POWERVR FURIAN

Graphics processor technologies are the de-facto standard for mobile and embedded GPUs.

Improving our already industry-leading architecture

The PowerVR Furian GPU architecture is an evolution to the industry-leading PowerVR Rogue graphics architecture. Furian brings about a large number of completely changed blocks for improved performance, power and density. It encompasses the significant improvements made to Rogue architecture throughout its life span.

Designed to meet a wide variety of applications

The PowerVR Furian architecture is designed to maintain and widen the performance/mW gap over competing solutions. It is a scalable architecture designed to meet a wide variety of market segments and applications, over multiple generations of products. It ensures ease of implementation, being specifically designed for ease of layout, including support for new sub-14nm process geometries.

Designed for high performance applications like AR and VR

The Furian architecture is also designed to scale efficiently to the higher performance points required for applications such as AR and VR. It does this in a manner that mitigates performance density issues which are seen in some competing solutions. As process nodes continue to shrink, it is important that the architecture can be efficiently laid out, especially as the design scales to higher performance levels.

Related blog articles

bseries imgic technology

Back in the high-performance game

My first encounter with the PowerVR GPU was helping the then VideoLogic launch boards for Matrox in Europe. Not long after I joined the company, working on the rebrand to Imagination Technologies and promoting both our own VideoLogic-branded boards and those of our partners using ST’s Kyro processors. There were tens of board partners but only for one brief moment did we have two partners in the desktop space: NEC and ST.

Read more »
pvrtune complete

What is PVRTune Complete?

PVR Tune Complete highlights exactly what the application is doing at the GPU level, helping to identify any bottlenecks in the compute stage, the renderer, and the tiler.

Read more »