The Vega GPU features a revolutionary memory controller capable of leveraging system memory, flash-based SSDs and new storage-class memory types to create incredible new possibilities for gaming, design, and compute analysis workloads. With a dedicated High-Bandwidth Cache, the Vega GPU utilizes HBM2, the latest in graphics memory technology, to provide incredible levels of power efficiency and memory performance.
The Vega GPU introduces the Next-Gen Compute Unit that is highly optimised for incredible peak throughput across a range of data types. This NCU makes it possible to deliver higher clock speeds and higher instructions per clock resulting in up to 2X throughput increase over previous design.
The First Vega-based GPU – The Radeon Vega Frontier Edition
The Radeon™ Vega Frontier Edition is built for the innovators, creators and pioneers of the world. They are the early adopters, the people whose passion is to pursue what is new and different:
- Scientists developing algorithms for artificial intelligence (AI) or deep learning on a desktop or workstation equipped with Radeon Vega Frontier Edition, who can then deploy on servers containing Radeon Instinct™ accelerators
- Developers, and creators working with design tools who want the superior performance and are fine with not having application certification. Those who wish to have certified app support can turn to the “Vega” versions of Radeon™ Pro WX coming Q3, 2017
- Game, CGI and VR designers looking for a solution that will meet their needs throughout their entire design workflow, from asset creation to performance optimization
Radeon Vega Frontier Edition Specifications
|Radeon Vega Frontier Edition||Radeon Vega Frontier Edition|
|Next-Gen Compute Units (nCUs)||64 (4096 Stream Processors)||64 (4096 Stream Processors)|
|Peak FP32 Compute Performance||13.1 TeraFLOPS||13.1 TeraFLOPS|
|Peak FP16 Compute Performance||26.2 TeraFLOPS||26.2 TeraFLOPS|
|Memory Bandwidth||483 GB/s||483 GB/s|
|Memory Configuration||16GB High Bandwidth Cache (HBC)||16GB High Bandwidth Cache (HBC)|
XENON will be architecting a range of Workstation and Servers solutions featuring AMD’s Vega GPU’s. To customise your solution please contact XENON.Talk to a Solutions Architect