NEXT-GEN “VEGA” ARCHITECTURE WITH THE

WORLD’S MOST ADVANCED MEMORY ARCHITECTURE

World’s most advanced GPU compute engine

and memory architecture built with cutting-

edge 14nm FinFET process, purpose-built

to handle big data sets and a diverse range of

computational workloads.

The Radeon Instinct MI25 accelerator brings in a new era of compute for the datacenter with its

Next-Gen “Vega” architecture delivering superior compute performance via its powerful parallel

compute engine and Next-Gen programmable geometry pipeline improving processing efficiencies,

Instinct MI25 provides increased performance density, while decreasing energy consumption per

operation making it the perfect solution for today’s demanding workloads in the datacenter.

The Radeon Instinct MI25 design is a passively-

cooled accelerator design for large-scale server

deployments.

The Radeon Instinct MI25 accelerator has

advanced out-of-band manageability circuitry

for simplified GPU monitoring in large scale

systems. The MI25’s manageability capabilities

provide accessibility via I2C, regardless of

what state the GPU is in, providing advanced

monitoring of a range of static and dynamic

GPU information using PMCI compliant data

structures including board part detail, serial

numbers, GPU temperature, power and other

information.

NEXT-GEN “VEGA”

ARCHITECTURE

PASSIVELY

COOLED

REMOTE MANAGEABILITY

CAPABILITIES

Combined with AMD’s state of the art Infinity

Memory Engine with a newly designed High

Bandwidth Cache (HBC) and controller, the

memory with 484 GB/s of memory bandwidth.

HBM2: ULTRA-HIGH

MEMORY BANDWIDTH

AMD’s ROCm platform provides a scalable,

fully open source software platform optimized

for large-scale heterogeneous system

deployments with an open source headless

Linux driver, HCC compiler, rich runtime based

on HSA, tools and libraries.

ROCm OPEN SOFTWARE

PLATFORM

Design support for AMD’s MxGPU SR-IOV

hardware virtualization technology, the

Radeon Instinct MI25 provides a VDI solution

with dedicated user GPU resources, data

security and version control. Plus, a cost

effective licensing model with no additional

hardware licensing fees, and a simplified native

driver model ensuring operating system and

application compatibility.

MxGPU SR-IOV HARDWARE

VIRTUALIZATION

©2017 Advanced Micro Devices, Inc. All rights reserved. AMD, the AMD arrow logo, Radeon, and combinations thereof, are trademarks of Advanced Micro Devices, Inc.

All other product names are for reference only and may be trademarks of their respective owners. “Vega” and “Vega 10” are internal architecture code names and not products.

For more information, visit:

Radeon.com/Instinct

ROCm.github.io

Up to 1.3x More FP16 TFLOPS

Up to 1.3x More FP32 TFLOPS

Nvidia Tesla P100-16

Radeon Instinct MI25

18.7

24.5

Peak FP16 TFLOPS

Nvidia Tesla P100-16

Radeon Instinct MI25

9.3

12.3

Peak FP32 TFLOPS