[sharelatex-git-integration Best effort strategy and virtual load for asynchronous...

[loba-papers.git] / loba-besteffort / loba-besteffort.tex

diff --git a/loba-besteffort/loba-besteffort.tex b/loba-besteffort/loba-besteffort.tex
index 6a766026a07a1a23c59851f4e4baf6aa3c7cf83a..c387ad6e62bee5e66175f1220104360944bb12e4 100644 (file)
--- a/loba-besteffort/loba-besteffort.tex
+++ b/loba-besteffort/loba-besteffort.tex
@@ -27,6 +27,9 @@
 
 \newcommand{\VAR}[1]{\textit{#1}}
 
 
 \newcommand{\VAR}[1]{\textit{#1}}
 

+\newcommand{\besteffort}{\emph{best effort}}
+\newcommand{\makhoul}{\emph{Makhoul}}
+

 \begin{document}
 
 \begin{frontmatter}
 \begin{document}
 
 \begin{frontmatter}


@@ -42,11 +45,13 @@
 \author{Arnaud Giersch\corref{cor}}
 \ead{arnaud.giersch@femto-st.fr}
 
 \author{Arnaud Giersch\corref{cor}}
 \ead{arnaud.giersch@femto-st.fr}
 

-\address{FEMTO-ST, University of Franche-Comté\\
- 19 avenue de Maréchal Juin, BP 527, 90016 Belfort cedex , France\\
-  % Tel.: +123-45-678910\\
-  % Fax: +123-45-678910\\
-}
+\address{%
+  Institut FEMTO-ST (UMR 6174),
+  Université de Franche-Comté (UFC),
+  Centre National de la Recherche Scientifique (CNRS),
+  École Nationale Supérieure de Mécanique et des Microtechniques (ENSMM),
+  Université de Technologie de Belfort Montbéliard (UTBM)\\
+  19 avenue du Maréchal Juin, BP 527, 90016 Belfort cedex, France}

 
 \cortext[cor]{Corresponding author.}
 
 
 \cortext[cor]{Corresponding author.}
 


@@ -57,7 +62,7 @@
   the most well known algorithm for which the convergence proof is given. From a
   practical point of view, when a node wants to balance a part of its load to
   some of its neighbors, the strategy is not described.  In this paper, we
   the most well known algorithm for which the convergence proof is given. From a
   practical point of view, when a node wants to balance a part of its load to
   some of its neighbors, the strategy is not described.  In this paper, we

-  propose a strategy called \emph{best effort} which tries to balance the load
+  propose a strategy called \besteffort{} which tries to balance the load

   of a node to all its less loaded neighbors while ensuring that all the nodes
   concerned by the load balancing phase have the same amount of load.  Moreover,
   asynchronous iterative algorithms in which an asynchronous load balancing
   of a node to all its less loaded neighbors while ensuring that all the nodes
   concerned by the load balancing phase have the same amount of load.  Moreover,
   asynchronous iterative algorithms in which an asynchronous load balancing


@@ -82,7 +87,8 @@
 Load  balancing algorithms  are  extensively used  in  parallel and  distributed
 applications in  order to  reduce the  execution times. They  can be  applied in
 different scientific  fields from high  performance computation to  micro sensor
 Load  balancing algorithms  are  extensively used  in  parallel and  distributed
 applications in  order to  reduce the  execution times. They  can be  applied in
 different scientific  fields from high  performance computation to  micro sensor

-networks.   They are  iterative by  nature.  In  literature many  kinds  of load
+networks.   They are  iterative by  nature.\FIXME{really?}
+In  literature many  kinds  of load

 balancing  algorithms  have been  studied.   They  can  be classified  according
 different  criteria:   centralized  or  decentralized,  in   static  or  dynamic
 environment,  with  homogeneous  or  heterogeneous load,  using  synchronous  or
 balancing  algorithms  have been  studied.   They  can  be classified  according
 different  criteria:   centralized  or  decentralized,  in   static  or  dynamic
 environment,  with  homogeneous  or  heterogeneous load,  using  synchronous  or


