[12] | 1 | \begin{figure}\leavevmode\center |
---|
| 2 | \includegraphics[width=.8\linewidth]{architecture-csg} |
---|
[105] | 3 | \caption{\label{archi-csg} Software architecture for digital system generation} |
---|
[12] | 4 | %\end{figure}\begin{figure}\leavevmode\center |
---|
| 5 | \mbox{}\vspace*{1ex}\\ |
---|
[21] | 6 | \includegraphics[width=1.0\linewidth]{architecture-hls} |
---|
[105] | 7 | \caption{\label{archi-hls} Software architecture of hardware accellerator synthesis} |
---|
[12] | 8 | %\end{figure}\begin{figure}\leavevmode\center |
---|
| 9 | \mbox{}\vspace*{1ex}\\ |
---|
| 10 | \includegraphics[width=.8\linewidth]{architecture-hpc} |
---|
[105] | 11 | \caption{\label{archi-hpc} Software architecture of HPC} |
---|
[12] | 12 | \end{figure} |
---|
[56] | 13 | %FIXME: la figure ne montre que l'aspect simulation. Intégrer la partie génération (PC API, PCIX, FPGA-IP, bridge vers VCI, SoC API) serait un plus, non ? |
---|
[12] | 14 | % |
---|
[33] | 15 | Figures~\ref{archi-csg}, \ref{archi-hls} and \ref{archi-hpc} |
---|
[65] | 16 | summarize the software architecture of the COACH framework we will develop. |
---|
[12] | 17 | In figures, the dotted boxes are the softwares or formats that COACH |
---|
[65] | 18 | has to provide and to support. |
---|
[99] | 19 | \parlf |
---|
[56] | 20 | For the system generation presented in figure~\ref{archi-csg}, the conductor |
---|
[21] | 21 | is the tool \verb!CSG! (COACH System Generator). Its inputs are a process |
---|
[119] | 22 | network describing the target application and the synthesis parameters. |
---|
[21] | 23 | The main parameters are the target hardware architectural template |
---|
[119] | 24 | with its instantiation parameters, the hardware/software mapping of the |
---|
[21] | 25 | tasks, the FPGA device and design constraints. |
---|
[119] | 26 | \verb+CSG+ thus requires an architectural template library, an operating system |
---|
[21] | 27 | library, two system hardware component (CPU, memories, BUS...) libraries |
---|
| 28 | (one for synthesis, one for simulation). |
---|
| 29 | For generating the coprocessor of a task mapped as hardware, \verb+CSG+ |
---|
| 30 | controls the HAS tools described below. |
---|
| 31 | From these inputs \verb!CSG! can generate the entire system (both software \& |
---|
[183] | 32 | hardware) either as a SystemC simulator (cycla accurate and/or TLM) to prototype and explore quickly the |
---|
[21] | 33 | design space or as a bitstream\footnote{COACH generates synthesizable VHDL, and |
---|
[134] | 34 | launch the \xilinx or \altera RTL synthesis tools.} directly downloadable on the |
---|
[56] | 35 | FPGA device\footnote{Additional partial bitstreams are generated in case of |
---|
| 36 | dynamic partial reconfiguration}. |
---|
[99] | 37 | \parlf |
---|
[21] | 38 | The software architecture for HAS is presented in figure~\ref{archi-hls}. |
---|
| 39 | The input is a single task of the process network. The HAS tools do not work |
---|
[12] | 40 | directly on the C++ task description but on an internal format called |
---|
[38] | 41 | \xcoach generated by a plugin into the GNU C compiler (GCC). |
---|
[183] | 42 | This will allow on the one hand to insure that all the tools will |
---|
[132] | 43 | accept the same C++ description and on the other hand make possible |
---|
[21] | 44 | their chaining. The front-end tools read a \xcoach description and generate |
---|
| 45 | a new \xcoach description that exibits more parallelism or implement |
---|
[132] | 46 | specific instructions for ASIP. The back-end tools read an \xcoach |
---|
| 47 | description and generate an \xcoachplus description. This is an \xcoach |
---|
| 48 | description annotated with hardware information (scheduling, binding) required by |
---|
[21] | 49 | the VHDL and systemC drivers. |
---|
| 50 | Furthermore, the back-end tools uses a macro-cell library (functional and memory |
---|
| 51 | unit). |
---|
[99] | 52 | \parlf |
---|
[21] | 53 | In addition to digital system design, HPC requires a supplementary |
---|
| 54 | partitioning step presented in figure~\ref{archi-hpc}. The designer |
---|
| 55 | splits the initial application (tag 1) in two parts: one still on the PC and the |
---|
| 56 | other running in a FPGA plugged on the PCI/X PC bus. The two parts exchange data |
---|
| 57 | through communication primitives (tag 2) implemented in a library. |
---|
[65] | 58 | To evaluate the relevance of the partitioning, the designer can build a |
---|
[21] | 59 | simulator. Once the partitioning is validated, the design of the FPGA part |
---|
| 60 | is done through \verb!CSG! (figure~\ref{archi-csg}). |
---|
[99] | 61 | \parlf |
---|
| 62 | The project is split into 8 tasks numbered from 1 to 8. There are described |
---|
[132] | 63 | below and detailled in section \ref{task-description}. |
---|
[99] | 64 | \begin{description} |
---|
| 65 | \item[Task-1: \textit{Project management}] |
---|
[132] | 66 | This task relates to the monitoring of the COACH project. |
