improving stability in large GPU clusters and cloud-scale environments.