@@ -104,17 +110,17 @@ Although  the Bertsekas  and Tsitsiklis'  algorithm describes  the  condition to
 ensure the convergence,  there is no indication or  strategy to really implement
 the load distribution. In other word, a node  can send a part of its load to one
 or   many  of   its  neighbors   while  all   the  convergence   conditions  are
 ensure the convergence,  there is no indication or  strategy to really implement
 the load distribution. In other word, a node  can send a part of its load to one
 or   many  of   its  neighbors   while  all   the  convergence   conditions  are

-followed. Consequently,  we propose a  new strategy called  \emph{best effort}
+followed. Consequently,  we propose a  new strategy called  \besteffort{}

 that tries to balance the load of  a node to all its less loaded neighbors while
 ensuring that all the nodes concerned  by the load balancing phase have the same
 amount of  load.  Moreover, when real asynchronous  applications are considered,
 using  asynchronous   load  balancing   algorithms  can  reduce   the  execution
 times. Most of the times, it is simpler to distinguish load information messages
 that tries to balance the load of  a node to all its less loaded neighbors while
 ensuring that all the nodes concerned  by the load balancing phase have the same
 amount of  load.  Moreover, when real asynchronous  applications are considered,
 using  asynchronous   load  balancing   algorithms  can  reduce   the  execution
 times. Most of the times, it is simpler to distinguish load information messages

-from  data  migration  messages.  Former  ones  allows  a  node to  inform  its
+from  data  migration  messages.  Former  ones  allow  a  node to  inform  its

 neighbors of its  current load. These messages are very small,  they can be sent
 neighbors of its  current load. These messages are very small,  they can be sent

-quite often.  For example, if an  computing iteration takes  a significant times
+quite often.  For example, if a computing iteration takes  a significant times

 (ranging from seconds to minutes), it is possible to send a new load information
 (ranging from seconds to minutes), it is possible to send a new load information

-message at each  neighbor at each iteration. Latter  messages contains data that
+message to each  neighbor at each iteration. Latter  messages contain data that

 migrates from one node to another one. Depending on the application, it may have
 sense or not  that nodes try to balance  a part of their load  at each computing
 iteration. But the time to transfer a load message from a node to another one is
 migrates from one node to another one. Depending on the application, it may have
 sense or not  that nodes try to balance  a part of their load  at each computing
 iteration. But the time to transfer a load message from a node to another one is


@@ -124,37 +130,36 @@ it can take this information into account  and it can consider that its new load
 is larger.   Consequently, it can  send a part  of it real  load to some  of its
 neighbors if required. We call this trick the \emph{virtual load} mechanism.
 
 is larger.   Consequently, it can  send a part  of it real  load to some  of its
 neighbors if required. We call this trick the \emph{virtual load} mechanism.
 

-
-
-So, in  this work, we propose a  new strategy for improving  the distribution of
-the  load  and  a  simple  but  efficient trick  that  also  improves  the  load
-balancing. Moreover, we have conducted  many simulations with SimGrid in order to
-validate our improvements are really efficient. Our simulations consider that in
-order  to send a  message, a  latency delays  the sending  and according  to the
-network  performance and  the message  size, the  time of  the reception  of the
+So, in this work, we propose a new strategy to improve the distribution of the
+load and a simple but efficient trick that also improves the load
+balancing. Moreover, we have conducted many simulations with SimGrid in order to
+validate that our improvements are really efficient. Our simulations consider
+that in order to send a message, a latency delays the sending and according to
+the network performance and the message size, the time of the reception of the

 message also varies.
 
 message also varies.
 

