top of page
Solving....
The surge in AI and HPC workloads is straining communication between CPUs and GPUs, exposing the limits of traditional network architectures in terms of scalability, performance, and energy efficiency.
BXI V3 Powers AI & HPC
Eviden launched BXI V3,the third generation of it's BullSequana eXascale Interconnect, co-developed with CEA. This "scale-out" network is designed for AI and HPC, offering ultra-low latency (<200ns), intelligent SmartNIC, and optimized CPU-GPU communicatio. By offloading tasks from processors and reducing energy use, BXI v3 boosts application speed (up to 35%) and lowers TCO.It's aligned with UEC's future Ethernet roadmap and helps meet Exascale perfomance demands sustainably.
EXASCALE
INTERCONNECT
bottom of page