PowerVR G6100: small is powerful
The PowerVR G6100 is the latest addition to the highly efficient Series6 family of graphics IP cores, making it the smallest member of the ‘Rogue’ area-optimized GPUs to date. A single core, single cluster GPU, it will enable mass-market adoption of OpenGL ES 3.0 on a wide range of computing platforms.
Products supporting the latest APIs from Khronos will become a significant and growing part of a billion unit market by 2014, and we expect PowerVR to be the leading GPU IP family driving the adoption of graphics and compute APIs from Khronos, Google, Microsoft or the HSA Foundation. PowerVR Series6 is one of the first GPU architectures to achieve full conformance for OpenGL ES 3.0*.
Launched at MWC in Barcelona, the PowerVR G6100 is the smallest and most efficient GPU available on the market that can handle both graphics APIs like OpenGL ES 3.0 and DirectX9 Level 3 as well as GPU compute applications accelerated through OpenCL, Renderscript Compute or Filterscript.
Make no mistake: when it comes to the three defining characteristics of any mobile-oriented processor – performance, power, area – it delivers a unique combination of innovative architectural characteristics optimized for low power and reduced area, giving it an important competitive advantage over other traditional multicore GPUs. Furthermore, software developed for previously released Series6 cores, including the PowerVR G6200 and G6400, will be fully compatible with the PowerVR G6100. This keeps development costs across the ‘Rogue’ family to a minimum and gives application designers the opportunity to easily bring the full functionality of OpenGL ES 3.0 to all current and upcoming PowerVR-enabled devices.
PowerVR Series6: the best graphics and compute performance
PowerVR G6100 is a 128bit graphics and compute powerhouse, designed to deliver a wide range of APIs at the lowest power consumption. Its area-optimized floating point and rounding algorithm implementations make it the most efficient OpenGL ES 3.0 graphics core available today for its corresponding silicon footprint.
Until now, all PowerVR Series6 cores have had pairs of two, four or six unified shader clusters (USCs). Thanks to the architectural flexibility of the Series6 ‘Rogue’ family, the G6100 is able to integrate just one USC and pair it with a texture mapping unit. The arrays of programmable computing elements that make up the unified cluster offer high performance and efficiency while minimizing power and bandwidth requirements.
The scalar ALU design allows for a far simpler compiler and more efficient resource balancing in contrast to competing vector-based multicore solutions. These architectural features provide superior fillrates for complex, high-res UIs as well as higher peak GFLOPS for an improved image quality and fluid experience when running resource-demanding applications like games or augmented reality (AR).
The PowerVR G6100 is able to comfortably deliver the same raw fillrates as a Series5XT MP2 design. Combining that with a higher GFLOPS rate, this small but powerful Series6 GPU offers the optimal balance between silicon area and performance for designing high-resolution, multi-screen mass market devices capable of providing OpenGL ES 3.0 functionality.
PowerVR G6100, G6200 and G6400: the area-optimized cores
Imagination has always offered a wide range of design points in terms of target performance and size. For example, at its launch, the PowerVR SGX520 was the smallest OpenGL ES 2.0 conformant GPU core for the mobile market.
The PowerVR G6100 – together with the G6200 and G6400 – are area-optimized Series6 GPUs designed to deliver the best performance at a very small area for one, two and four cluster architectures respectively.
All PowerVR GPUs carefully match features and performance to market requirements, avoiding the all too common trap of delivering excessive feature sets on low end performance points just for the sake of bragging rights. Each core in the PowerVR family takes into account a target device’s feature set, efficiency, and battery requirements. We see the PowerVR G6100 ushering in a new plethora of thin and light, mass market smartphones or tablets as well as exciting new products, from affordable 4K smart TVs and set-top boxes to foldable or wearable computers, smart watches and Google Glass-like designs.
Imagination’s ‘Rogue’ architecture: the lowest power
Power consumption has become the final frontier of the mobile and embedded space. Real world measurements performed by Imagination’s customers as well as independent journalists and analysts have always placed PowerVR GPUs significantly ahead of competing solutions.
The enhanced scheduling architecture and dedicated housekeeping embedded processor together with the highly-efficient PVRTC texture compression technology keep PowerVR G6100’s interactions with the main CPU and system memory to a minimum, therefore reducing overall power consumption.
Here’s just one example of how complex, high-end OpenGL ES 3.0 graphics run comfortably on Imagination’s PowerVR Series6 GPUs.
This is because, when it comes to PowerVR GPUs, we avoid simply relying on frequency and process nodes to solve all our problems and instead optimize for lowest power, focusing on architectural gains in performance and power. For example, the PowerVR G6100 has a simpler, smaller TMU (Texture Mapping Unit). It allows it to support a wide range of DirectX9.3 texture standards, but keeps the complexity of the design vastly reduced, for specific products that do not really require the full DirectX10 feature set.
Excited about our new PowerVR Series6 line-up of graphics cores? Then don’t forget to follow us on Twitter (@ImaginationPR and @PowerVRInsider) for the latest announcements from Imagination, its partners and the mobile computing market.
*Conformance submission has been achieved on Imagination’s PowerVR Series6 test chip on Linux 3.2.0. Product is based on a published Khronos Specification, current conformance status can be found at http://www.khronos.org/conformance/adopters/conformant-products#opengles.