-\documentclass[10pt, conference, compsocconf]{IEEEtran}
+\documentclass[10pt, peerreview, compsocconf]{IEEEtran}
%\usepackage{latex8}
%\usepackage{times}
\usepackage[utf8]{inputenc}
\newcommand{\tab}{\ \ \ }
-
\begin{document}
-\title{Using FPGAs for high speed and real time cantilever deflection estimation}
+\title{A new approach based on least square methods to estimate in real time cantilevers deflection with a FPGA}
\author{\IEEEauthorblockN{Raphaël Couturier\IEEEauthorrefmark{1}, Stéphane Domas\IEEEauthorrefmark{1}, Gwenhaël Goavec-Merou\IEEEauthorrefmark{2} and Michel Lenczner\IEEEauthorrefmark{2}}
\IEEEauthorblockA{\IEEEauthorrefmark{1}FEMTO-ST, DISC, University of Franche-Comte, Belfort, France\\
\{raphael.couturier,stephane.domas\}@univ-fcomte.fr}
-\maketitle
+%\maketitle
\thispagestyle{empty}
-{\it keywords}: FPGA, cantilever, interferometry.
+
\end{abstract}
+\begin{IEEEkeywords}
+FPGA, cantilever, interferometry.
+\end{IEEEkeywords}
+
+
+\IEEEpeerreviewmaketitle
+
\section{Introduction}
Cantilevers are used inside atomic force microscope (AFM) which provides high
In this paper our attention is focused on a method based on interferometry to
measure cantilevers' displacements. In this method cantilevers are illuminated
-by an optic source. The interferometry produces fringes on each cantilevers
+by an optic source. The interferometry produces fringes on each cantilever
which enables to compute the cantilever displacement. In order to analyze the
fringes a high speed camera is used. Images need to be processed quickly and
then a estimation method is required to determine the displacement of each
-cantilever. In~\cite{AFMCSEM11}, the authors have used an algorithm based on
+cantilever. In~\cite{AFMCSEM11}, authors have used an algorithm based on
spline to estimate the cantilevers' positions.
- The overall process gives
-accurate results but all the computation are performed on a standard computer
-using labview. Consequently, the main drawback of this implementation is that
-the computer is a bootleneck in the overall process. In this paper we propose to
-use a method based on least square and to implement all the computation on a
-FGPA.
+The overall process gives accurate results but all the computations
+are performed on a standard computer using LabView. Consequently, the
+main drawback of this implementation is that the computer is a
+bootleneck. In this paper we propose to use a method based on least
+square and to implement all the computation on a FGPA.
The remainder of the paper is organized as follows. Section~\ref{sec:measure}
describes more precisely the measurement process. Our solution based on the
\section{Measurement principles}
\label{sec:measure}
-
-
-
-
-
-
-
\subsection{Architecture}
\label{sec:archi}
%% description de l'architecture générale de l'acquisition d'images
interferometry is sensitive to the optical path difference induced by the
vertical displacement of the cantilever.
-The system build by authors of~\cite{AFMCSEM11} has been developped based on a
-Linnick interferomter~\cite{Sinclair:05}. It is illustrated in
-Figure~\ref{fig:AFM}. A laser diode is first split (by the splitter) into a
-reference beam and a sample beam that reachs the cantilever array. In order to
-be able to move the cantilever array, it is mounted on a translation and
-rotational hexapod stage with five degrees of freedom. The optical system is
-also fixed to the stage. Thus, the cantilever array is centered in the optical
-system which can be adjusted accurately. The beam illuminates the array by a
-microscope objective and the light reflects on the cantilevers. Likewise the
-reference beam reflects on a movable mirror. A CMOS camera chip records the
-reference and sample beams which are recombined in the beam splitter and the
-interferogram. At the beginning of each experiment, the movable mirror is
-fitted manually in order to align the interferometric fringes approximately
-parallel to the cantilevers. When cantilevers move due to the surface, the
-bending of cantilevers produce movements in the fringes that can be detected
-with the CMOS camera. Finally the fringes need to be
-analyzed. In~\cite{AFMCSEM11}, the authors used a LabView program to compute the
-cantilevers' movements from the fringes.
