source: anr/section-1.tex @ 23

% les objectifs globaux, 
A digital system is an application integrated into one or several chips.
These chips can be embedded in devices such as a personal digital assistant 
(PDA), ambiant computing component, wireless sensor network (WSN). They can 
also be used on a board connected to a PC to accelerate an application like 
in High-Performance Computing (HPC) and in High-Speed Signal Processing (HSSP). 
Digital system design has  been investigated since eighties by using Applications 
Specific Integrated Circuits (ASIC), Digital Signal Processing (DSP) and parallel computing on 
multiprocessor machines or networks.  More recently, since the end of nineties, 
other technologies appeared like Very Large Instruction Word (VLIW), Application 
Specific Instruction Processors (ASIP), System on Chip (SoC), Multi-Processors 
SoC (MPSoC).
\\
During these last decades, digital systems are more and more reserved 
to major companies targeting high volume market due to the design and fabrication
costs of ASIC technologies due to increasing NRE (Non Recurring-Engineering) charges.
Nowadays Field Programmable Gate Arrays (FPGA), like Virtex5 from Xilinx
and Stratix4 from Altera, can implement a complete SoC with multiple processors and
several coprocessors for less than 10K euros per device.
In addition, Electronic System Level (ESL) design methodologies (Virtual Prototyping, 
Co-design, High-Level Synthesis...) become mature and allow to
automate design and to drastically decrease its cost in terms of man power.
Thus, coupling both FPGA and ESL methodologies will soon allow small and medium 
enterprises (SMEs) to get into new and low-volume markets, to design highly innovative devices, 
to prototype complete complex embedded systems, to realize HPC or HSSP applications.
\par
The objective of COACH is to provide an environment to design emmbedded systems and
HPC applications on FPGA devices. The COACH framework will allow designer to explore various 
software/hardware partitioning scenario of the target application through timing and functional
simulations and to generate automatically both the software and the
synthesizable description of the hardware. Exploration and design are mainly 
driven by throughput, latency and/or power consumption criteria.
The main contributions of the project are:
\begin{itemize} 
\item Targeted hardware architecture and technology: 
COACH will handle both Altera and Xilinx FPGA technologies. COACH will define 
architectural templates that can be customized by additional dedicated coprocessors and ASIPs.
The parameters of the architectural templates will be the number of CPU, the operating  system...  
%the coprocessors, the number and the size of the FIFO communication channels 
Basically, the 3 following architectural templates will be provided:
A COACH architectural template based on the MIPS of the TSAR ANR project and a VCI ring bus,
An Altera architectural template based on the NIOS and the AVALON bus,
%FIXME
% The following point has to be confirmed by XILINX
% Microblaze+OPB => ARM+Amba ???
A Xilinx architectural template based on the MICROBLAZE and the OPB bus. 
Moreover, the specification of the application will be independant of both the template 
architecture and the selected technology. 
\item Design space exploration: The COACH environment will allow to select and parametrize 
the target architecture, to define hardware/software partitioning and to profile the application. 
For each point of design space exploration, metrics such as throughput, latency, power consumption, 
area, memory allocation and data locality will be provided.
This criteria will be evaluated by using virtual prototyping and high-level estimation methodologies.
\item Hardware accelerators synthesis (HAS): COACH will allow to generate automatically hardware accelerators 
when required. Hence, High-Level Synthesis (HLS) tools, ASIP design environement and 
source-level transformations (loop transformations and memory optimisation) will be provided. 
This will allow to further explore the micro-architectural design space.
\end{itemize}
%In HPC, the kind of targeted application is an existing one running on PC.
%COACH helps designer to accelerate it by migrating critical parts into a
%SoC implemented on a FPGA plugged to the PC bus.\\
%FIXME licence a speficier
The COACH environment will be designed to abstract the hardware as much as possible to the end user.
It will thus be mainly dedicated to system designers.
Finally COACH will be developped under the General Public Licence for the software tools.
and USAGE LIBRE NON COMMERCIAL for the COACH arhitecture.
%The COACH architectural templates will be freely distributed.
%
% verrous scientifiques et techniques
\mbox{}\vspace*{.9ex}\par
System design is a very complex task this project will simplify as much as possible. 
For this purpose the following scientific and technological barriers will be addressed:
\begin{itemize}
\item The clock frequency of the coprocessors generated by the HLS must respect
the frequency of the processors and the system bus.
\item 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 must have a 
common language and coding style to avoid engineering work to the designer.
\item The main problem in HPC comes from timing performance and implementation of the communication 
between the PC and the FPGA.
%FIXME: a completer loop tranfrom?, ASIP?, ...
\end{itemize}
%
% le programme de travail
\vspace*{.9ex}\par
COACH is the result of the will of several laboratories to unify their know
hows and skills in the following domains: Operating system and hardware
communication (TIMA and CITI), SoC and MPSoC (LIP6 and TIMA), ASIP (IRISA) and
HLS (LIP6 and Lab-STICC)  and loop tranformations (LIP).
COACH does not start from scratch but relies
on the SocLib platform~\cite{soclib} with the MUTEX and DNA/OS operating system for
SoC and MPSoC prototyping, on GAUT~\cite{gaut08} and UGH~\cite{ugh08} for HLS, on 
ROMA~\cite{roma} for ASIP, on SYNTOL~\cite{syntol} and BEE~\cite{bee} for loop tranformations.
The project objective is to enhance and seamlessly integrate these tools into
a unique open source framework. 
%masking these domains and its different tools to the system designer. 
The main steps of this project are:
1) Definition of the user inputs: application description as set of communicating tasks, each
task beeing described in C++ language; architectural template with its parameters; design constraints. 
2) Definition of the internal \xcoach format for representing a task.
3) Development of a GCC pluggin for generating the \xcoach representation of a C++ task.
4) Adaptation of the existing HLS tools to read and write the \xcoach format. This will allow to 
swap from one tool to an other one and to chain them.
5) Modification of the Design System eXplorator DSX of the SocLib platform to let the user
explore the design space and then to generate the bitstream.
%FIXME : a completer
\par
The role of the industrial partners BULL, THALES, XXX is to provide real
benchmarks to guide the design of the framework and to prove that COACH is
usuable and cover a large spectrum of applications.
%
% les retombées scientifiques, techniques et économiques
\vspace*{.9ex}\par
The main scientific contributions of the project are:
to make high-level synthesis an elementary tool of system design,
to unify various synthesis techniques (same input and output formats)
allowing the designer to swap from one to an other and even to chain them
without rewritting effort,
to provide a system description independent of the target architecture and 
the FPGA family.
\par
The market of embedded system and HPC is about 4,600 M\$ today and is
estimated to 5,600 M\$ in 2012.
This market is dominated by Multi-core CPUs based solution and is controlled
by major companies that can support the very high Non Recurring Engineering (NRE)
costs involved in designing such system.
Small and medium companies can only be present in this market with GPUs based solutions that have
low NRE costs but limit the application domains.\\
COACH reduces the NRE costs to the design costs (the FPGA device being only a few
K\euro) and reduces drastically them.
So one can expect that tools targeting FPGA and dedicated to software developpers
will gain market share over Multi-core CPUs and GPUs HPC based solutions.
Moreover this market can also be boosted by small and even very small new companies
that will be able to propose embedded system and accelerating solutions for standard
software applications with acceptable prices.\\
The two major FPGA companies Altera and Xilinx expect this by supporting
and participating in this project.