THE WORLD’S FASTEST TRAINING ACCELERATOR

The Radeon Instinct™ MI25 accelerator, designed with the most

advanced Next-Gen “Vega” GPU architecture, is the ultimate

training accelerator for large scale machine intelligence and

deep learning, along with being an optimized open compute

workhorse for single-precision HPC-class system workloads.

The MI25 delivers leadership in FP16 and FP32 performance in

a passively-cooled single GPU server card with 24.6 TFLOPS of

FP16 and 12.3 TFLOPS of FP32 peak performance through its 64

The Radeon Instinct MI25’s powerful compute engine and

advanced memory architecture, combined with AMD’s ROCm

open software platform and ecosystem, provides a powerful,

flexible heterogeneous compute solution that allows datacenter

designers to meet the challenges of a new era of compute and

Machine Intelligence.

The Next Era of Compute and Machine Intelligence

Highlights

Key Features

GPU Architecture:

AMD “Vega10”

Stream Processors:

4,096

GPU Memory:

16GB HBM2

Memory Bandwidth:

Up to 484 GB/s

Performance:

Half-Precision (FP16)

24.6 TFLOPS

Single-Precision (FP32)

12.3 TFLOPS

Double-Precision (FP64)

768 GFLOPS

ECC:

MxGPU Capability:

Yes

Board Form Factor:

Full-Height, Duel-Slot

Length:

10.5”

Thermal Solution:

Passively Cooled

Standard Max Power:

300W TDP

AMD’s Radeon Instinct™ MI25, combined with

the ROCm open software platform and MIOpen

libraries, delivers superior performance per

watt for deep learning training deployments in

OS Support:

Linux® 64-bit

ROCm Software Platform:

Yes

Programing Environment:

ISO C++, OpenCL™, CUDA (via AMD’s HIP conversion tool)

For more information, visit: Radeon.com/Instinct

Bus Interface:

PCIe® Gen 3 x16

Superior Performance Per Watt

Nvidia Tesla P100-SMX2

Nvidia Tesla P100-16

71

75

Peak FP16 GFLOPS/Watt

Radeon Instinct MI25

82

Warranty:

Note: “Vega 10” is an internal codename only