+The system build by these authors is based on a Linnick
+interferomter~\cite{Sinclair:05}. It is illustrated in
+Figure~\ref{fig:AFM}. A laser diode is first split (by the splitter)
+into a reference beam and a sample beam that reachs the cantilever
+array. In order to be able to move the cantilever array, it is
+mounted on a translation and rotational hexapod stage with five
+degrees of freedom. The optical system is also fixed to the stage.
+Thus, the cantilever array is centered in the optical system which can
+be adjusted accurately. The beam illuminates the array by a
+microscope objective and the light reflects on the cantilevers.
+Likewise the reference beam reflects on a movable mirror. A CMOS
+camera chip records the reference and sample beams which are
+recombined in the beam splitter and the interferogram. At the
+beginning of each experiment, the movable mirror is fitted manually in
+order to align the interferometric fringes approximately parallel to
+the cantilevers. When cantilevers move due to the surface, the
+bending of cantilevers produce movements in the fringes that can be
+detected with the CMOS camera. Finally the fringes need to be
+analyzed. In~\cite{AFMCSEM11}, authors used a LabView program to
+compute the cantilevers' deflections from the fringes.
\begin{figure}
\begin{center}
\subsection{Cantilever deflection estimation}
\label{sec:deflest}
-As shown on image \ref{img:img-xp}, each cantilever is covered by
-interferometric fringes. The fringes will distort when cantilevers are
-deflected. Estimating the deflection is done by computing this
-distortion. For that, (ref A. Meister + M Favre) proposed a method
-based on computing the phase of the fringes, at the base of each
-cantilever, near the tip, and on the base of the array. They assume
-that a linear relation binds these phases, which can be use to
-"unwrap" the phase at the tip and to determine the deflection.\\
+\begin{figure}
+\begin{center}
+\includegraphics[width=\columnwidth]{lever-xp}
+\end{center}
+\caption{Portion of an image picked by the camera}
+\label{fig:img-xp}
+\end{figure}
-More precisely, segment of pixels are extracted from images taken by a
-high-speed camera. These segments are large enough to cover several
-interferometric fringes and are placed at the base and near the tip of
-the cantilevers. They are called base profile and tip profile in the
-following. Furthermore, a reference profile is taken on the base of
-the cantilever array.
+As shown on image \ref{fig:img-xp}, each cantilever is covered by
+several interferometric fringes. The fringes will distort when
+cantilevers are deflected. Estimating the deflection is done by
+computing this distortion. For that, authors of \cite{AFMCSEM11}
+proposed a method based on computing the phase of the fringes, at the
+base of each cantilever, near the tip, and on the base of the
+array. They assume that a linear relation binds these phases, which
+can be use to "unwrap" the phase at the tip and to determine the deflection.\\
-The pixels intensity $I$ (in gray level) of each profile is modelized by :
+More precisely, segment of pixels are extracted from images taken by
+the camera. These segments are large enough to cover several
+interferometric fringes. As said above, they are placed at the base
+and near the tip of the cantilevers. They are called base profile and
+tip profile in the following. Furthermore, a reference profile is
+taken on the base of the cantilever array.
+
+The pixels intensity $I$ (in gray level) of each profile is modelized by:
\begin{equation}
\label{equ:profile}
to the maximum precision ever obtained experimentally on the
architecture, i.e. 0.3nm. Finally, the latency between an image
entering in the unit and the deflections must be as small as possible
-(NB : future works plan to add some control on the cantilevers).\\
+(NB: future works plan to add some control on the cantilevers).\\
If we put aside some hardware issues like the speed of the link
between the camera and the computation unit, the time to deserialize
Obviously, some cache effects and optimizations on
huge amount of computations can drastically increase these
-performances : peak efficiency is about 2.5Gflops for the considered
+performances: peak efficiency is about 2.5Gflops for the considered
CPU. But this is not the case for phase computation that used only few
tenth of values.\\
In order to evaluate the original algorithm, we translated it in C
-language. Profiles are read from a 1Mo file, as if it was an image
-stored in a device file representing the camera. The file contains 100
-profiles of 21 pixels, equally scattered in the file. We obtained an
-average of 10.5$\mu$s by profile (including I/O accesses). It is under
-are requirements but close to the limit. In case of an occasional load
-of the system, it could be largely overtaken. A solution would be to
-use a real-time operating system but another one to search for a more
-efficient algorithm.
