The NVIDIA MCX653105A-HDAT ConnectX-6 InfiniBand smart adapter card delivers up to 200Gb/s connectivity in a single-port QSFP56 form factor. Featuring PCIe Gen 4.0 x16 host interface, this low-profile stand-up card provides extremely low latency, high message rate (215 million messages/sec), and advanced hardware offloads that accelerate HPC, AI, and enterprise data center workloads.

Built on the NVIDIA Quantum InfiniBand platform, ConnectX-6 integrates In-Network Computing engines, RDMA with end-to-end flow control, and block-level XTS-AES encryption (FIPS compliant). The MCX653105A-HDAT supports both InfiniBand and Ethernet protocols, enabling flexible deployment across compute and storage infrastructures with reduced CPU overhead and lower total cost of ownership.

• Up to 200Gb/s connectivity per port
• Up to 215 million messages/sec
• Ultra-low latency
• Block-level XTS-AES encryption (256/512-bit key)
• FIPS compliant
• PCIe Gen 4.0 and Gen 3.0 compatible
• In-Network Compute acceleration engines
• RDMA, send/receive semantics
• Hardware-based congestion control
• NVMe-of target and initiator offloads
• RoHS compliant, ODCC compatible
• Supports both 50G SerDes (PAM4) and 25G SerDes (NRZ)

The ConnectX-6 adapter offloads collective operations and MPI tag matching to the network, freeing CPU cycles for computation. Combined with NVIDIA PeerDirect® RDMA (GPUDirect), it enables direct GPU-to-GPU communication for AI and machine learning workloads.

Hardware offload of IEEE AES-XTS encryption/decryption saves latency and CPU utilization, with dedicated encryption keys for multi-tenant protection.

Target and initiator offloads optimize NVMe over Fabrics, enhancing CPU utilization and scalability for storage environments.

InfiniBand (200/100/50/40/25/10Gb/s) and Ethernet (200/100/50/40/25/10/1GbE).

• High-performance computing (HPC) clusters
• AI and machine learning training infrastructure
• Hyperscale cloud and enterprise data centers
• NVMe-of storage systems
• Virtualized networks including NFV
• x86, Power, Arm, GPU, and FPGA-based platforms

• Switches: NVIDIA Quantum-2, Quantum, Spectrum (InfiniBand/Ethernet)
• Host Interface: PCIe Gen 4.0 / Gen 3.0 x16
• Cables: QSFP56 passive copper, active optical, optical modules
• Operating Systems: RHEL, SLES, Ubuntu, Windows, FreeBSD, VMware
• Software: OFED, WinOF-2, DPDK, OpenMPI, MVAPICH, MPICH
• Virtualization: SR-IOV (up to 1K VFs), ASAP2, OVS offload

Up to 200Gb/s per port using InfiniBand or Ethernet (with appropriate switch/cable).

Yes, the adapter supports both protocols and can be configured accordingly.

Block-level XTS-AES 256/512-bit hardware encryption, FIPS compliant.

SR-IOV (1K VFs), ASAP2, Open vSwitch offload, and NPAR virtualization hierarchies.

Yes, it supports PCIe Gen 4.0 and is backward compatible with Gen 3.0.

• Observe ESD handling precautions during installation
• Verify PCIe slot compatibility (x16 physical)
• Use NVIDIA-qualified QSFP56 optics and cables
• Ensure adequate airflow around the adapter
• Install appropriate drivers (OFED/WinOF-2) before use

• 1-year standard warranty
• Technical support via email / phone
• Firmware updates and driver support
• RMA services for defective units
• Global logistics and delivery