Specifications
Orderable Part Number
MQM9790-NS2F
NVIDIA SKU
920-9B210-00FN-0D0
System
NVIDIA Quantum-2 QM9790-based NDR InfiniBand switch
Ports
64 × NDR 400 Gb/s on 32 OSFP connectors
Port Split
Up to 128 × 200 Gb/s ports
Management
Externally managed (unmanaged switch; fabric run by UFM or an external subnet manager)
Airflow
Power-to-Connector (P2C), forward
Switching Capacity
25.6 Tb/s unidirectional / 51.2 Tb/s bidirectional
Packet Rate
> 66.5 billion packets per second (BPPS)
Latency
130 ns
Switch IC
NVIDIA Quantum-2
In-Network Computing
NVIDIA SHARP / SHARPv3 (fabric-managed)
Routing
Adaptive routing, congestion control, RDMA
Power Supply
1+1 redundant, hot-swappable; 200–240 VAC
Typical Power
747 W
Form Factor
1U rack mount
HPE SKU Reference
P45691-B21
Operating Temperature
0 °C to 40 °C (forward airflow)
Trade-off
Lower cost per port than the managed QM9700; requires an external fabric manager
Overview
The NVIDIA Quantum-2 QM9790 MQM9790-NS2F is the externally managed variant of the 64-port NDR switch. The hardware is the same Quantum-2 platform as the QM9700 — 64 NDR 400 Gb/s ports on 32 OSFP connectors, 51.2 Tb/s bidirectional, 130 ns latency, SHARPv3 support — but the subnet manager is not on board.
That single difference changes both cost and architecture. Removing the management complex lowers the per-port price, and it centralises fabric control in UFM or a dedicated subnet-manager host instead of distributing it across every switch. Large fabrics that already run NVIDIA UFM for orchestration typically standardise on the unmanaged build for the leaf layer and reserve managed switches where local intelligence is needed.
QS Compute supplies the QM9790 NS2F and NS2R airflow variants alongside the rest of the Quantum-2 QM97XX line, with OSFP transceivers, DAC/AOC cabling and rail kits.
Key Benefits
Lower cost per 400G port than the managed QM9700 — same silicon, no on-board subnet manager; centralised fabric control via NVIDIA UFM fits large orchestrated deployments; 51.2 Tb/s and 130 ns identical to the managed build; P2C forward airflow for front-of-rack installation; 1+1 hot-swap PSUs.
Applications
Leaf-layer fabrics in large AI and HPC clusters running UFM; cost-optimised NDR topologies; scale-out storage fabrics; research clusters where fabric policy is centralised.
Request a Quote — NVIDIA QUANTUM-2 QM9790 MQM9790-NS2F — 64-PORT NDR 400G UNMANAGED SWITCH
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →