AMD Vega release date, specs, price news: next-gen GPU with HBM2 technology can handle large volumes of data

AMD has revealed preliminary details on its next-generation graphics processing unit (GPU) architecture, Vega, in anticipation of its arrival in the first half of 2017. According to the tech company, Vega was in development for over five years to ensure that it provides new experiences when it comes to gaming, virtual reality (VR), professional design and machine intelligence.

AMD's Vega architecture features a groundbreaking memory subsystem that allows the GPU to address large volumes of data sets across different memory types. Vega-based graphics cards also come with a new high-bandwidth cache and cache controller that is equipped with HBM2 technology. This allows the GPU to transfer terabytes of data within seconds while only using up half the footprint of the traditional graphics double data rate type five synchronous (GDDR5) memory.

"We designed the Vega architecture to build on this ability, with the flexibility to address the extraordinary breadth of problems GPUs will be solving not only today but also five years from now. Our high-bandwidth cache is a pivotal disruption that has the potential to impact the whole GPU market," said Raja Koduri, senior vice president and chief architect of Radeon Technologies Group, AMD.

AMD has not yet announced the full capabilities and technical specifications of its new GPU architecture but it did preview some of Vega's advancements including a next-generation geometry pipeline that makes processing complex geometry more efficient and balances the computer's load to deliver consistent performance. There is also a next-generation compute engine that can process 8-bit, 16-bit, 32-bit or 64-bit operations. Lastly, there is also an advanced pixel engine that is designed to improve performance and power efficiency.

Leaked presentation slides obtained by Videocardz also provide an overview of the technical specifications of AMD's upcoming Vega-based GPUs.

AMD Vega 10 will reportedly feature a 14 nm GFX9 architecture with 64 next-generation compute units (NCUs) with 12.5 teraflops in floating point performance, two HBM2 stacks with up to 16 GB of random-access memory (RAM) and a bandwidth of 512 GB/s and consumes 225 watts of power. A dual Vega 10 GPU is said to follow in the second half of 2017.

Pricing will be revealed during the launch of the Vega-based GPUs.