Changeset 330


Ignore:
Timestamp:
Jan 26, 2011, 7:29:20 PM (13 years ago)
Author:
coach
Message:
 
Location:
anr
Files:
2 added
3 edited

Legend:

Unmodified
Added
Removed

  • anr/section-consortium-people.tex

    r327 r330  
    4444
    4545\peopletabularentry{\lip}
    46 responsible & Alias       & Christophe  & junior reasearcher  & HLS         & ... & Memory management, HLS \\\hline
    47 contributor & Feautrier   & Paul        & emeritus professor  & Compil.     & ... & Process scheduling and building.
     46responsible & Alias       & Christophe  & Permanent researcher at Inria  & Compilers, HPC & 15 & Memory management and FIFO construction. Irregular extensions.\\\hline
     47contributor & Feautrier   & Paul        & emeritus professor  & Compilers     & 13 & Process scheduling and building.
    4848                                                                                    Automatic Parallelization \\ \hline
    4949

  • anr/section-ressources.tex

    r325 r330  
    7575  workstations is evaluated to 4\% of the total requested ANR funding.
    7676\item [Personnel costs]
    77   The faculty members involved in the project are an emeritus
    78   professor at ENS Lyon (Paul Feautrier) and a research associate
    79   (CR2) at INRIA Rh\^one-Alpes (Christophe Alias).  The non-permanent
    80   personel required is a PhD student that will work on network process
    81   generation from polyhedral loops, then on extensions to
    82   non-polyhedral loops.  We are looking for a student with both
    83   theoretical and practical skills, that will be able to get a
    84   sufficient understanding of the polyhedral techniques and to produce
    85   a working implementation.
    86   \parlf
    87   The table below summarizes the \hommemois by
    88   deliverables and tasks for both permanent and non-permanent
    89   personnels.  The non-permanent personnels costs represent 26\% of
    90   the personnal costs. The requested funding for non
    91   permanent personnels is 100\% of the total ANR requested funding.
    92   \begin{center}\input{table_inria_compsys_full.tex}\end{center}
     77  The faculty members involved in the project are Christophe Alias
     78  (INRIA researcher) and Paul Feautrier (emeritus professor at
     79  ENS-Lyon). The non-permanent personel required is a post-doc that
     80  will work on FIFO construction, then on extensions of process
     81  construction and memory optimization to non-polyhedral loops.  We
     82  are looking for a candidate with both theoretical and practical
     83  skills, that will be able to get a sufficient understanding of the
     84  polyhedral techniques to produce a working implementation.  \parlf
     85  The table below summarizes the \hommemois by deliverables and tasks
     86  for both permanent and non-permanent personnels.  The non-permanent
     87  personnels costs represents ??\% of the personnal costs. The
     88  requested funding for non permanent personnels is 100\% of the
     89  total ANR requested
     90  funding.  \begin{center}\input{table_compsys_short.tex}\end{center}
    9391\item [Subcontracting]
    9492  No subcontracting costs.
    … …  
    299297\end{figure}
    300298\end{landscape}
     299%
     300\begin{landscape}
     301\begin{figure}
     302\begin{small}
     303\settowidth\desclen{XILINX RTL optimisation (5)}
     304\def\resstablestyletitle#1{\parbox{\desclen}{\textit{#1}}}
     305\hfill
     306\begin{minipage}[b]{.47\linewidth}\center
     307\input{table_compsys_full.tex}\vspace{.5ex}\\     \lip    \vspace{2.5ex}\\
     308\end{minipage}\hfill
     309\end{small}
     310\caption{\label{table-livrables-2} Man power in \hommemois for the deliverables (3)}
     311\end{figure}
     312\end{landscape}

  • anr/task-frontend.tex

    r328 r330  
    5959
    6060 \subtask{Automatic parallelization and memory optimization}
    61 
    6261This sub-task aims at providing a source-level optimizer in front the
    63 HLS back-end tools. The optimization are twofold:
     62HLS back-end tools. The optimizations are threefold:
    6463\begin{itemize}
    6564\item Extraction of parallelism in polyhedral loops and conversion
    … …  
    7069We will design these methods by using polyhedral techniques, as we did
    7170in the past for pure HPC optimizations. The program model is typically
    72 regular programs with affine loop bound conditions and array index
    73 functions. In a second part, we will extend the program model by using
    74 conservative approximations.
     71regular programs where loop bounds, conditions and array indices are
     72affine functions. In a second part, we will extend the program model
     73by using conservative approximations.
    7574   \begin{livrable}
    7675    \itemV{0}{6}{d}{\Slip}{Method, Preliminary Definition}
    … …  
    8584    \itemV{6}{12}{d}{\Slip} {Arrays and FIFO}
    8685    Description and specification of the FIFO construction method
    87     and local memory optimization algorithms.
     86    and the local memory optimization method.
    8887    \itemV{12}{18}{d+x}{\Slip} {Arrays and FIFO}
    8988    Preliminary implementation in the Bee framework.
    9089    Conversion of the input and output of Bee to the \xcoach format.
    91     At this step, only memory optimizations will be available.
     90    At this step, only local memory optimization will be available.
    9291    \itemL{18}{30}{d+x}{\Slip}{Arrays and FIFO}{10:0:9}
    9392    Final assessment of the method and improved version of the specification.
Note: See TracChangeset for help on using the changeset viewer.