NVIDIA Mellanox MQM8700-HS2F Quantum HDR 200Gb/s InfiniBand Switch - 40 x QSFP56 ports, 1U, internally managed, P2C airflow. The MQM8700-HS2F is the internally managed member of the NVIDIA Quantum QM8700 series (NVIDIA ordering P/N 920-9B110-00FH-0MD; confirm on the label). It provides 40 HDR 200Gb/s QSFP56 ports in 1U with 16Tb/s of non-blocking switching capacity and sub-130ns port-to-port latency, and carries an x86 dual-core CPU running MLNX-OS with an on-board subnet manager for fabrics of up to 2,048 nodes, so it can bring a cluster up with no external SM appliance. The F suffix denotes P2C forward airflow. InfiniBand only. Refurbished, supplied by T.E.S IT-SOLUTIONS.
Engineering Context
The NVIDIA Quantum HDR switch device forwards 7.2 billion packets per second - 390 million pps per port - at sub-130ns port-to-port latency across 40 QSFP56 ports of 200Gb/s. With ConnectX-6 adapters it supports HDR100 by using two pairs of two lanes per physical port, presenting up to 80 ports of 100Gb/s from one 1U chassis, enough for more than 40 servers per rack and often enough to avoid a third fat-tree tier. Quantum silicon implements SHARP in-network computing for MPI and NCCL collective offload, plus adaptive routing, advanced congestion management and SHIELD self-healing. Management is on the front panel: micro-USB, one RJ45 Ethernet management port and an RJ45 console, driving MLNX-OS via CLI, Web UI and SNMP.
Deployment and Use Cases
Leaf, spine and standalone positions in HDR fat-tree fabrics where the switch itself owns the subnet manager. AI training fabrics running NCCL collectives over GPUDirect RDMA for distributed PyTorch, Megatron-LM and DeepSpeed on A100-generation GPU nodes. HPC clusters running MPI (HPC-X, Open MPI, MPICH) for CFD, molecular dynamics, weather modelling and seismic processing. NVMe-oF and parallel filesystem backends - Lustre, IBM Spectrum Scale, BeeGFS - at 200Gb/s per host. HDR100 double-density racks fanning out to 80 x 100Gb/s endpoints. Small and medium clusters needing a complete fabric from one box.
Technical Specifications
Brand / MPN: NVIDIA Mellanox, MQM8700-HS2F
NVIDIA ordering P/N: 920-9B110-00FH-0MD (confirm on label)
Platform: NVIDIA Quantum, QM8700 series, 1U fixed configuration
Ports: 40 x QSFP56, HDR 200Gb/s per port
Aggregate capacity: 16Tb/s non-blocking
Port-to-port latency: sub-130ns
Packet rate: 7.2 billion packets/s (390 million pps per port)
HDR100 split: up to 80 x 100Gb/s ports with ConnectX-6
Native port speeds: 40, 56, 100, 200Gb/s
Protocol: InfiniBand only
In-network computing: SHARP, adaptive routing, congestion control, SHIELD
Management: internally managed - x86 dual-core CPU, MLNX-OS, CLI / Web UI / SNMP
Subnet manager: on-board SM for fabrics of up to 2,048 nodes; UFM optional
Management ports: front micro-USB, 1 x RJ45 Ethernet management, RJ45 console
Airflow: P2C (power-to-connector) - air enters at the PSU side, exits at the port side
Power supplies: 2 x hot-swap redundant AC, C14 inlets
Fans: 6 x hot-swap, N+1 redundant
Physical: 1U, 44 x 433 x 591mm, 12.5kg
Condition: Refurbished - tested by T.E.S IT-SOLUTIONS
Compatibility and Parent Systems
Endpoint adapters: ConnectX-6 and ConnectX-6 VPI at HDR 200Gb/s and HDR100, ConnectX-7 at HDR, BlueField-2 InfiniBand DPUs; ConnectX-5 connects at EDR 100Gb/s. Cabling is QSFP56: MCP1650-H passive DACs, MFS1S00-H AOCs, MMA1T00-HS transceivers and MCP7H50-H / MFS1S50-H HDR-to-2x HDR100 splitters. Companion switch: MQM8790, the externally managed sibling, which relies on this switch or UFM for subnet management. Long-haul extension of the same generation: MetroX-2 MTQ8100-HS2F and MTQ8200-HS2F. Predecessor: Switch-IB 2 SB7800 / SB7890 at EDR. Successor: Quantum-2 QM9700 / QM9790 at NDR 400Gb/s on OSFP cages.
What This Is NOT
NOT the externally managed MQM8790. MQM8790-HS2F has no management CPU and no management ports and needs an external subnet manager.
NOT an Ethernet or VPI switch. Quantum is InfiniBand only; the Ethernet equivalent is Spectrum-2 / Spectrum-3.
NOT an NDR switch. Per-port maximum is HDR 200Gb/s on QSFP56; NDR 400Gb/s requires Quantum-2 with OSFP cages and there is no upgrade path.
NOT a C2P airflow unit. The F in -HS2F is P2C; the reverse-airflow model is MQM8700-HS2R. PSU and fan modules are airflow-specific.
NOT unlimited in subnet-manager scale. The on-board SM is rated to 2,048 nodes; larger fabrics need NVIDIA UFM.
NOT supplied with transceivers, cabling, rail kit or software entitlements unless listed on the quotation.
Frequently Asked Questions
Q. Is the MQM8700-HS2F managed, and how many nodes can its subnet manager handle?
Yes - it is the internally managed model. An x86 dual-core CPU runs MLNX-OS with an on-board subnet manager rated for out-of-the-box fabric bring-up of up to 2,048 nodes, plus CLI, Web UI and SNMP over a dedicated RJ45 management port. Larger fabrics, or sites wanting central telemetry, run NVIDIA UFM instead.
Q. What is the difference between MQM8700-HS2F and MQM8790-HS2F?
Management only. Both use the same Quantum silicon, 40 QSFP56 ports, 16Tb/s of non-blocking capacity and sub-130ns latency. MQM8700 adds the x86 CPU, MLNX-OS, the on-board 2,048-node subnet manager and the Ethernet and console ports. MQM8790 omits them and must be driven by an external SM or UFM.
Q. How many ports and how much capacity does the MQM8700-HS2F have?
40 QSFP56 ports at HDR 200Gb/s, 16Tb/s non-blocking capacity, 7.2 billion packets per second and sub-130ns port-to-port latency. With ConnectX-6 adapters and splitter cabling it presents up to 80 ports of HDR100 100Gb/s.
Q. Is the MQM8700-HS2F port-to-connector or connector-to-power airflow?
P2C - power-to-connector forward airflow. Air enters at the power-supply side and exhausts at the cabled port side. The reverse-airflow model is MQM8700-HS2R. PSU and fan modules are airflow-specific and every module in a chassis must match.
Q. Does the MQM8700-HS2F support SHARP in-network computing?
Yes. The NVIDIA Quantum switch device implements SHARP collective offload, executing MPI and NCCL reduction operations inside the switch silicon, alongside adaptive routing, congestion management and SHIELD self-healing.
Why T.E.S IT-SOLUTIONS
T.E.S IT-SOLUTIONS (tes-itsolutions.com) supplies refurbished NVIDIA Mellanox InfiniBand switching to B2B, HPC and AI customers across the EU and worldwide. Every Quantum switch is bench-verified before shipment: power-on, port link verification where test topology allows, PSU and fan redundancy check and cosmetic inspection. EU stock, 1-3 business day handling and technical consultation on subnet manager placement and airflow planning.