Hard Real-Time
Embedded 
Multiprocessor Platform – RTEMP

Contact: Prof. Martin SchoeberlProf. Jens Sparsø

Many products and systems have built-in intelligence, realized by computers and software. It denotes those “embedded systems,” and much of the product innovation happens in these years relies on the use of embedded systems. Insulin pens and ABS brakes are a few examples that are also safety-critical – i.e. they may not be able to perform unintended actions and to react in time. It is known that computers have become faster and faster.

Many of the mechanisms we have implemented to achieve this, unfortunately, means that it is virtually impossible to predict the running time of a program and that there is a significant difference in the time you will typically see, and the time can be guaranteed. Moreover, there is no correlation between them. The computers you use, is often unnecessarily cumbersome, and the temporal analyses carried out, is very resource intensive.

The project takes the root of the problem by developing a computer platform that is optimized to reduce a program’s worst-case running time and to make it much easier to analyze the running time. The project focuses on so-called multi-core platforms and aims realization using.

FPGA chips – standard chips that can be configured to implement any hardware design. The project’s focus on embedded systems and FPGA technology is highly relevant to a wide range of Danish companies.

Partners

Project funded by:

The Danish Council for Independent Research | Technology and Production Sciences (FTP)
Project duration: April 1st, 2013– April 30th, 2016

Researchers at DTU Compute

Find out more by visiting the project website.