Overview

Personal computing currently faces a rapid trend from desktop machines towards mobile services, accessed via tablets, smartphones and similar terminal devices.
With respect to computing power, today´s handheld devices are similar to Cray-2 supercomputers from the 1980s. Due to higher computational load (e.g. via multimedia apps) and the variety of radio interfaces (such as WiFi, 3G, and LTE), modern terminals are getting increasingly energy hungry. For instance, a single UMTS upload or a video recording process on today´s smartphones may consume as much as 1.5 Watts, i.e. roughly 50% of the maximal device power. In the near future, higher data rates and traffic, advanced media codecs, and graphics applications will ask for even more energy than the battery can deliver.
At the same time, the power density limit might lead to a significant share of “Dark Silicon” at 22nm CMOS and below. Obviously, disruptive energy optimizations are required that go well beyond traditional technologies like DVFS (dynamic voltage and frequency scaling) and power-down of temporarily unused components.
The GEMSCLAIM project aims at introducing novel approaches for reducing this “greed for energy”, thereby improving the user experience and enabling new opportunities for mobile computing.

The focus is on three novel approaches:

  1. cross layer energy optimization, ranging from the compiler over the operating system down to the target HW platform
  2. efficient programming support for energy-optimized heterogeneous Multicore platforms based on energy-aware service level agreements (SLAs) and energy-sensitive tunable parameters
  3. introducing energy awareness into Virtual Platforms for the purpose of dynamically customizing the HW architecture for energy optimization and online energy monitoring and accounting.
GEMSCLAIM will provide new methodologies and tools in these domains and will quantify the potential energy savings via benchmarks and a HW platform prototype.


1. GemsClaim Objectives
In a world of de-facto standards as well as huge amounts of legacy HW and SW, it is very difficult to achieve real breakthrough in system-wide energy savings beyond fragmented point solutions, e.g. at the HW or OS level.
GEMSCLAIM´s mission is to overcome this hurdle by a novel cross layer energy optimization approach that combines the following major research activities:

  1. Energy-aware optimizing and parallelizing compiler: A programming environment capable of distributing the data and work of parallel programs in an energy-optimized way on heterogeneous Multicore platforms, based on Service-Level Agreements (SLAs) negotiated with the OS and dynamic feedback about the energy behavior from the lower system layers.
  2. Component aware energy-efficient operating system: A new energy-efficient OS kernel providing energy-aware resource management via SLAs, and capable of energy-optimized task-to-processor dispatching, multitasking, and memory management, based on energy monitoring at component (e.g. processor, memory block) level.
  3. Customizable HW modelling with energy monitoring facilities: A target HW platform, capable of dynamic configuration of HW components for energy-efficiency and providing component level power metrics. The OS derives power estimations based on the metrics. For exploring the most energy-efficient HW configuration and to interact with the software tool-flow from early design phases, a virtual prototype (SW model) of the target platform will also be developed.
2. GemsClaim Outcomes
The following concrete project outcomes are expected for GEMSCLAIM:
  1. New insights into the potential of energy savings in the digital HW platforms of mobile terminals via cross layer optimization, with a targeted energy savings of 30% compared to the present state-of-the-art.
  2. New methodologies and SW tool prototypes, including an energy-aware OpenMP+ compiler, HW supported energy optimizing runtime resource manager, and Virtual Platform technology capable of energy monitoring at component level.
  3. An FPGA based HW platform demonstrator, with capabilities for dynamic HW customization for energy and real-life energy measurement for representative embedded benchmark applications.