-
-But the main drawback is the latency of such a solution : since each
+language. As said further, for 20 pixels, it does about 1550
+operations, thus an estimated execution time of $1550/155
+=$10$\mu$s. For a more realistic evaluation, we constructed a file of
+1Mo containing 200 profiles of 20 pixels, equally scattered. This file
+is equivalent to an image stored in a device file representing the
+camera. We obtained an average of 10.5$\mu$s by profile (including I/O
+accesses). It is under are requirements but close to the limit. In
+case of an occasional load of the system, it could be largely
+overtaken. A solution would be to use a real-time operating system but
+another one to search for a more efficient algorithm.
+
+But the main drawback is the latency of such a solution: since each
profile must be treated one after another, the deflection of 100
cantilevers takes about $200\times 10.5 = 2.1$ms, which is inadequate
for an efficient control. An obvious solution is to parallelize the
\section{Proposed solution}
\label{sec:solus}
-Project Oscar aims to provide an hardware and software architecture to
-estimate and control the deflection of cantilevers. The hardware part
-consists in a high-speed camera, linked on an embedded board hosting
-FPGAs. By the way, the camera output stream can be pushed directly
-into the FPGA. The software part is mostly the VHDL code that
-deserializes the camera stream, extracts profile and computes the
-deflection. Before focusing on our work to implement the phase
-computation, we give some general informations about FPGAs and the
-board we use.
+Project Oscar aims to provide a hardware and software architecture to estimate
+and control the deflection of cantilevers. The hardware part consists in a
+high-speed camera, linked on an embedded board hosting FPGAs. By the way, the
+camera output stream can be pushed directly into the FPGA. The software part is
+mostly the VHDL code that deserializes the camera stream, extracts profile and
+computes the deflection. Before focusing on our work to implement the phase
+computation, we give some general information about FPGAs and the board we use.
\subsection{FPGAs}
A field-programmable gate array (FPGA) is an integrated circuit designed to be
-configured by the customer. A hardware description language (HDL) is used to
-configure a FPGA. FGPAs are composed of programmable logic components, called
-logic blocks. These blocks can be configured to perform simple (AND, XOR, ...)
-or complex combinational functions. Logic blocks are interconnected by
-reconfigurable links. Modern FPGAs contains memory elements and multipliers
-which enables to simplify the design and increase the speed. As the most complex
-operation operation on FGPAs is the multiplier, design of FGPAs should not used
-complex operations. For example, a divider is not an available operation and it
-should be programmed using simple components.
-
-FGPAs programming is very different from classic processors programming. When
-logic block are programmed and linked to performed an operation, they cannot be
-reused anymore. FPGA are cadenced more slowly than classic processors but they can
-performed pipelined as well as parallel operations. A pipeline provides a way
-manipulate data quickly since at each clock top to handle a new data. However,
-using a pipeline consomes more logics and components since they are not
-reusable, nevertheless it is probably the most efficient technique on FPGA.
-Parallel operations can be used in order to manipulate several data
-simultaneously. When it is possible, using a pipeline is a good solution to
-manipulate new data at each clock top and using parallelism to handle
-simultaneously several data streams.
-
-%% parler du VHDL, synthèse et bitstream
+configured by the customer. FGPAs are composed of programmable logic components,
+called configurable logic blocks (CLB). These blocks mainly contains look-up
+tables (LUT), flip/flops (F/F) and latches, organized in one or more slices
+connected together. Each CLB can be configured to perform simple (AND, XOR, ...)
+or complex combinational functions. They are interconnected by reconfigurable
+links. Modern FPGAs contain memory elements and multipliers which enable to
+simplify the design and to increase the performance. Nevertheless, all other
+complex operations, like division, trigonometric functions, $\ldots$ are not
+available and must be done by configuring a set of CLBs. Since this
+configuration is not obvious at all, it can be done via a framework, like
+ISE~\cite{ISE}. Such a software can synthetize a design written in a hardware
+description language (HDL), map it onto CLBs, place/route them for a specific
+FPGA, and finally produce a bitstream that is used to configre the FPGA. Thus,
+from the developper point of view, the main difficulty is to translate an
+algorithm in HDL code, taking account FPGA resources and constraints like clock
+signals and I/O values that drive the FPGA.
