AMD Radeon VEGA: A New Graphics Architecture
“After the uprising” was the teaser trailer from AMD to introduce Radeon VEGA, the trailer showcases “make some noise” theme, to AMD’s end of bargain, VEGA does make some noise. Radeon VEGA is a set of graphics architecture, designed to go beyond gaming related applications. It covers gaming, VR, professional design applications and machine intelligence areas.
We covered the leak about AMD Radeon Vega that came few days before the official convention of AMD at CES 2017. Now, all details are unfolded by AMD itself as what exactly AMD VEGA is capable off. Senior Vice President and Chief architect Raja Koduri said:
It is incredible to see GPUs being used 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.
In Nutshell following are the highlights of official convention about Radeon VEGA:
- The world’s most advanced GPU memory architecture:
VEGA enables new memory hierarchy for GPUs, it comes in the form of new high bandwidth cache and its controller. The feature doubles the bandwidth per-pin. It enables greater capacity at less than half the footprint of GDDR5 memory. VEGA is optimized for streaming large datasets, plus it can work with variety of memory types. Upto 512 TB of virtual address space.
- Next-generation geometry pipeline:
Vega’s next generation pipeline enables the programmer to extract incredible efficiency in processing today’s complex geometry. Plus, delivering more than 200% of the throughput-per-clock over previous Radeon architectures. It also contains improved load balancing to deliver consistent performance.
- Next-generation compute engine :
At the very core of Radeon VEGA architecture is a new next-gen compute engine built on flexible compute units that can natively process 8-bit, 16-bit, 32-bit and 64-bit operations in each clock.
- Advanced pixel engine:
The new Radeon VEGA pixel engine employs a Draw Stream Binning Rasterizer. Meaning, it is designed to improve performance and power efficiency. VEGA’s pixel engine is now a client of the onboard L2 cache, enabling considerable overhead reduction for graphics workload which perform frequent read-after-write operations.
The very first GPU based on this architecture according to AMD will reach market in first half of 2017.