-In the following of this paper, Section~\ref{BT algo} describes the Bertsekas
-and Tsitsiklis' asynchronous load balancing algorithm. Moreover, we present a
-possible problem in the convergence conditions.  Section~\ref{Best-effort}
-presents the best effort strategy which provides an efficient way to reduce the
-execution times.  This strategy will be compared with other ones, presented in
-Section~\ref{Other}.  In Section~\ref{Virtual load}, the virtual load mechanism
-is proposed.  Simulations allowed to show that both our approaches are valid
-using a quite realistic model detailed in Section~\ref{Simulations}.  Finally we
-give a conclusion and some perspectives to this work.
+In the following of this paper, Section~\ref{sec.bt-algo} describes the
+Bertsekas and Tsitsiklis' asynchronous load balancing algorithm. Moreover, we
+present a possible problem in the convergence conditions.
+Section~\ref{sec.besteffort} presents the best effort strategy which provides an
+efficient way to reduce the execution times.  This strategy will be compared
+with other ones, presented in Section~\ref{sec.other}.  In
+Section~\ref{sec.virtual-load}, the virtual load mechanism is proposed.
+Simulations allowed to show that both our approaches are valid using a quite
+realistic model detailed in Section~\ref{sec.simulations}.  Finally we give a
+conclusion and some perspectives to this work.

 
 
 
 \section{Bertsekas  and Tsitsiklis' asynchronous load balancing algorithm}
 
 
 
 \section{Bertsekas  and Tsitsiklis' asynchronous load balancing algorithm}

-\label{BT algo}
+\label{sec.bt-algo}

 
 In  order  prove  the  convergence  of  asynchronous  iterative  load  balancing
 Bertsekas         and        Tsitsiklis         proposed         a        model
 in~\cite{bertsekas+tsitsiklis.1997.parallel}.   Here we  recall  some notations.
 Consider  that  $N={1,...,n}$  processors   are  connected  through  a  network.
 
 In  order  prove  the  convergence  of  asynchronous  iterative  load  balancing
 Bertsekas         and        Tsitsiklis         proposed         a        model
 in~\cite{bertsekas+tsitsiklis.1997.parallel}.   Here we  recall  some notations.
 Consider  that  $N={1,...,n}$  processors   are  connected  through  a  network.

-Communication links  are represented by  a connected undirected  graph $G=(N,V)$
-where $V$ is the set of links connecting different processors. In this work, we
+Communication links  are represented by  a connected undirected  graph $G=(N,A)$
+where $A$ is the set of links connecting different processors. In this work, we

 consider that  processors are  homogeneous for sake  of simplicity. It  is quite
 easy to tackle the  heterogeneous case~\cite{ElsMonPre02}. Load of processor $i$
 at  time $t$  is  represented  by $x_i(t)\geq  0$.   Let $V(i)$  be  the set  of
 consider that  processors are  homogeneous for sake  of simplicity. It  is quite
 easy to tackle the  heterogeneous case~\cite{ElsMonPre02}. Load of processor $i$
 at  time $t$  is  represented  by $x_i(t)\geq  0$.   Let $V(i)$  be  the set  of


@@ -170,7 +175,7 @@ amount of  load received by processor $j$  from processor $i$ at  time $t$. Then
 the amount of load of processor $i$ at time $t+1$ is given by:
 \begin{equation}
 x_i(t+1)=x_i(t)-\sum_{j\in V(i)} s_{ij}(t) + \sum_{j\in V(i)} r_{ji}(t)
 the amount of load of processor $i$ at time $t+1$ is given by:
 \begin{equation}
 x_i(t+1)=x_i(t)-\sum_{j\in V(i)} s_{ij}(t) + \sum_{j\in V(i)} r_{ji}(t)

-\label{eq:ping-pong}
+\label{eq.ping-pong}

 \end{equation}
 
 
 \end{equation}
 
 


@@ -187,15 +192,15 @@ Nevertheless,  we  think that  this  condition may  lead  to  deadlocks in  some
 cases. For example, if we consider  only three processors and that processor $1$
 is linked to processor $2$ which is  also linked to processor $3$ (i.e. a simple
 chain which 3 processors). Now consider we have the following values at time $t$:
 cases. For example, if we consider  only three processors and that processor $1$
 is linked to processor $2$ which is  also linked to processor $3$ (i.e. a simple
 chain which 3 processors). Now consider we have the following values at time $t$:

