Skip to main content

GPU-WEAR, Ultra-low power heterogeneous Graphics Processing Units for Wearable/IoT devices

Periodic Reporting for period 3 - GPU-WEAR (GPU-WEAR, Ultra-low power heterogeneous Graphics Processing Units for Wearable/IoT devices)

Reporting period: 2018-06-01 to 2018-11-30

End users keep demanding better wearable and IoT product experiences. These depend on extended battery lifetime, higher performance, and quality of the visual experience. Over 64% of today’s end-users are not satisfied with their mobile-device battery lifetime and 30% of the first generation smartwatch buyers returned their watch because the battery lifetime was limited to 25 hours of regular use per day. High performance and vibrant display technology in sensor-based mobile and data-acquisition devices come at the expense of a short battery lifetime. The majority of todays’ available devices and their embedded SoCs are not optimized to meet user expectations in terms of power consumption.
The GPU-WEAR concept is based on a holistic approach (HW, SW, API, and compiler level) to reduce power consumption in wearable/IoT devices and enable developers to deliver power efficient applications for embedded GPUs. The new family of heterogeneous, multicore GPUs is driven by the graphics characteristics of wearable/IoT devices and displays. Both core types of NEMA heterogeneous GPUs are powered by a single and “morphable” low-power green ISA, thus a single executable and a single software/compiler toolchain. Moreover, in GPU-WEAR project we developed Display-aware and Content-aware graphics technology.
Display-aware graphics operations aim to exploit the relation between target screen properties (display size/resolution and display type technology) and accuracy of the graphics operations (precision of computations and storage data types), thus the goal is to reduce power without adversely affecting the visual quality. Context-aware technology targets to exploit the inherent imbalance in the graphics workloads among different applications. This imbalance in the graphics workload can be inherently exploited by a heterogeneous GPU leading to significant power savings.
Thanks to the GPU-WEAR project an array of innovative ultra-low power Graphics-Processing-Unit (GPU) IP-technology, together with a complete ecosystem of development-tools became a reality. The ambitious goals set back in 2015 were mostly achieved or were re-defined and successfully executed, towards satisfying technical, commercial and customer requirements.
Overview of the results
The following list reveals the achievements implemented within the GPU-WEAR project:
• Transparent “display-aware” and “QoS-aware” graphics libraries
o NEMA®|GFX-API – graphics library to accelerate Graphics User Interfaces (GUI)
o NEMA®|GUI-Builder - drag & drop tool to design Graphics User Interfaces (GUI) in a fraction of time and generates automatic C++ code.
o NEMA®|Bits – Pre-silicon FPGA evaluation kit
o NEMA®|Pix-Presso - asset management and image optimization tool for optimal visual appearance and efficient memory utilization.
o NEMA®|SHADER-Edit – Vertex and Fragment-Shader editor with compiler for easily debugging input shaders.
o GLOVE™ - Middleware which translates OpenGL® ES calls to Vulkan® API.
• Heterogeneous GPU
• NEMA®|xNN - Artificial intelligence EDGE inference accelerator
Exploitation and dissemination activities
From the beginning of the project Think Silicon has participated in 16 highly prestigious Industrial Exhibitions, 8 of which are the biggest exhibitions of the ICT sector (CES Las Vegas – 2017 and 2018, Mobile World Congress – 2017 and 2018, ARC Processor Summit 2017 and 2018, Linley Processor Conference 2017), achieving to showcase its products and technology generated with the support of the GPU-WEAR project to more than 657,000 individuals. Think Silicon also attended 18 Conferences and Workshops, which enhanced considerably the scientific impact of the project’s results, reaching more than 3,330 attendees worldwide.
Also, in total, 14 press releases have been published reaching more than 20,000 total pick-ups from Canada, China, France, Germany, Israel, Taiwan, UK and US and generating more than 4,500 web site visits. Also, more than 220,000 Think Silicon followers have received at least once an electronic notification about the GPU-WEAR related products either via our social media channels and/or our corporate newsletter. The extensive electronic publicity of the last 2.5 years has led the Think Silicon website appearing in the 1st position in the most popular search engines whenever searching for phrases like “NEMA GPU”, “GPU-WEAR”, “ultra-low power, heterogeneous, multicore GPU”, “ultra-low power vivid graphics”, etc.
Regarding the exploitation activities performed, the company was able to file ten (10) international patents (6 granted, 4 pending). Also, several meetings were held with several potential customers and/or partners achieving 3 planned production license agreements, 5 signed development license agreements, 4 signed evaluation license agreements, 1 signed marketing license agreement and 2 tape-outs. Also, 9 new releases of our products have been introduced in the market, NEMA®|p, NEMA®|GFX, NEMA®|Bits, NEMA®|GUI-Builder, NEMA®|SHADER-Edit, NEMA®|PIX-Presso, GLOVE™, the software back-end of NEMA |GFX and NEMA | TSCTMD while three products are currently finalized NEMA®|tS, NEMA®|dpu, and NEMA®|xNN (TRL 6).
One of the main achievements of the project beyond the current state of the art was to develop holistic power reduction techniques by reducing power not only on ASIC level, but also on the entire system (Display device). To achieve this, the company developed and implemented multiple techniques such as value- memorization, new image/texture/z-buffer compression methods, smart clock gating, power gating, and adaptive backlighting, to name a few.
In summary, thanks to the GPU-WEAR project Think Silicon managed to scale up expanding its product and patent portfolio, sales channel, number of employees as well as its customer base increasing its visibility and appearance on a worldwide level. As far as its product and patent portfolio is concerned Think Silicon managed, through the GPU-WEAR project and among other resources, to triple its annual turnover during the last 3 years. Also, a strategic decision was also made within this period, moving Think Silicon to the AI market with the objective to develop an ultra low power Inference Accelerator.
In the same time the company managed to grow its workforce from 14 to 29 employees, leading to a complete organizational restructuring. The GPU-WEAR project was extremely instrumental for Think Silicon to build and expand its sales channel as the company managed to open sales offices in Canada, Europe, North America, Taiwan and Japan, managing to get even closer to its customers. Also, throughout the project duration Think Silicon was extremely active towards expanding its customer base, achieving 3 planned production license agreements, 5 signed development license agreements, 4 signed evaluation license agreements, 1 signed marketing license agreement and 2 tape-outs.
The company also had the chance within the project to do a thorough market research so as to identify all potential market segments for the company’s products, selecting to focus on those with the highest potentials. In a nutshell, the GPU-WEAR project had a significant and critical social and economic impact not only on the company itself but also to its local and global business ecosystem as it boosted the company’s visibility and appearance on a worldwide level.