NVIDIA ConnectX-5 Ex EN - Network adapter - PCIe 4.0 x16 - 100 Gigabit QSFP28 x 2

NVIDIA ConnectX-5 InfiniBand adapter cards provide a high performance and flexible solution with up to two ports of InfiniBand and Ethernet connectivity, low latency, and a high message rate, plus an embedded PCIe switch and NVMe over fabrics offloads. These intelligent remote direct memory access (RDMA)-enabled adapters provide advanced application offload capabilities for high-performance computing (HPC), cloud hyperscale, and storage platforms.
ConnectX-5 adapter cards for PCIe Gen 3 and Gen 4 servers are available as stand-up PCIe cards and open compute project (OCP) Spec 3.0 form factors. Selected models also offer NVIDIA multi-host and NVIDIA socket direct technologies.

• Tag matching and rendezvous offloads
• Adaptive routing on reliable transport
• Burst buffer offloads for background checkpointing
• Embedded PCIe switch
• PCIe Gen 4 support
• RoHS compliant

• HPC environments
  ConnectX-5 offers enhancements to HPC infrastructures by providing MPI and SHMEM/PGAS and rendezvous tag matching offload, hardware support for out-of-order RDMA write and read operations, as well as additional network atomic and PCIe atomic operations support. ConnectX-5 enhances RDMA network capabilities by completing the switch adaptive-routing capabilities and supporting data delivered out-of-order, while maintaining ordered completion semantics, providing multipath reliability, and efficient support for all network topologies, including DragonFly and DragonFly+. ConnectX-5 also supports burst buffer offload for background checkpointing without interfering in the main CPU operations, and the innovative dynamic connected transport (DCT) service to ensure extreme scalability for compute and storage systems.
• Storage environments
  NVMe storage devices are gaining popularity, offering very fast storage access. The NVMe over Fabrics (NVMe-oF) protocol leverages RDMA connectivity for remote access. ConnectX-5 offers further enhancements by providing NVMe-oF target offloads, enabling highly efficient NVMe storage access with no CPU intervention, and thus improved performance and lower latency. Standard block and file access protocols can leverage RDMA for high-performance storage access. A consolidated compute and storage network achieves significant cost-performance advantages over multi-fabric networks.
• Adapter card portfolio
  ConnectX-5 InfiniBand adapter cards are available in several form factors, including low-profile stand-up PCIe, open compute project (OCP) Spec 3.0. NVIDIA multi-host technology allows multiple hosts to be connected into a single adapter by separating the PCIe interface into multiple and independent interfaces. The portfolio also offers NVIDIA socket direct configurations that enable servers without x16 PCIe slots to split the card's 16-lane PCIe bus into two 8-lane buses on dedicated cards connected by a harness. This provides 100 Gb/s port speed even to servers without a x16 PCIe slot. Socket direct also enables NVIDIA GPU Direct RDMA for all CPU/GPU pairs by ensuring that all GPUs are linked to CPUs close to the adapter card, and enables Intel DDIO on both sockets by creating a direct connection between the sockets and the adapter card.

NVIDIA ConnectX-5 InfiniBand adapter cards provide a high performance and flexible solution with up to two ports of InfiniBand and Ethernet connectivity, low latency, and a high message rate, plus an embedded PCIe switch and NVMe over fabrics offloads. These intelligent remote direct memory access (RDMA)-enabled adapters provide advanced application offload capabilities for high-performance computing (HPC), cloud hyperscale, and storage platforms.