-\begin{eqnarray*}
-x_1(t)=10   \\
-x_2(t)=100   \\
-x_3(t)=99.99\\
- x_3^2(t)=99.99\\
-\end{eqnarray*}
-In this case, processor $2$ can  either sends load to processor $1$ or processor
-$3$.   If  it  sends  load  to  processor $1$  it  will  not  satisfy  condition
-(\ref{eq:ping-pong})  because  after the  sending  it  will  be less  loaded  that
+\begin{align*}
+  x_1(t)   &= 10    \\
+  x_2(t)   &= 100   \\
+  x_3(t)   &= 99.99 \\
+  x_3^2(t) &= 99.99 \\
+\end{align*}
+In this case, processor $2$ can either sends load to processor $1$ or processor
+$3$.  If it sends load to processor $1$ it will not satisfy condition
+\eqref{eq.ping-pong} because after the sending it will be less loaded that

 $x_3^2(t)$.  So we consider that the \emph{ping-pong} condition is probably to
 strong. Currently, we did not try to make another convergence proof without this
 condition or with a weaker condition.
 $x_3^2(t)$.  So we consider that the \emph{ping-pong} condition is probably to
 strong. Currently, we did not try to make another convergence proof without this
 condition or with a weaker condition.


@@ -209,15 +214,15 @@ that they are sufficient to ensure the convergence of the load-balancing
 algorithm.
 
 \section{Best effort strategy}
 algorithm.
 
 \section{Best effort strategy}

-\label{Best-effort}
+\label{sec.besteffort}

 
 In this section we describe a new load-balancing strategy that we call
 
 In this section we describe a new load-balancing strategy that we call

-\emph{best effort}.  First, we explain the general idea behind this strategy,
+\besteffort{}.  First, we explain the general idea behind this strategy,

 and then we describe some variants of this basic strategy.
 
 \subsection{Basic strategy}
 
 and then we describe some variants of this basic strategy.
 
 \subsection{Basic strategy}
 

-The general idea behind the \emph{best effort} strategy is that each processor,
+The general idea behind the \besteffort{} strategy is that each processor,

 that detects it has more load than some of its neighbors, sends some load to the
 most of its less loaded neighbors, doing its best to reach the equilibrium
 between those neighbors and himself.
 that detects it has more load than some of its neighbors, sends some load to the
 most of its less loaded neighbors, doing its best to reach the equilibrium
 between those neighbors and himself.


@@ -281,17 +286,17 @@ potentially wrong decision has a lower impact.
 
 Concretely, once $s_{ij}$ has been evaluated as before, it is simply divided by
 some configurable factor.  That's what we named the ``parameter $k$'' in
 
 Concretely, once $s_{ij}$ has been evaluated as before, it is simply divided by
 some configurable factor.  That's what we named the ``parameter $k$'' in

-Section~\ref{Results}.  The amount of data to send is then $s_{ij}(t) = (\bar{x}
-- x^i_j(t))/k$.
-\FIXME[check that it's still named $k$ in Sec.~\ref{Results}]{}
+Section~\ref{sec.results}.  The amount of data to send is then $s_{ij}(t) =
+(\bar{x} - x^i_j(t))/k$.
+\FIXME[check that it's still named $k$ in Sec.~\ref{sec.results}]{}

 
 \section{Other strategies}
 
 \section{Other strategies}

-\label{Other}
+\label{sec.other}

 
 Another load balancing strategy, working under the same conditions, was
 previously developed by Bahi, Giersch, and Makhoul in
 \cite{bahi+giersch+makhoul.2008.scalable}.  In order to assess the performances
 
 Another load balancing strategy, working under the same conditions, was
 previously developed by Bahi, Giersch, and Makhoul in
 \cite{bahi+giersch+makhoul.2008.scalable}.  In order to assess the performances

-of the new \emph{best effort}, we naturally chose to compare it to this anterior
+of the new \besteffort{}, we naturally chose to compare it to this anterior

 work.  More precisely, we will use the algorithm~2 from
 \cite{bahi+giersch+makhoul.2008.scalable} and, in the following, we will
 reference it under the name of Makhoul's.
 work.  More precisely, we will use the algorithm~2 from
 \cite{bahi+giersch+makhoul.2008.scalable} and, in the following, we will
 reference it under the name of Makhoul's.


