NVIDIA Mellanox Unveils MQM8790-HS2F InfiniBand Switch, Doubling Interconnect Speed for Demanding HPC and AI Workloads

The NVIDIA Mellanox MQM8790-HS2F is strategically positioned as a high-density, cost-optimized solution for organizations building or expanding accelerated computing infrastructures. It bridges the gap between premium NDR solutions and previous generations, offering a balanced performance-per-watt and performance-per-dollar ratio. This makes the MQM8790-HS2F InfiniBand switch solution an ideal choice for IT directors and network engineers aiming to future-proof their fabric without exceeding budget constraints.

At the heart of the MQM8790-HS2F lies robust HDR 200Gb/s technology. The switch provides 40 non-blocking QSFP56 ports, enabling flexible and scalable fabric design. Key features detailed in the official MQM8790-HS2F datasheet include:

• 200Gb/s HDR Speeds: Provides double the bandwidth per port compared to previous EDR generations, efficiently handling data-intensive traffic.
• Advanced In-Network Computing: Integrates NVIDIA SHARPv2 technology, offloading collective operations from CPUs to significantly accelerate AI and HPC applications.
• Enhanced Fabric Management: Supports full integration with NVIDIA's UFM® platform for automated provisioning, monitoring, and predictive analytics.
• Cut-Through Switching & Adaptive Routing: Ensures the lowest possible latency and maximizes bandwidth utilization for unpredictable traffic patterns.

These specifications establish the MQM8790-HS2F 200Gb/s HDR 40-port QSFP56 switch as a powerful and efficient building block for modern data centers. It is backward compatible with EDR and FDR InfiniBand, protecting existing investments.

The design of the MQM8790-HS2F directly addresses the needs of several critical deployment scenarios:

Enterprise AI and Mid-Scale HPC Clusters: For research institutions and enterprises running complex simulations or medium-scale AI training, this switch provides the necessary latency-sensitive fabric to keep GPU servers in sync, reducing job completion times.

High-Frequency Data Analytics Platforms: In financial services, telecommunications, and real-time analytics environments, the predictable microsecond-level latency of the MQM8790-HS2F ensures rapid data movement and decision-making.

Consolidated Storage and Compute Fabrics: The switch's high bandwidth and RDMA optimization make it excellent for unifying storage (NVMe-oF) and compute networks, simplifying infrastructure and reducing costs.

For procurement teams evaluating the MQM8790-HS2F for sale, its operational benefits translate into a compelling total cost of ownership through reduced power consumption, streamlined management, and enhanced application performance.

The NVIDIA Mellanox MQM8790-HS2F emerges as a versatile and powerful engine for the next wave of data center innovation. By delivering optimized performance for RDMA, HPC, and AI clusters, it enables organizations to extract maximum value from their compute investments.

To explore detailed benchmark data, configuration guides, or obtain price information, interested parties are encouraged to contact NVIDIA or an authorized distributor. The MQM8790-HS2F stands ready to empower efficient and scalable accelerated computing infrastructures worldwide.