What This Document Is
This is a detailed technical paper exploring machine description languages and generic machine models specifically for hardware/software co-design. It delves into the challenges of creating accurate processor models that bridge the gap between high-level software development and detailed hardware implementation. The work focuses on a novel language, LISA, and its associated machine model, designed to address the needs of co-design and co-simulation environments. It’s a research-level exploration of modeling techniques for embedded systems.
Why This Document Matters
This paper is valuable for graduate students, researchers, and engineers working in the field of embedded systems design, particularly those focused on hardware/software co-optimization. It’s most useful when you need a deeper understanding of processor modeling methodologies, the trade-offs involved in different abstraction levels, and the requirements for effective co-design workflows. Professionals involved in compiler design, architecture exploration, and DSP system development will also find this a relevant resource. Accessing the full content will provide a comprehensive understanding of the LISA approach.
Topics Covered
* The challenges of bridging the gap between software and hardware modeling.
* Instruction-level hardware/software co-design methodologies.
* Behavioral pipeline modeling techniques.
* Data and control hazard detection in processor pipelines.
* Pipeline control mechanisms and parameterization strategies.
* Application of machine models to real-world signal processors.
* The role of abstract processor models in HW/SW co-design environments.
What This Document Provides
* A presentation of the LISA machine description language.
* A generic machine model framework for hardware/software co-design.
* An exploration of ASAP (As Soon As Possible) sequencer parameterization.
* An extension of standard pipeline descriptions using operation descriptors.
* A case study applying the LISA model to a specific signal processor architecture.
* Detailed discussion of the need for models tailored to specific co-design needs.
* Insights into the limitations of existing machine models and description languages.
