As the complexity of embedded systems grows, so does the attraction of using object-oriented dynamic languages, like Python or Smalltalk, to implement them. To make this practical, a reduction in the energy and cost of the required computing resources for such languages has become a hot research topic. This project addresses these issues by designing a processor specifi-cally for Smalltalk-80, by optimizing this processor for adaptive compilation, by using both finegrained and coarse grained parallelism to do more work at lower clock speeds and by taking advantage for the reconfigurability of Field Programmable Gate Arrays (which are increasingly present in embedded systems) to adapt the hardware at runtime to variable computing loads.
As the complexity of embedded systems grows, so does the attraction of using object-oriented dynamic languages, like Python or Smalltalk, to implement them. To make this practical, a reduction in the energy and cost of the required computing resources for such languages has become a hot research topic. This project addresses these issues by designing a processor specifi-cally for Smalltalk-80, by optimizing this processor for adaptive compilation, by using both finegrained and coarse grained parallelism to do more work at lower clock speeds and by taking advantage for the reconfigurability of Field Programmable Gate Arrays (which are increasingly present in embedded systems) to adapt the hardware at runtime to variable computing loads.