EU-projekt skal fokusere på low-power grafik og processorer
I et ambitiøst nyt EU-projekt bliver der fokuseret på, hvordan man kan forbedrede powereffektiviteten i grafiske processeringsenheder i mobile, batteridrevne produkter (in english).
Graphics researchers at Samsung Electronics UK have teamed up with mobile graphics specialists Codeplay, Think Silicon and TU Berlin to develop a tool for enabling smartphone batteries to last longer while running advanced video games and using the camera.The EU Commission has awarded a European GPU consortium a grant of 2.97 million Euros to research and develop a novel tool chain for analysing, visualizing, and improving the power efficiency of applications on mobile Graphics Processor Units (GPUs).
The consortium includes three European technology companies: Codeplay, the Edinburgh based GPU technology company, Think Silicon (a Greek low gate-count Graphics Semiconductor IP Core company) and Samsung Electronics UK Ltd. TU-Berlin (Germany), a European University, completes the group.
The key objectives of the 2 and a half years research project are:
- Define new industry standards for resource and performance monitoring to be widely adopted by embedded hardware GPU vendors (Khronos group)
- Define a methodology for accurate power estimations for embedded GPU.
- Enhance existing Dynamic Voltage and Frequency Scaling (DVFS) mechanism for optimum power management with sustained performance.
- To improve the power efficiency of compute and graphics applications running on mobile GPUs
- Build a unique power and performance visualization tool which informs application and GPU device driver developers of potential power and performance improvements.
Andrew Richards, Codeplay CEO explains:
- Working within this expert team across business and academia to analyse power consumption of videogames and camera processing is a fabulous opportunity for us at Codeplay. It will enable us to solve a very challenging problem: lengthening battery life of smartphones while running the most advanced graphics processing.
Ben Juurlink, project coordinator and professor at TU Berlin, adds:
- Searching for the performance and energy bottlenecks in applications running on embedded GPUs is like searching for a needle in a haystack. It is absolutely crucial that application developers are supported in this challenging task by smart analysis and visualization tools. Current embedded GPUs are powerful enough to execute immersive applications that we could only dream of a few years ago. However, all this compute power is good-for-nothing if the battery lasts for a few minutes only.
