Click to open expanded view
HPE (Mellanox) 834972-B25 Compatible 10m (33ft) 100G InfiniBand HPC Active Optical Cable (QSFP28 to QSFP28, 7.5mm)
Mellanox’s active optical cables are a cost-effective solution for high bandwidth fabrics, extending the benefits of Mellanox’s high-performance InfiniBand and Ethernet adapters and switches. Mellanox QSFP+ cable is an assembly of 4 full-duplex lanes, where each lane is capable of transmitting data at rates up to 28Gb/s per direction, providing an aggregated rate of 100Gb/s.NADDOD offers HPE (Mellanox) 834972-B25 Compatible 10m (33ft) 100G QSFP28 Active Optical Cable (AOC) for SB7800/SB7890, MCX556A-ECAT/MCX653106A-HDAT in 100 Gigabit Ethernet and Infiniband EDR Applications.
With NVIDIA/Mellanox InfiniBand NDR/HDR/EDR all series switches and NICs, NADDOD Test Center have extensively tested every part by live GPU and NVIDIA/Mellanox switching systems to guarantee its performance and ensure our InfiniBand networking products' 100% compatibility.
NADDOD's 100G InfiniBand HDR AOC provides exactly the same technology and performance standards including low power consumption, high bandwidth, high density, lowest latency, and insertion loss as NVIDIA/Mellanox, getting 120% of the performance at 50% of the price of the original. Bandwidth tests and latency tests in the HDR network environment are verified by NVIDIA/Mellanox equipment connection, and the practical cases of supercomputing application solutions associated with InfiniBand products of NADDOD service are verified to be fully comparable to the performance and quality of the original.
NADDOD continuously achieves efficient integration with NVIDIA products and solutions in cloud computing, artificial intelligence, HPC, and other areas. In data-intensive applications, NADDOD's high-performance networking products have been deployed to the core of national well-known supercomputing centers in January 2022, and the solutions include high-performance switches, optical connectors and NICs to support renowned supercomputing centers and undertake various large-scale scientific and engineering computing tasks with high-performance networking.