The NVIDIA® ConnectX®-8 SuperNIC™ is optimized to supercharge hyperscale AI computing workloads. Engineered with the Riding Heat Sink (RHS) architecture for enhanced thermal efficiency, it supports 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.
Click to open expanded view






Product Highlights
NVIDIA Mellanox 900-9X81E-00EX-ST0 ConnectX®-8 VPI Adapter Card 800GbE XDR/2 x 400GbE, Single-Cage OSFP, PCIe 6.0 x16, Crypto and Secure Boot, Tall&Short Bracket
Product Spotlights
RDMA and RoCEv2 accelerations
GPUDirect RDMA
GPUDirect Storage
Advanced, programmable congestion control
High-speed packet reordering
In-network computing
MPI accelerations (Burst-buffer, Collective, Atomic, Rendezvous)
NCCL, HPC-X®, DOCA UCC/UCX®, Open MPI, MVAPICH2
Ethernet Accelerated Switching and Packet Processing™ (ASAP²) for SDN and VNF
Single Root I/O Virtualization (SR-IOV)
Open vSwitch (OVS) acceleration
Overlay network accelerations: VXLAN, GENEVE, NVGRE
Hierarchical QoS, header rewrite, flow mirroring, flow-based statistics, flow aging
Connection tracking (L4 firewall) and NAT
Stateless TCP offloads: IP/TCP/UDP checksum
LSO, LRO, GRO, TSS, RSS
Secure boot with hardware root of trust
Secure firmware update
Flash encryption
Device attestation (SPDM 1.1)
Inline crypto accelerations: IPsec, MACsec, PSP