@@ -301,19 +306,19 @@ a load balancing decision, it starts by sorting its neighbors by increasing
 order of their load.  Then, it computes the difference between its own load, and
 the load of each of its neighbors.  Finally, taking the neighbors following the
 order defined before, the amount of load to send $s_{ij}$ is computed as
 order of their load.  Then, it computes the difference between its own load, and
 the load of each of its neighbors.  Finally, taking the neighbors following the
 order defined before, the amount of load to send $s_{ij}$ is computed as

-$1/(N+1)$ of the load difference, with $N$ being the number of neighbors.  This
+$1/(n+1)$ of the load difference, with $n$ being the number of neighbors.  This

 process continues as long as the node is more loaded than the considered
 neighbor.
 
 
 \section{Virtual load}
 process continues as long as the node is more loaded than the considered
 neighbor.
 
 
 \section{Virtual load}

-\label{Virtual load}
+\label{sec.virtual-load}

 
 In this section,  we present the concept of \emph{virtual load}.  In order to
 use this concept, load balancing messages must be sent using two different kinds
 of  messages:  load information  messages  and  load  balancing messages.   More
 
 In this section,  we present the concept of \emph{virtual load}.  In order to
 use this concept, load balancing messages must be sent using two different kinds
 of  messages:  load information  messages  and  load  balancing messages.   More

-precisely, a node  wanting to send a part  of its load to one  of its neighbors,
-can first send  a load information message containing the load  it will send and
+precisely, a node wanting to send a part of its load to one of its neighbors
+can first send a load information message containing the load it will send, and

 then it can send the load  balancing message containing data  to be transferred.
 Load information  message are really  short, consequently they will  be received
 very quickly.  In opposition, load  balancing messages are often bigger and thus
 then it can send the load  balancing message containing data  to be transferred.
 Load information  message are really  short, consequently they will  be received
 very quickly.  In opposition, load  balancing messages are often bigger and thus


@@ -330,31 +335,30 @@ load  to  some of  its  neighbors  without waiting  the  reception  of the  load
 balancing message.
 
 Doing  this, we  can  expect a  faster  convergence since  nodes  have a  faster
 balancing message.
 
 Doing  this, we  can  expect a  faster  convergence since  nodes  have a  faster

-information of the load they will receive, so they can take in into account.
+information of the load they will receive, so they can take it into account.

 
 \FIXME{Est ce qu'on donne l'algo avec virtual load?}
 
 \FIXME{describe integer mode}
 
 \section{Simulations}
 
 \FIXME{Est ce qu'on donne l'algo avec virtual load?}
 
 \FIXME{describe integer mode}
 
 \section{Simulations}

-\label{Simulations}
+\label{sec.simulations}

 
 In order to test and validate our approaches, we wrote a simulator
 using the SimGrid
 
 In order to test and validate our approaches, we wrote a simulator
 using the SimGrid

-framework~\cite{casanova+legrand+quinson.2008.simgrid}.  This
+framework~\cite{simgrid.web,casanova+giersch+legrand+al.2014.simgrid}.  This

 simulator, which consists of about 2,700 lines of C++, allows to run
 the different load-balancing strategies under various parameters, such
 as the initial distribution of load, the interconnection topology, the
 characteristics of the running platform, etc.  Then several metrics
 are issued that permit to compare the strategies.
 
 simulator, which consists of about 2,700 lines of C++, allows to run
 the different load-balancing strategies under various parameters, such
 as the initial distribution of load, the interconnection topology, the
 characteristics of the running platform, etc.  Then several metrics
 are issued that permit to compare the strategies.
 

-The simulation model is detailed in the next section (\ref{Sim
-  model}), and the experimental contexts are described in
-section~\ref{Contexts}.  Then the results of the simulations are
-presented in section~\ref{Results}.
+The simulation model is detailed in the next section (\ref{sec.model}), and the
+experimental contexts are described in section~\ref{sec.exp-context}.  Then the
+results of the simulations are presented in section~\ref{sec.results}.

