|
|
| |
|
PhD.
Research (Thesis abstract) |
|
|
|
Telecom systems include
components for broadband networks, mobile network infrastructures and interactive video-on-demand servers. These and other emerging applications
in similar areas are among the fastest growing segments of the system industry today.
The increasing complexity of telecom systems and the decreasing reduction of
their design time require changes in currently used design methodologies for
such systems, which are (partly) realized as an embedded solution to satisfy area/power/performance costs.
This thesis addresses digital systems, characterized by large data storage
requirements, dynamic (de)allocation of data, and intensive data transfers,
such as protocol processing applications found in telecom systems like ATM or
the system layer in multimedia applications like MPEG4. In this thesis, these systems are named {\em dynamic data-dominated} applications.
In order to reduce design time, higher levels of abstraction are needed. This thesis presents a concurrent object-oriented language and model
that meets the requirements needed for specifying dynamic data-dominated applications.
In the embedded processor design for dynamic data-dominated applications, a large part of the area is due to memory units.
Also the system power is heavily dominated by the storage and transfers. Given the importance of the data storage and transfers,
this thesis presents a systematic design methodology in which the storage related issues are globally optimized as a first step. This
methodology bridges the gap from the system-level specification down to the low-level hardware and software
levels from where existing tool chains are available.
In the complete design flow, a crucial stage is the virtual memory management
of dynamically (de)allocated data. This thesis focuses on a virtual memory management methodology for power and area exploration of
dynamic data-dominated applications.
In this thesis, two industrial applications illustrate the use of the developed specification language,
model, and the exploration methodology. These experiments demonstrate the applicability of this specification
language and model. They also demonstrate the effectiveness of this exploration methodology.
|
| |
|
|