The case against general-purpose processors
With a large number of emerging applications such as implantables, wearables, printed electronics, and IoT have ultra-low area and power constraints, and these applications relying on ultra-low-power general purpose microcontrollers and microprocessors, there are drawbacks, researchers at the University of Illinois and the University of Minnesota reminded.
While general purpose processors have several advantages, such as amortized development cost across many applications, they are significantly over-provisioned for many area- and power-constrained systems, which tend to run only one or a small number of applications over their lifetime. To this end, the team makes a case for bespoke processor design, an automated approach that tailors a general purpose processor IP to a target application by removing all gates from the design that can never be used by the application.
Since removed gates are never used by an application, the researchers assert that bespoke processors can achieve significantly lower area and power than their general-purpose counterparts without any performance degradation. In addition, they said gate removal can expose additional timing slack that can be exploited to increase area and power savings or performance of a bespoke de- sign. Bespoke processor design reduces area and power by 62% and 50%, on average, while exploiting exposed timing slack improves average power savings to 65%.
Simulating neuron functions
A transistor that simulates some of the functions of neurons has been invented based on experiments and models developed by researchers at the Federal University of São Carlos in São Paulo State, Brazil, Würzburg University in Germany, and the University of South Carolina in the United States.
The device, which has micrometric as well as nanometric parts, can see light, count, and store information in its own structure, dispensing with the need for a complementary memory unit.
The team said transistors based on quantum dots can perform complex operations directly in memory, which can lead to the development of new kinds of device and computer circuit in which memory units are combined with logical processing units, economizing space, time, and power consumption.
They produced the transistor by a technique called epitaxial growth, which consists of coating a crystal substrate with thin film. On this microscopic substrate, nanoscopic droplets of indium arsenide act as quantum dots, confining electrons in quantized states. Memory functionality is derived from the dynamics of electrical charging and discharging of the quantum dots, creating current patterns with periodicities that are modulated by the voltage applied to the transistor’s gates or the light absorbed by the quantum dots.
The key feature of the device is its intrinsic memory stored as an electric charge inside the quantum dots, they said, and the challenge was to control the dynamics of the charges so that the transistor can manifest different states. Its functionality consists of the ability to count, memorize, and perform the simple arithmetic operations normally done by calculators, but using incomparably less space, time, and power.
While the transistor is not likely to be used in quantum computing because this requires other quantum effects, such as entanglement — which occurs when pairs or groups of particles are generated or interact in such a way that the quantum state of each particle cannot be described independently but depends on both or all of the particles, however far apart they may be — it could lead to the development of a platform for use in equipment such as counters or calculators, with memory intrinsically linked to the transistor itself and all functions available in the same system at the nanometric scale, with no need for a separate space for storage.