+
+Indeed, HDL programming is very different from classic languages like
+C. A program can be seen as a state-machine, manipulating signals that
+evolve from state to state. By the way, HDL instructions can execute
+concurrently. Basic logic operations are used to agregate signals to
+produce new states and assign it to another signal. States are mainly
+expressed as arrays of bits. Fortunaltely, libraries propose some
+higher levels representations like signed integers, and arithmetic
+operations.
+
+Furthermore, even if FPGAs are cadenced more slowly than classic
+processors, they can perform pipeline as well as parallel
+operations. A pipeline consists in cutting a process in sequence of
+small tasks, taking the same execution time. It accepts a new data at
+each clock top, thus, after a known latency, it also provides a result
+at each clock top. However, using a pipeline consumes more logics
+since the components of a task are not reusable by another
+one. Nevertheless it is probably the most efficient technique on
+FPGA. Because of its architecture, it is also very easy to process
+several data concurrently. When it is possible, the best performance
+is reached using parallelism to handle simultaneously several
+pipelines in order to handle multiple data streams.
+
\subsection{The board}
The board we use is designed by the Armadeus compagny, under the name
SP Vision. It consists in a development board hosting a i.MX27 ARM
processor (from Freescale). The board includes all classical
-connectors : USB, Ethernet, ... A Flash memory contains a Linux kernel
+connectors: USB, Ethernet, ... A Flash memory contains a Linux kernel
that can be launched after booting the board via u-Boot.
The processor is directly connected to a Spartan3A FPGA (from Xilinx)
tunnel. By default, the WEIM interface provides a clock signal at
100MHz that is connected to dedicated FPGA pins.
-The Spartan6 is an LX100 version. It has 15822 slices, equivalent to
-101261 logic cells. There are 268 internal block RAM of 18Kbits, and
-180 dedicated multiply-adders (named DSP48), which is largely enough
-for our project.
+The Spartan6 is an LX100 version. It has 15822 slices, each slice
+containing 4 LUTs and 8 flip/flops. It is equivalent to 101261 logic
+cells. There are 268 internal block RAM of 18Kbits, and 180 dedicated
+multiply-adders (named DSP48), which is largely enough for our
+project.
Some I/O pins of Spartan6 are connected to two $2\times 17$ headers
that can be used as user wants. For the project, they will be
classical least square method but suppose that frequency is already
known.
-\subsubsection{Spline algorithm}
+\subsubsection{Spline algorithm (SPL)}
\label{sec:algo-spline}
Let consider a profile $P$, that is a segment of $M$ pixels with an
intensity in gray levels. Let call $I(x)$ the intensity of profile in $x
computed. Finding intersections of $I^s$ and this line allow to obtain
the period thus the frequency.
-The phase is computed via the equation :
+The phase is computed via the equation:
\begin{equation}
\theta = atan \left[ \frac{\sum_{i=0}^{N-1} sin(2\pi f x^s_i) \times I^s(x^s_i)}{\sum_{i=0}^{N-1} cos(2\pi f x^s_i) \times I^s(x^s_i)} \right]
\end{equation}
-Two things can be noticed :
+Two things can be noticed:
\begin{itemize}
\item the frequency could also be obtained using the derivates of
spline equations, which only implies to solve quadratic equations.
computation of $\theta$.
\end{itemize}
-\subsubsection{Least square algorithm}
+\subsubsection{Least square algorithm (LSQ)}
Assuming that we compute the phase during the acquisition loop,
-equation \ref{equ:profile} has only 4 parameters :$a, b, A$, and
+equation \ref{equ:profile} has only 4 parameters: $a, b, A$, and
$\theta$, $f$ and $x$ being already known. Since $I$ is non-linear, a
-least square method based an Gauss-newton algorithm must be used to
+least square method based on a Gauss-newton algorithm can be used to
determine these four parameters. Since it is an iterative process
ending with a convergence criterion, it is obvious that it is not
particularly adapted to our design goals.