 
 \subsection{Simulation model}
 
 \subsection{Simulation model}

-\label{Sim model}
+\label{sec.model}

 
 In the simulation model the processors exchange messages which are of
 two kinds.  First, there are \emph{control messages} which only carry
 
 In the simulation model the processors exchange messages which are of
 two kinds.  First, there are \emph{control messages} which only carry


@@ -378,7 +382,7 @@ For the sake of simplicity, a few details were voluntary omitted from
 these descriptions.  For an exhaustive presentation, we refer to the
 actual source code that was used for the experiments%
 \footnote{As mentioned before, our simulator relies on the SimGrid
 these descriptions.  For an exhaustive presentation, we refer to the
 actual source code that was used for the experiments%
 \footnote{As mentioned before, our simulator relies on the SimGrid

-  framework~\cite{casanova+legrand+quinson.2008.simgrid}.  For the
+  framework~\cite{casanova+giersch+legrand+al.2014.simgrid}.  For the

   experiments, we used a pre-release of SimGrid 3.7 (Git commit
   67d62fca5bdee96f590c942b50021cdde5ce0c07, available from
   \url{https://gforge.inria.fr/scm/?group_id=12})}, and which is
   experiments, we used a pre-release of SimGrid 3.7 (Git commit
   67d62fca5bdee96f590c942b50021cdde5ce0c07, available from
   \url{https://gforge.inria.fr/scm/?group_id=12})}, and which is


@@ -491,10 +495,11 @@ iteratively runs the following operations:
 \end{algorithm}
 
 \paragraph{}\FIXME{ajouter des détails sur la gestion de la charge virtuelle ?
 \end{algorithm}
 
 \paragraph{}\FIXME{ajouter des détails sur la gestion de la charge virtuelle ?

-par ex, donner l'idée générale de l'implémentation.  l'idée générale est déja décrite en section~\ref{Virtual load}}
+  par ex, donner l'idée générale de l'implémentation.  l'idée générale est déja
+  décrite en section~\ref{sec.virtual-load}}

 
 \subsection{Experimental contexts}
 
 \subsection{Experimental contexts}

-\label{Contexts}
+\label{sec.exp-context}

 
 In order to assess the performances of our algorithms, we ran our
 simulator with various parameters, and extracted several metrics, that
 
 In order to assess the performances of our algorithms, we ran our
 simulator with various parameters, and extracted several metrics, that


@@ -503,7 +508,7 @@ we will describe in this section.
 \subsubsection{Load balancing strategies}
 
 Several load balancing strategies were compared.  We ran the experiments with
 \subsubsection{Load balancing strategies}
 
 Several load balancing strategies were compared.  We ran the experiments with

-the \emph{Best effort}, and with the \emph{Makhoul} strategies.  \emph{Best
+the \besteffort{}, and with the \makhoul{} strategies.  \emph{Best

   effort} was tested with parameter $k = 1$, $k = 2$, and $k = 4$.  Secondly,
 each strategy was run in its two variants: with, and without the management of
 \emph{virtual load}.  Finally, we tested each configuration with \emph{real},
   effort} was tested with parameter $k = 1$, $k = 2$, and $k = 4$.  Secondly,
 each strategy was run in its two variants: with, and without the management of
 \emph{virtual load}.  Finally, we tested each configuration with \emph{real},


@@ -511,7 +516,7 @@ and with \emph{integer} load.
 
 To summarize the different load balancing strategies, we have:
 \begin{description}
 
 To summarize the different load balancing strategies, we have:
 \begin{description}

-\item[\textbf{strategies:}] \emph{Makhoul}, or \emph{Best effort} with $k\in
+\item[\textbf{strategies:}] \makhoul{}, or \besteffort{} with $k\in

   \{1,2,4\}$
 \item[\textbf{variants:}] with, or without virtual load
 \item[\textbf{domain:}] real load, or integer load
   \{1,2,4\}$
 \item[\textbf{variants:}] with, or without virtual load
 \item[\textbf{domain:}] real load, or integer load


@@ -555,7 +560,7 @@ algorithms currently do not handle heterogeneous computing resources, the
 processor speeds were normalized, and we arbitrarily chose to fix them to
 1~GFlop/s.
 
 processor speeds were normalized, and we arbitrarily chose to fix them to
 1~GFlop/s.
 

