source: anr/section-1.tex @ 337

% les objectifs globaux, 
The market of digital systems is about 4,600 M\$ today and is estimated to
5,600 M\$ in 2012. However the ever growing application complexity involves
integration of heterogeneous technologies and requires the design of
complex Multi-Processors System on Chip (MPSoC).
\\
During the last decade, the use of ASICs (Application Specific
Integrated Circuits) appeared to be more and more reserved to high volume markets, because
the design and fabrication costs of such components exploded, due to increasing NRE (Non
Recurring-Engineering) costs.
Fortunately, FPGA (Field Programmable Gate Array) components, such as the
Virtex5 family from \xilinx or the Stratix4 family from \altera, can nowadays
implement a complete MPSoC with multiple processors and several dedicated
coprocessors for a few Keuros per device.
\\
Many applications are initially captured 
algorithmically in High-Level Languages (HLLs) such as C/C++. This has led to growing interest 
in tools that can provide an implementation path directly from HLLs to hardware. 
Thus, Electronic System Level (ESL) design methodologies (Virtual Prototyping,
Co-design, High-Level Synthesis...) are now mature and allow the automation of
a system-level design flow. Unfortunately, ESL tool development to date has primarily focused 
on the design of hard-wired devices i.e. ASICs and ASSPs (Application Specific Standard Product). 
However, the increasing sophistication of FPGAs has accelerated the need for FPGA-based ESL design 
methodologies. ESL methodologies hold the promise of streamlining the design approach by accepting 
designs written in the C/C++ language and implementing the function directly into FPGA. 
We believe that coupling FPGA technologies and ESL methodologies
will allow both SMEs (Small and Medium Enterprise) and major companies to design innovative
devices and to enter new, low and medium volume markets.
Furthermore, today there is an increasing industrial interest into IC
that integrates both hardwired CPU cores or MPSoC and a configurable area (FPGA)
such as the ATOM E600C chip (Intel).
In few a years, one can expect that such chips will become current. Even standard
general purpose CPU cores will contains a configurable area
bringing an explosion in low and medium volume markets.
\parlf
The objective of COACH is to provide an integrated design flow for the design of
multi-processors digital systems targeting FPGA devices.
It will be dedicated to system/software designers, and hide as much as possible
the hardware characteristics to the end-user.
COACH will mainly target three kinds of digital systems: 
1) embedded and autonomous application such as  personal digital assistants (PDA),
   ambient computing components, or wireless sensor networks (WSN);
2) PCI/E extension boards connected to a PC to accelerate a specific application,
   it is the domain of High-Performance Computing (HPC) and High-Speed Signal Processing (HSSP);
3) sub-system application for generating an IP to a larger system.
\parlf
%verrous scientifiques et techniques
The COACH environment will integrate several hardware and software technologies:
\begin{description}
\item[Design Space Exploration:]
    The COACH environment will allow to describe an application as a process 
        network i.e. a set of tasks communicating through FIFO channels.
        COACH will allow to map the application on a shared-memory, MPSoC architecture. 
    It will permit to easily explore the design space to help the system designer 
        to define the proper hardware/software partitioning of the application.
    For each point in the design space, metrics such as throughput, latency, power
    consumption, silicon area, memory allocation and data locality will be provided.
\item[Hardware Accelerators Synthesis (HAS):]
    COACH will allow the automatic generation of hardware accelerators when required.
    Hence, High-Level Synthesis (HLS) tools, Application Specific Instruction Processor
    (ASIP) design environments and source-level transformation tools (loop transformations
    and memory optimization) will be provided.
    This will allow further exploration of the micro-architectural design space.
    HLS tools are sensitive to the coding style of the input specification and the domain
    they target (control vs. data dominated).
    The HLS tools of COACH will support a common language and coding style to avoid
    re-engineering by the designer.
\item[Platform based design:] 
    COACH will handle both \altera and \xilinx FPGA devices.
    COACH will define architectural templates that can be customized by adding
    dedicated coprocessors and ASIPs and by fixing template parameters such as
    the number of embedded processors, the number and size of embedded memory banks
    or the embedded operating system.
    However, the specification of the application will be independent of both the
    architectural template and the target FPGA device.
    Basically, the following three architectural templates will be provided:
    \begin{enumerate}
    \item A Neutral architectural template based on the SoCLib IP core library and the
      VCI/OCP communication infrastructure.
