Electronic system development efforts are spent primarily in software design. MDE addresses the productivity and cost challenges of software development. Virtual platforms can only be transitional technologies, since they are limited to the traditional vision of software development and hardware/software verification and do not embrace MDE.
Virtual platforms also heavily rely on the availability of models for each constituting hardware component. This may considerably delay the platform’s availability for software developers and thus go against the objective of early software development, as well as jeopardize its affordability.
Virtual systems are next-generation system-level paradigms that enable the realization of a common hardware and software model-based design process. Virtual systems are the realization, or “mapping” in MDA terms, of executable specifications that represent the system’s functionality and time constraints onto a high-level description of the platform resources (see Figure 1).
Virtual system technology should provide the following requirements:
• A smooth transition to MDE, with support for traditional C/C++ development.
• A fast, high-level common HW/SW and ESL/MDE simulation platform.
• An open infrastructure for integrating heterogeneous models of different types (functional, architectural) described in various formats (Matlab/Simulink, UML, SysML, SystemC, IP-XACT, others), which is called “model transformation” in MDE terms.
• A library of customizable generic models for accelerating the availability of the virtual system when specialized IP blocks are not available.
• Support for all dimensions of non-functional system requirements: performance, power consumption, memory footprint, cost, area, etc.
• The ability to automate software (code generation) and hardware (high-level synthesis) design.
For the common HW/SW and ESL/MDE simulation platform listed in the second requirement above, SystemC is best candidate. Its simulation kernel already provides the necessary base data types, threading, and time management primitives. It supports transaction-level modeling. As an extension to C++, it is natively compatible with most software programs. But SystemC alone is insufficient, since it does not provide generic system-level runtime elements necessary for describing platform resources (RTOS, bus, memory, interconnect, processor, multicore), and specialized IP models have to be developed.
Virtual system technology overcomes many of the limitations of virtual platforms. It is accessible when specialized IP models are not available. It provides the simulation capability, analysis requirements, and links to code generation and high-level synthesis. Companies that have adopted a virtual system design approach are experiencing the benefits of both ESL and MDE today.
Electronic system development efforts are spent primarily in software design. MDE addresses the productivity and cost challenges of software development. Virtual platforms can only be transitional technologies, since they are limited to the traditional vision of software development and hardware/software verification and do not embrace MDE.
Virtual platforms also heavily rely on the availability of models for each constituting hardware component. This may considerably delay the platform’s availability for software developers and thus go against the objective of early software development, as well as jeopardize its affordability.
Virtual systems are next-generation system-level paradigms that enable the realization of a common hardware and software model-based design process. Virtual systems are the realization, or “mapping” in MDA terms, of executable specifications that represent the system’s functionality and time constraints onto a high-level description of the platform resources (see Figure 1).
Virtual system technology should provide the following requirements:
• A smooth transition to MDE, with support for traditional C/C++ development.
• A fast, high-level common HW/SW and ESL/MDE simulation platform.
• An open infrastructure for integrating heterogeneous models of different types (functional, architectural) described in various formats (Matlab/Simulink, UML, SysML, SystemC, IP-XACT, others), which is called “model transformation” in MDE terms.
• A library of customizable generic models for accelerating the availability of the virtual system when specialized IP blocks are not available.
• Support for all dimensions of non-functional system requirements: performance, power consumption, memory footprint, cost, area, etc.
• The ability to automate software (code generation) and hardware (high-level synthesis) design.
For the common HW/SW and ESL/MDE simulation platform listed in the second requirement above, SystemC is best candidate. Its simulation kernel already provides the necessary base data types, threading, and time management primitives. It supports transaction-level modeling. As an extension to C++, it is natively compatible with most software programs. But SystemC alone is insufficient, since it does not provide generic system-level runtime elements necessary for describing platform resources (RTOS, bus, memory, interconnect, processor, multicore), and specialized IP models have to be developed.
Virtual system technology overcomes many of the limitations of virtual platforms. It is accessible when specialized IP models are not available. It provides the simulation capability, analysis requirements, and links to code generation and high-level synthesis. Companies that have adopted a virtual system design approach are experiencing the benefits of both ESL and MDE today.
Electronic system development efforts are spent primarily in software design. MDE addresses the productivity and cost challenges of software development. Virtual platforms can only be transitional technologies, since they are limited to the traditional vision of software development and hardware/software verification and do not embrace MDE.
Virtual platforms also heavily rely on the availability of models for each constituting hardware component. This may considerably delay the platform’s availability for software developers and thus go against the objective of early software development, as well as jeopardize its affordability.
Virtual systems are next-generation system-level paradigms that enable the realization of a common hardware and software model-based design process. Virtual systems are the realization, or “mapping” in MDA terms, of executable specifications that represent the system’s functionality and time constraints onto a high-level description of the platform resources (see Figure 1).
Virtual system technology should provide the following requirements:
• A smooth transition to MDE, with support for traditional C/C++ development.
• A fast, high-level common HW/SW and ESL/MDE simulation platform.
• An open infrastructure for integrating heterogeneous models of different types (functional, architectural) described in various formats (Matlab/Simulink, UML, SysML, SystemC, IP-XACT, others), which is called “model transformation” in MDE terms.
• A library of customizable generic models for accelerating the availability of the virtual system when specialized IP blocks are not available.
• Support for all dimensions of non-functional system requirements: performance, power consumption, memory footprint, cost, area, etc.
• The ability to automate software (code generation) and hardware (high-level synthesis) design.
For the common HW/SW and ESL/MDE simulation platform listed in the second requirement above, SystemC is best candidate. Its simulation kernel already provides the necessary base data types, threading, and time management primitives. It supports transaction-level modeling. As an extension to C++, it is natively compatible with most software programs. But SystemC alone is insufficient, since it does not provide generic system-level runtime elements necessary for describing platform resources (RTOS, bus, memory, interconnect, processor, multicore), and specialized IP models have to be developed.
Virtual system technology overcomes many of the limitations of virtual platforms. It is accessible when specialized IP models are not available. It provides the simulation capability, analysis requirements, and links to code generation and high-level synthesis. Companies that have adopted a virtual system design approach are experiencing the benefits of both ESL and MDE today.
Author Resource:-
CoFluent Studio is the only tool that offers true capabilities for hardware/software real-time co-modelingand application-level design space exploration based on prospective performance analysis allowing for architecture decisions very early on the project.
![[Valid RSS feed]](http://www.articlesinsight.com/images/icon_Rss.png "XML Feed For RSS"){kind=icon}
Category Rss Feed
Print This Article
Add To Favorites
