The past six months have certainly been an exciting time for the wearable industry. We’ve seen new platform announcements from Ingenic, Ineda and Microchip, the rollout of several smartwatches and smart glasses from worldwide sellers, and the release of the new MIPS M5100 and M5150 CPUs designed for wearables and IoT. More recently, Archos has deployed our PowerVR GPUs in a Rockchip-based Android tablet which acts as a controller and multimedia hub for the company’s Smart Home platform.
There is a certain category of wearables and IoT devices that require advanced graphics and video processing alongside the sensors, connectivity and general-purpose processors; these devices feature displays and need to run multimedia-rich apps that require smooth OpenGL® ES graphics.
Even though the potential for this market is quite impressive, there have been several issues facing wide adoption of IoT and wearables devices. This is because most of these devices are currently powered by sub-optimal solutions derived from smartphone chips which cause power consumption problems.
To address these issues, Imagination is introducing PowerVR GX5300, a GPU specifically designed for ultra-low power, low area OpenGL ES graphics rendering.
PowerVR GX5300 is a GPU optimized for extremely low area and power requirements
PowerVR GX5300 is designed to address the ultra-low power requirements of wearables. Whereas previous generation smartwatches could render graphics using a software-only approach or a simple 2D engine, newer wearable devices require fully featured GPUs to drive higher resolution user interfaces. PowerVR GX5300 can comfortably drive 480p and 720p resolutions representative of modern smartwatches, while offloading the main CPU and reducing software complexity.
Additionally, PowerVR GX5300 has been optimized to occupy the lowest silicon area possible: at only 0.55mm2 (post layout) when implemented in a 28nm process node from TSMC, it is the smallest OpenGL ES 2.0 GPU capable of running Android, Android Wear and other Linux-based operating systems. This makes it up to 2x smaller than competing single-core designs manufactured in a similar process node.
PowerVR GX5300 is based on the highly successful PowerVR Series5 family, which features a number of advantages over competing solutions, especially when it comes to power consumption:
- Unified shaders: Our TBDR graphics architecture offers unified shaders where vertex, pixel and GPU compute resources are scaled simultaneously. This approach makes our GPU adapt to any workload requirements, whereas competing solutions suffer from bottlenecks that impact processing efficiency.
- Low power and high precision graphics: All PowerVR GPUs offer a mix of low power (FP16) and high precision (FP32) rendering and implement the full OpenGL ES 2.0 specification whereas some competing solutions offer no support for FP32 which results in visual artefacts and lack of compatibility.
- Reduced memory footprint: PowerVR GX5300 supports PVRTC, a texture compression format widely adopted in the mobile industry which reduces memory bandwidth and decreases power consumption. Competing solutions offer ETC-only implementations that suffer from lack of compatibility, higher bandwidth usage that lead to power costs. PVRTC helps silicon vendors reduce memory costs, a vital aspect in wearables and IoT, where smaller is cheaper and lower power.
Use cases for GPUs in wearables and IoT
Embedded Linux or Android-based connected home systems
System architects can use PowerVR GX5300 GPUs for platforms that require graphics rendering for interactive home automation controls. For example, electronic appliances can be equipped with multimedia displays that show readings such as temperature or humidity but can also connect to the internet to bring up cake recipes, best storage practices for certain products or washing tips and tricks for clothes.
The Discovery IQ smart oven from Dacor uses a PowerVR GPU to run the Android GUI
Imagination was one of the official launch partners of Android Wear. We have been working closely with our ecosystem to optimize our processor IP and tools for operating systems designed for wearables. Silicon vendors can combine MIPS, PowerVR and Ensigma processors to create low power solutions for smartwatches, smart glasses and other connected devices.
PowerVR GX5300 is a graphics processor optimized for wearable and IoT use cases. The extremely low area and reduced power consumption makes it an ideal solution for devices that require multimedia processing yet have strict heat dissipation profiles and need to run off much smaller batteries.
We’ve made updates to our tools and IP to help system architects integrate the design much faster and ensure companies get to market faster. The many unique features of our PowerVR GX5300 GPU enable architects to create a complete multimedia platform that incorporates the smallest yet most powerful graphics hardware IP for this category of devices. We are very excited to see how this new GPU will get integrated into some next generation exciting products that we will wear on our wrists, on our clothes or in our pockets.Follow @ImaginationTech