AMD’s Vega graphics architecture will handle “virtually unlimited” workloads

January 09, 2017 // By Graham Prophet
Disclosing a completely new, high performance GPU architecture that is designed to remove many traditional constraints from gaming, VR, professional design and machine intelligence arenas, AMD has issued preliminary details of its forthcoming GPU architecture, Vega.

The Vega architecture enables new possibilities in PC gaming, professional design and machine intelligence that traditional GPU architectures have not been able to address effectively, AMD asserts. Data-intensive workloads are becoming the new normal, and the parallel nature of the GPU lends itself ideally to tackling them. However, processing these huge new datasets requires fast access to massive amounts of memory. The Vega architecture's memory subsystem enables GPUs to address very large data sets spread across a mix of memory types. The high-bandwidth cache controller in Vega-based GPUs can access on-package cache and off-package memories in a flexible, programmable fashion using fine-grained data movement.

 

“...to solve gigabyte-scale data problems in gaming to exabyte-scale data problems in machine intelligence... 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, Radeon Technologies Group, AMD.

 

The Vega GPU architecture’s advancements include:

- The most advanced GPU memory architecture: The Vega architecture enables a new memory hierarchy for GPUs. This approach comes in the form of a new high-bandwidth cache and its controller. The cache features leading-edge HBM2 technology which is capable of transferring terabytes of data every second, doubling the bandwidth-per-pin over the previous generation HBM technology. HBM2 also enables much greater capacity at less than half the footprint of GDDR5 memory. Vega architecture is optimized for streaming very large datasets and can work with a variety of memory types with up to 512TB of virtual address space.

- Next-generation geometry pipeline: Today’s games and professional applications make use of incredibly complex geometry enabled by the extraordinary increase in the resolutions of data acquisition devices. The hundreds of millions of polygons in