Fortunatly, it is quite simple to reduce the number of parameters to
only $\theta$. Let $x^p$ be the coordinates of pixels in a segment of
size $M$. Thus, $x^p = 0, 1, \ldots, M-1$. Let $I(x^p)$ be their
-intensity. Firstly, we "remove" the slope by computing :
+intensity. Firstly, we "remove" the slope by computing:
\[I^{corr}(x^p) = I(x^p) - a.x^p - b\]
Since linear equation coefficients are searched, a classical least
-square method can be used to determine $a$ and $b$ :
+square method can be used to determine $a$ and $b$:
\[a = \frac{covar(x^p,I(x^p))}{var(x^p)} \]
-Assuming an overlined symbol means an average, then :
+Assuming an overlined symbol means an average, then:
\[b = \overline{I(x^p)} - a.\overline{{x^p}}\]
\[A = \frac{max(I^{corr}) - min(I^{corr})}{2}\]
-Then, the least square method to find $\theta$ is reduced to search the minimum of :
+Then, the least square method to find $\theta$ is reduced to search the minimum of:
\[\sum_{i=0}^{M-1} \left[ cos(2\pi f.i + \theta) - \frac{I^{corr}(i)}{A} \right]^2\]
-It is equivalent to derivate this expression and to solve the following equation :
+It is equivalent to derivate this expression and to solve the following equation:
\begin{eqnarray*}
2\left[ cos\theta \sum_{i=0}^{M-1} I^{corr}(i).sin(2\pi f.i) + sin\theta \sum_{i=0}^{M-1} I^{corr}(i).cos(2\pi f.i)\right] \\
- A\left[ cos2\theta \sum_{i=0}^{M-1} sin(4\pi f.i) + sin2\theta \sum_{i=0}^{M-1} cos(4\pi f.i)\right] = 0
\end{eqnarray*}
-Several points can be noticed :
+Several points can be noticed:
\begin{itemize}
\item As in the spline method, some parts of this equation can be
computed before the acquisition loop. It is the case of sums that do
- not depend on $\theta$ :
+ not depend on $\theta$:
\[ \sum_{i=0}^{M-1} sin(4\pi f.i), \sum_{i=0}^{M-1} cos(4\pi f.i) \]
\item The simplest method to find the good $\theta$ is to discretize
$[-\pi,\pi]$ in $nb_s$ steps, and to search which step leads to the
result closest to zero. By the way, three other lookup tables can
- also be computed before the loop :
+ also be computed before the loop:
\[ sin \theta, cos \theta, \]
\end{itemize}
-Finally, the whole summarizes in an algorithm (called LSQ in the following) in two parts, one before and one during the acquisition loop :
-\begin{algorithm}[h]
+Finally, the whole summarizes in an algorithm (called LSQ in the following) in two parts, one before and one during the acquisition loop:
+\begin{algorithm}[htbp]
\caption{LSQ algorithm - before acquisition loop.}
\label{alg:lsq-before}
}
\end{algorithm}
-\begin{algorithm}[ht]
+\begin{algorithm}[htbp]
\caption{LSQ algorithm - during acquisition loop.}
\label{alg:lsq-during}
\subsubsection{Comparison}
-We compared the two algorithms on the base of three criterions :
+We compared the two algorithms on the base of three criteria:
\begin{itemize}
\item precision of results on a cosinus profile, distorted with noise,
\item number of operations,
deflection, which is smaller than the best precision they achieved,
i.e. 0.3nm.
-For each test, we add some noise to the profile : each group of two
+For each test, we add some noise to the profile: each group of two
pixels has its intensity added to a random number picked in $[-N,N]$
(NB: it should be noticed that picking a new value for each pixel does
not distort enough the profile). The absolute error on the result is
evaluated by comparing the difference between the reference and
-computed phase, out of $2\pi$, expressed in percents. That is : $err =
+computed phase, out of $2\pi$, expressed in percents. That is: $err =
100\times \frac{|\theta_{ref} - \theta_{comp}|}{2\pi}$.
