Procyon GPU Server Series


Accelerate AI and HPC with NVIDIA®

GPU-accelerated computing can speed up highly parallel computing applications by 100X compared to CPU-only computing.


A single GPU node can replace several CPU cluster nodes, achieving higher power efficiency and lower overall infrastructure cost.


Atipa's Procyon G-Series servers offer flexible choices with GPU densities of up to 4 GPUs per 1U of rack space. 

Procyon SE428-32G8SXM4
  • 4U rackmount server

  • Dual-socket AMD Epyc™ 7003 processors

  • 32x DDR4 3200MHz DIMM slots

  • 2x 10GbaseT RJ45 Ports

  • 1x RJ45 dedicated IPMI LAN port

  • IPMI 2.0 with virtual media over LAN and KVM-over-LAN support

  • 8x A100-SXM4 GPUs with NVLINK interconnect

  • 8x PCI-E 4.0 x16 slots for add-on cards 

  • 6x 2.5” SATA/NVMe

  • 2200W/3000W Platinum Level Redundant Power Supplies