Flopper.io Flopper.io
EF

Enflame Cloud Blazer POD P2110

rack
64× Yunsui i20
Two 25Gb/s links between compute nodes, as filed. The prospectus describes the cabinet as four to eight AI servers and several network switches joined by high-speed cabling; it gives no scale-up fabric for the accelerators.
Power
—
kW Total
Memory
1024
GB Total

Enflame's single-cabinet inference appliance, branded Cloud Blazer POD and described in the company's STAR Market prospectus as a single-cabinet, single-compute-node device that forms the smallest unit of a compute cluster. It integrates 64 Yunsui i20 accelerators across four to eight AI servers, each server carrying eight cards. The filing lists it for inference, air-cooled, with a PUE of 1.25 or better. Its 1TB of accelerator memory matches 64 cards at 16 GB exactly. Enflame publishes no throughput figure for this product or for the cards inside it.

Ask us where to buy this Spec sheet

We will point you at suppliers who have it. Free, and no signup.

System Details

GPU Configuration

GPU Model: Enflame Yunsui i20
GPU Count: 64 GPUs
Architecture: GCU / DTU 2.5
Interconnect: Two 25Gb/s links between compute nodes, as filed. The prospectus describes the cabinet as four to eight AI servers and several network switches joined by high-speed cabling; it gives no scale-up fabric for the accelerators.

System Specifications

Form Factor: rack
Total Power: —
Total Memory: 1024 GB
Memory Bandwidth: 52416 GB/s

Powered by Enflame Yunsui i20

This system utilizes 64 × Enflame Yunsui i20 GPUs, each delivering exceptional performance for AI and HPC workloads.

Per GPU TDP

150W

Per GPU Memory

16 GB

Process Node

12nm

Architecture

GCU / DTU 2.5

View Yunsui i20 Details →

Related Systems

Enflame Cloud Blazer POD P2010

32× Yunsui T20
Latest

Documentation & Resources

Official Datasheet

Enflame Cloud Blazer POD P2110 technical specifications

Download PDF

Typical Use Cases

AI/ML Training
High-Performance Computing
Data Analytics

The Enflame Cloud Blazer POD P2110 runs 64× Yunsui i20 GPUs over Two 25Gb/s links between compute nodes, as filed. The prospectus describes the cabinet as four to eight AI servers and several network switches joined by high-speed cabling; it gives no scale-up fabric for the accelerators..