Table \ref{tab:algo_prec} gives the maximum and average error for the two algorithms and increasing values of $N$.
\begin{figure}[ht]
\begin{center}
- \includegraphics[width=9cm]{intens-noise20-spl}
+ \includegraphics[width=\columnwidth]{intens-noise20}
\end{center}
\caption{Sample of worst profile for N=10}
\label{fig:noise20}
\begin{figure}[ht]
\begin{center}
- \includegraphics[width=9cm]{intens-noise60-lsq}
+ \includegraphics[width=\columnwidth]{intens-noise60}
\end{center}
\caption{Sample of worst profile for N=30}
\label{fig:noise60}
We assume that $M=20$, $nb_s=1024$, $k=4$, all possible parts are
already in lookup tables and a limited set of operations (+, -, *, /,
-<, >) is taken account. Translating the two algorithms in C code, we
+$<$, $>$) is taken account. Translating the two algorithms in C code, we
obtain about 430 operations for LSQ and 1550 (plus few tenth for
$atan$) for SPL. This result is largely in favor of LSQ. Nevertheless,
considering the total number of operations is not really pertinent for
-an FPGA implementation : it mainly depends on the type of operations
+an FPGA implementation: it mainly depends on the type of operations
and their
ordering. The final decision is thus driven by the third criterion.\\
-The Spartan 6 used in our architecture has hard constraint : it has no
-built-in floating point units. Obviously, it is possible to use some
-existing "black-boxes" for double precision operations. But they have
-a quite long latency. It is much simpler to exclusively use integers,
-with a quantization of all double precision values. Obviously, this
-quantization should not decrease too much the precision of
-results. Furthermore, it should not lead to a design with a huge
-latency because of operations that could not complete during a single
-or few clock cycles. Divisions are in this case and, moreover, they
-need an varying number of clock cycles to complete. Even
-multiplications can be a problem : DSP48 take inputs of 18 bits
-maximum. For larger multiplications, several DSP must be combined,
-increasing the latency.
-
-Nevertheless, the hardest constraint does not come from the FPGA
-characteristics but from the algorithms. Their VHDL implentation will
-be efficient only if they can be fully (or near) pipelined. By the
-way, the choice is quickly done : only a small part of SPL can be.
-Indeed, the computation of spline coefficients implies to solve a
-tridiagonal system $A.m = b$. Values in $A$ and $b$ can be computed
-from incoming pixels intensity but after, the back-solve starts with
-the lastest values, which breaks the pipeline. Moreover, SPL relies on
-interpolating far more points than profile size. Thus, the end
-of SPL works on a larger amount of data than the beginning, which
-also breaks the pipeline.
-
-LSQ has not this problem : all parts except the dichotomial search
-work on the same amount of data, i.e. the profile size. Furthermore,
-LSQ needs less operations than SPL, implying a smaller output
-latency. Consequently, it is the best candidate for phase
-computation. Nevertheless, obtaining a fully pipelined version
-supposes that operations of different parts complete in a single clock
-cycle. It is the case for simulations but it completely fails when
-mapping and routing the design on the Spartan6. By the way,
-extra-latency is generated and there must be idle times between two
-profiles entering into the pipeline.
-
-%%Before obtaining the least bitstream, the crucial question is : how to
+The Spartan 6 used in our architecture has a hard constraint: it has no built-in
+floating point units. Obviously, it is possible to use some existing
+"black-boxes" for double precision operations. But they have a quite long
+latency. It is much simpler to exclusively use integers, with a quantization of
+all double precision values. Obviously, this quantization should not decrease
+too much the precision of results. Furthermore, it should not lead to a design
+with a huge latency because of operations that could not complete during a
+single or few clock cycles. Divisions are in this case and, moreover, they need
+a varying number of clock cycles to complete. Even multiplications can be a
+problem: DSP48 take inputs of 18 bits maximum. For larger multiplications,
+several DSP must be combined, increasing the latency.
+
+Nevertheless, the hardest constraint does not come from the FPGA characteristics
+but from the algorithms. Their VHDL implentation will be efficient only if they
+can be fully (or near) pipelined. By the way, the choice is quickly done: only a
+small part of SPL can be. Indeed, the computation of spline coefficients
+implies to solve a tridiagonal system $A.m = b$. Values in $A$ and $b$ can be
+computed from incoming pixels intensity but after, the back-solve starts with
+the lastest values, which breaks the pipeline. Moreover, SPL relies on
+interpolating far more points than profile size. Thus, the end of SPL works on a
+larger amount of data than the beginning, which also breaks the pipeline.
