← Back to Products
Professional
微算-P
Professional-grade compute center for mid-scale AI training & inference, industrial edge computing. Multi-node cluster with EBOF, up to 12 PFLOPS.
Specifications
ComputeMulti CPU+GPU Cluster
Storage16×3.84TB NVMe SSD, EBOF
Network100G RDMA, RoCEv2
OutputUp to 12 PFLOPS
Deployment48-72h
ScalingHot-swap, zero downtime
Use Cases
Mid-scale AI Training
Enterprise model fine-tuning, multi-GPU parallel
Industrial Edge
AI visual inspection, predictive maintenance, digital twin
Autonomous Driving
Perception algorithm training, data stays local
Medical Imaging
CT/MRI AI diagnosis, zero patient data leaving hospital