---|
[99] | 67 | \item[Task-2: \textit{\Backbone}] This task tackles the fundamental points of the |
---|
[21] | 68 | project such as the defintion of the COACH inputs and outputs, |
---|
| 69 | the internal formats (e.g. \xcoach), the architectural templates and |
---|
| 70 | the design flow. |
---|
[183] | 71 | \item[Task-3: \textit{System generation}] This task addresses the prototyping and |
---|
| 72 | the generation of digital system. Apart from HAS that belongs to task 3 |
---|
[21] | 73 | and 4, its components are those presented figure~\ref{archi-csg} |
---|
| 74 | (e.g. \verb!CSG!, operating systems). |
---|
[99] | 75 | \item[Task-4: \textit{HAS front-end}] This task mainly focusses on four functionalities: |
---|
[21] | 76 | optimization of the memory usage, parallelism enhancement through loop |
---|
| 77 | transformations, coarse grain parallelization and ASIP generation. |
---|
[99] | 78 | \item[Task-5: \textit{HAS back-end}] This task groups two functionalities: |
---|
[21] | 79 | High-Level Synthesis of data dominated description and HLS of control |
---|
| 80 | dominated description. |
---|
| 81 | This task contains also the development of a frequency adaptator |
---|
| 82 | that will allow the coprocessors to respect the processor \& the bus |
---|
| 83 | frequency. |
---|
[99] | 84 | \item[Task-6: \textit{PC/FPGA communication middleware}] |
---|
[132] | 85 | This task pools the features dedicated to HPC. These are mainly the |
---|
[183] | 86 | validation of the partitioning (see figure~\ref{archi-hpc}), the sytem drivers for |
---|
[56] | 87 | both PC and FPGA-SoC sides, the hardware communication components and |
---|
[183] | 88 | the support for dynamic partial reconfiguration. |
---|
[99] | 89 | \item[Task-7: \textit{Industrial demonstrators}] |
---|
[33] | 90 | This task groups the demonstrators of the COACH project. |
---|
[132] | 91 | Most of them are industrial applications that will be developped within |
---|
| 92 | the COACH framework. |
---|
| 93 | Others consist in integrating the COACH framework as a driver of |
---|
| 94 | industrial proprietary design tools. |
---|
[183] | 95 | \item[Task 8: \textit{Dissemination}] |
---|
[132] | 96 | This task concerns the diffusion of the project results. |
---|
[99] | 97 | It mainly consists of the production of 4 COACH releases (\verb!T0+12!, \verb!T0+18!, |
---|
[183] | 98 | \verb!T0+24! and \verb!T0+36!), the publication of a tutorial and user manuals on a WEB site, the publication |
---|
| 99 | of research papers in international journals and conferences and the organization of workshops and tutorials in |
---|
| 100 | international conferences. |
---|
[99] | 101 | \end{description} |
---|
[21] | 102 | % |
---|
[12] | 103 | \begin{figure}\leavevmode\center |
---|
[21] | 104 | %\includegraphics[width=.4\linewidth]{dependence-task} |
---|
| 105 | \includegraphics[width=0.70\linewidth]{dependence-task-h} |
---|
| 106 | \caption{\label{dependence-task}Task dependencies} |
---|
[12] | 107 | \end{figure} |
---|
[65] | 108 | Figure~\ref{dependence-task} presents the tasks dependencies. |
---|
[99] | 109 | "$T_N \longrightarrow T_M$" means that $T_N$ impacts the $T_M$. |
---|
[132] | 110 | The more bold the arrow, the more important is the impact. |
---|
[21] | 111 | The graph shows: |
---|
| 112 | \begin{itemize} |
---|
[132] | 113 | \item Notwithstanding that $T4$ and $T5$ functionalities are complementary, |
---|
| 114 | their developments are independent (thanks to the \xcoach internal format). |
---|
[99] | 115 | \item $T3$ slightly depends on $T4$ and $T5$. Indeed, $T3$ may works |
---|
[132] | 116 | without $T4$ and $T5$ if we limit ourselves to digital systems without hardware |
---|
| 117 | accelerators. |
---|
| 118 | \item $T3$ strongly impacts $T6$ but $T3$ does not depend at all on |
---|
| 119 | $T6$. Hence demonstrators ($T7$) of embedded system would not be impacted if |
---|
[99] | 120 | $T6$ would fail. |
---|
[132] | 121 | \item $T2$ drives all the tasks ($T3$, $T4$, $T5$, $T6$) and is at the heart of |
---|
[65] | 122 | the COACH project. |
---|
[132] | 123 | \item The demonstrators developped in $T7$, of course strongly depends on the achievements |
---|
[105] | 124 | of the previous tasks ($T2$, $T3$, $T4$, $T5$, $T6$). |
---|
[99] | 125 | \item $T8$ and $T1$ respectively depends on and impacts all the other tasks. |
---|
[21] | 126 | \end{itemize} |
---|
[33] | 127 | This organisation offers enough robustness to insure the success of the |
---|
[99] | 128 | project except for the specification task $T2$. |
---|
| 129 | The only critical task in this chart is $T2$. \label{xcoach-problem} |
---|
[33] | 130 | However, the partners met |
---|
| 131 | 10 times (a one day meeting per month) during the last year to prepare the |
---|
| 132 | specification and the project proposal. This gives us a degree of confidence |
---|
[99] | 133 | that $T2$ will be completed in time. |
---|