    \item An \altera architectural template based on the \altera IP core library, the
      AVALON system bus and the NIOS processor.
    \item A \xilinx architectural template based on the \xilinx IP core library,
      the \xilinxbus system bus and the \xilinxcpu processor.
    \end{enumerate}
\item[Hardware/Software communication middleware:]
    COACH will implement an homogeneous HW/SW communication infrastructure and
    communication APIs (Application Programming Interface), that will be used for 
    communications between software tasks running on embedded processors and 
    dedicated hardware coprocessors.
\item[Interaction with the industrial world:]
    COACH will not be a closed framework but it will be opened to the industrial
    world by using the IP-XACT format \cite{IP-XACT-08} for describing the components of the
    architectural template and by providing the IP-XACT description of the generated MPSoC.
    This should facilitate the enhancement of the architectural template with IP and the
    integration of the IP produced by COACH in larger design.
\end{description}
%From the end user point of view, the specification of the application will be
%independant from both the architectural template and from the selected FPGA
%family.
\parlf
% le programme de travail
%The COACH project targets fundamental issues related to design methodologies for 
%digital systems by providing estimation, exploration and design tools targeting both 
%performance and power optimization at all the abstraction levels of the flow (system, 
%architecture, algorithm and logic).
To reach this ambitious goal, the project will rely on the experience and the
%complementariness
synergy of the partners in the following domains:
Operating system and communication middleware (\tima, \upmc),
MPSoC architectures (\tima, \ubs, \upmc),
ASIP architectures (\inria),
High Level Synthesis (\tima, \ubs, \upmc), and compilation (\lip),
HPC (\bull, \thales, \lip), tools integration in IP-XACT flow (\mds).
\\
The COACH project does not start from scratch.
It relies
on the Magillem industrial platform for the integration into IP-XACT flows,
on the SoCLib platform~\cite{soclib} for prototyping and operating systems (DNA/OS),
on the GAUT~\cite{gaut08} and UGH~\cite{ugh08} tools for HLS, 
on the ROMA~\cite{roma, RAFFIN:2010:INRIA-00539874:1} project for ASIP,
on the SYNTOL~\cite{syntol} and BEE~\cite{bee} tools for source-level analysis and
transformations,
and on the \xilinx and \altera IP core libraries.
Finally it will use the \xilinx and \altera logic and physical synthesis tools
to generate the FPGA configuration bitstreams.
%The main development steps of the COACH project are: 
%\begin{enumerate}
%   \item Definition of the end user inputs:
%    The coarse grain parallelism of the application will be described as a communicating
%    task graph, each task being described in C language.
%    Similarly the architectural templates with their parameters and the design constraints
%    will be specified.
%  \item Definition of an internal format for representing task.
%  \item Development of the GCC pluggin for generating the internal format of a
%    C task.
%  \item Adaptation of the existing HAS tools (BEE, SYNTOL, UGH, GAUT) to read and write
%    the internal format. This will allow to swap from one tool to another one, and to
%    chain them if necessary.
%  \item Modification of the DSX tool (Design Space eXplorer) of the SocLib
%    platform to generate the bitstream for the various FPGA families and architectural
%    templates.
%  \item Development of new tools such as ASIP compiler, HPC design environment and
%    dynamic reconfiguration of FPGA devices.
%\end{enumerate}
\parlf
The role of the industrial partners \bull, \thales and \mds is to provide
real use cases to benchmark the COACH design environment and to analyze the designer productivity 
improvements. 
\parlf
The COACH project will deliver an open and freely distributed infrastructure.
The architectural templates and most of the software tools will be distributed under the
GPL-like license.
The VHDL synthesizable models for the neutral architectural template
will also be freely available for non commercial use.
For industrial exploitation the technology providers are ready to propose commercial licenses,
directly to the end user, or through a third party.
\parlf
\mustbecompleted{LIST NON A JOUR}
The major FPGA companies (\xilinx and \altera) have expressed their interest for
this project.
Finally, the COACH project is already supported by a large number of SMEs, as demonstrated by the 
"letters of interest" (see Annex B), that have been collected during the preparation of the project :
ADACSYS, MDS, INPIXAL, CAMKA System, ATEME, ALSIM, SILICOMP-AQL,
ABOUND Logic, EADS-ASTRIUM.