- Mia
- July 21, 2023
- 8:07 am
In Superpod networking, if I configure 4 NDR200 cards on each server, can I directly connect to the same switch with a 1-to-4 cable? Or do I have to use two 1-to-2 cables to connect to different switches?
FiberMall
Answered on 8:07 am
Not recommended. This connection mode does not comply with SuperPoD networking rules. Considering the NCCL/SHARP performance, Leaf switches need to connect NDR200 ports of different servers in the SU with one-quarter cables to form different communication rings.
People Also Ask
Related Articles
800G SR8 and 400G SR4 Optical Transceiver Modules Compatibility and Interconnection Test Report
Version Change Log Writer V0 Sample Test Cassie Test Purpose Test Objects：800G OSFP SR8/400G OSFP SR4/400G Q112 SR4. By conducting corresponding tests, the test parameters meet the relevant industry standards,
The Revolutionary Nvidia DGX GH200: Powering the Future of AI Supercomputers
Nvidia DGX GH200 represents a paradigm shift in artificial intelligence (AI) and machine learning, ushering in a new chapter for AI supercomputers. It has been designed as a cutting-edge system
Ethernet-Based GPU Scale-UP Networks
The recent launch of Intel’s Gaudi-3, which utilizes RoCE for Scale-UP interconnection, along with Jim Keller’s discussions on replacing NVLink with Ethernet, has brought attention to this innovative approach. Notably,
NVIDIA H100 vs A100: Unveiling the Best GPU for Your Needs
Within artificial intelligence (AI) and high-performance computing (HPC), there is a fast-changing world where the perfect graphical processing unit (GPU) can make or break your compute-intensive application’s performance. Two of
Unlock the Power of AI with Nvidia H100: The Ultimate Deep Learning GPU
Unlock the Power of AI with Nvidia H100: The Ultimate Deep Learning GPU In the fast-changing world of artificial intelligence (AI) and deep learning, there has been a spike in
How NVIDIA GB200 Utilizes 800G/1.6T DAC/ACC
NVIDIA has released the latest GB200 series compute systems, with significantly improved performance. These systems utilize both copper and optical interconnects, leading to much discussion in the market about the
NVIDIA GB200 Analysis: Interconnect Architecture and Future Evolution
GB200 Interconnect Architecture Analysis NVLink Bandwidth Calculation NVIDIA has a lot of confusion in the calculation of NVLink transmission bandwidth and the concepts of SubLink/Port/Lane. Typically, the NVLink bandwidth of