+
+LSQ has not this problem: all parts except the dichotomial search work on the
+same amount of data, i.e. the profile size. Furthermore, LSQ needs less
+operations than SPL, implying a smaller output latency. Consequently, it is the
+best candidate for phase computation. Nevertheless, obtaining a fully pipelined
+version supposes that operations of different parts complete in a single clock
+cycle. It is the case for simulations but it completely fails when mapping and
+routing the design on the Spartan6. By the way, extra-latency is generated and
+there must be idle times between two profiles entering into the pipeline.
+
+%%Before obtaining the least bitstream, the crucial question is: how to
%%translate the C code the LSQ into VHDL ?
\section{Experimental tests}
+In this section we explain what we have done yet. Until now, we could not perform
+real experiments since we just have received the FGPA board. Nevertheless, we
+will include real experiments in the final version of this paper.
+
\subsection{VHDL implementation}
+
+
% - ecriture d'un code en C avec integer
% - calcul de la taille max en bit de chaque variable en fonction de la quantization.
% - tests de quantization : équilibre entre précision et contraintes FPGA
% - en parallèle : simulink et VHDL à la main
-%
+
+
+From the LSQ algorithm, we have written a C program that uses only
+integer values. We use a very simple quantization by multiplying
+double precision values by a power of two, keeping the integer
+part. For example, all values stored in lut$_s$, lut$_c$, $\ldots$ are
+scaled by 1024. Since LSQ also computes average, variance, ... to
+remove the slope, the result of implied euclidian divisions may be
+relatively wrong. To avoid that, we also scale the pixel intensities
+by a power of two. Futhermore, assuming $nb_s$ is fixed, these
+divisions have a knonw denominator. Thus, they can be replaced by
+their multiplication/shift counterpart. Finally, all other
+multiplications or divisions by a power of two have been replaced by
+left or right bit shifts. By the way, the code only contains
+additions, substractions and multiplications of signed integers, which
+is perfectly adapted to FGPAs.
+
+As said above, hardware constraints have a great influence on the VHDL
+implementation. Consequently, we searched the maximum value of each
+variable as a function of the different scale factors and the size of
+profiles, which gives their maximum size in bits. That size determines
+the maximum scale factors that allow to use the least possible RAMs
+and DSPs. Actually, we implemented our algorithm with this maximum
+size but current works study the impact of quantization on the results
+precision and design complexity. We have compared the result of the
+LSQ version using integers and doubles and observed that the precision
+of both were similar.
+
+Then we built two versions of VHDL codes: one directly by hand coding
+and the other with Matlab using the Simulink HDL coder
+feature~\cite{HDLCoder}. Although the approach is completely different
+we obtained VHDL codes that are quite comparable. Each approach has
+advantages and drawbacks. Roughly speaking, hand coding provides
+beautiful and much better structured code while Simulink allows to
+produce a code faster. In terms of throughput and latency,
+simulations shows that the two approaches are close with a slight
+advantage for hand coding. We hope that real experiments will confirm
+that.
+
\subsection{Simulation}
+Currently, we have only simulated our VHDL codes with GHDL and GTKWave (two free
+tools with linux). Both approaches led to correct results. At the beginning of
+our simulations, our pipiline could compute a new phase each 33 cycles and the
+length of the pipeline was equal to 95 cycles. When we tried to generate the
+corresponding bitsream with ISE environment we had many problems because many
+stages required more than the 10$n$s required by the clock frequency. So we
+needed to decompose some part of the pipeline in order to add some cycles and
+simplify some parts between a clock top.
% ghdl + gtkwave
% au mieux : une phase tous les 33 cycles, latence de 95 cycles.
% mais routage/placement impossible.
\subsection{Bitstream creation}
+Currently both approaches provide synthesable bitstreams with ISE. We expect
+that the pipeline will have a latency of 112 cycles, i.e. 1.12$\mu$s and it
+could accept new profiles of pixel each 48 cycles, i.e. 480$n$s.
+
% pas fait mais prévision d'une sortie tous les 480ns avec une latence de 1120
\label{sec:results}
