Open Platforms

Compute
MX1 OAI (Open Accelerator Infrastructure) UBB System
  • inter-module communication for scale-up
  • input/output bandwidth for scale-out
  • support disparate network architectures through OAM direct connection
  • capability to build 16, 32, 64, and 128 OAM clusters
  • OCP interconnect topologies:
    • Hybrid Cube Mesh (HCM)
    • Fully Connected (FC)
Compute
NF8260M5 (2U 4-Socket)
  • Scalable 4P Performance in a Dense 2U Form Factor
  • Support for up to two 2000W Power Supplies for increased redundancy and reduced downtime
  • Support up to eight PCI-E expansion cards, including graphic processor units and network adapters, offering increased expandability
Compute
NF8380M5 (3U 4-Socket)
  • 4-socket, up to 4* Intel Xeon Processor
  • M.2 riser flash array configurations for GPU, FPGA or NVMe
  • 10* PCIe x16
  • Up to 24 AEP memory
Storage
ON5266M5 (2OU OCP JBOD)
  • 34* hard drive 2OU JBOD
  • NVMe SSD/HDD support
  • Flexible Architecture
GPU
ON5388M5 (3OU JBOG)
  • NVlink-enabled architecture
  • Flexible topology for different applications
  • High density
Compute
ON5293M5
  • I/O expansion
  • Ideal for data acceleration
Compute
ON5163M5 (1OU 3-Node)
  • High Compute Density OCP Compute Node
  • Energy Conservation
  • Converged Architecture
Storage
SN3116M4 (1U 1-Node)
  • Open rack storage node
  • High density
  • Ultra-low power consumption
Rack
OCP Rack
  • Flexible configuration
  • Standard design
  • High density
  • Central power supply
Compute
SN5161M5 (1U 1-Node)
  • High density, support 2S SKL CPU and 12 LFF
  • Balanced compute and storage
  • Front maintenance
Compute
ON5263M5 (2OU 3-Node)
  • Tool-less design for increased efficiency
  • Flexible 1OU, 2OU form factors
  • Powerful and efficient

OCP Accepted
Compute
ON5488M5 (4OU OCP AI Node)
  • Ultra-high GPU Density
  • Energy Conservation
  • Optimization Architecture for AI
Compute
ON5273M5 (2OU 3-Node)
  • I/O Balance OCP Compute Node
  • 2OU 3-Node with 100GB Ethernet Support
  • Ideal for NFV application scenarios
Compute
ON5283M5 (2OU 3-Node)
  • High Expansion OCP Compute Node
  • Ideal for data acceleration
  • I/O expansion
Rack
ODCC (SR-AI Rack)
  • Built for scale-out applications
  • Consolidate compute, storage, networking, power and cooling into a rack-scale solution
  • supports up to 16 NVIDIA Tesla GPU cards, providing better expansion capacity
  • Can be cascade up to 4 GPU Boxes via the PCI-E Switch for up to 64 GPUs in one daisy chain, for a gigantic compressing computing resource pool