-Then we derived each sort of platform with four different number of computing
+Then we derived each kind of platform with four different numbers of computing

 nodes: 16, 64, 256, and 1024 nodes.
 
 \subsubsection{Configurations}
 nodes: 16, 64, 256, and 1024 nodes.
 
 \subsubsection{Configurations}


@@ -602,6 +607,7 @@ Anyway, all these the experiments represent more than 240 hours of computing
 time.
 
 \subsubsection{Metrics}
 time.
 
 \subsubsection{Metrics}

+\label{sec.metrics}

 
 In order to evaluate and compare the different load balancing strategies we had
 to define several metrics.  Our goal, when choosing these metrics, was to have
 
 In order to evaluate and compare the different load balancing strategies we had
 to define several metrics.  Our goal, when choosing these metrics, was to have


@@ -643,10 +649,10 @@ With these constraints in mind, we defined the following metrics:
 
 
 \subsection{Experimental results}
 
 
 \subsection{Experimental results}

-\label{Results}
+\label{sec.results}

 
 In this section, the results for the different simulations will be presented,
 
 In this section, the results for the different simulations will be presented,

-and we'll try to explain our observations.
+and we will try to explain our observations.

 
 \subsubsection{Cluster vs grid platforms}
 
 
 \subsubsection{Cluster vs grid platforms}
 


@@ -660,7 +666,7 @@ Nevertheless their relative performances remain generally identical.
 This suggests that the relative performances of the different strategies are not
 influenced by the characteristics of the physical platform.  The differences in
 the convergence times can be explained by the fact that on the grid platforms,
 This suggests that the relative performances of the different strategies are not
 influenced by the characteristics of the physical platform.  The differences in
 the convergence times can be explained by the fact that on the grid platforms,

-distant sites are interconnected by links of smaller bandwith.
+distant sites are interconnected by links of smaller bandwidth.

 
 Therefore, in the following, we'll only discuss the results for the grid
 platforms.
 
 Therefore, in the following, we'll only discuss the results for the grid
 platforms.


@@ -699,7 +705,7 @@ initially on an only node, while the results on figure~\ref{fig.resultsN} are
 when the load to balance is initially randomly distributed over all nodes.
 
 On both figures, the computation/communication cost ratio is $10/1$ on the left
 when the load to balance is initially randomly distributed over all nodes.
 
 On both figures, the computation/communication cost ratio is $10/1$ on the left

-column, and $1/10$ on the right column.  With a computatio/communication cost
+column, and $1/10$ on the right column.  With a computation/communication cost

 ratio of $1/1$ the results are just between these two extrema, and definitely
 don't give additional information, so we chose not to show them here.
 
 ratio of $1/1$ the results are just between these two extrema, and definitely
 don't give additional information, so we chose not to show them here.
 


@@ -707,62 +713,131 @@ On each of the figures~\ref{fig.results1} and~\ref{fig.resultsN}, the results
 are given for the process topology being, from top to bottom, a line, a torus or
 an hypercube.
 
 are given for the process topology being, from top to bottom, a line, a torus or
 an hypercube.
 

-\FIXME{explain how to read the graphs}
+Finally, on the graphs, the vertical bars show the measured times for each of
+the algorithms.  These measured times are, from bottom to top, the average idle
+time, the average convergence date, and the maximum convergence date (see
+Section~\ref{sec.metrics}).  The measurements are repeated for the different
+platform sizes.  Some bars are missing, specially for large platforms.  This is
+either because the algorithm did not reach the convergence state in the
+allocated time, or because we simply decided not to run it.

 
 

-each bar -> times for an algorithm
-recall the different times
-no bar -> not run or did not converge in allocated time
+\FIXME{annoncer le plan de la suite}

 
 

-repeated for the different platform sizes.
+\subsubsection{The \besteffort{} and  \makhoul{} strategies without virtual load}

 
 

-\FIXME{donner les premières conclusions, annoncer le plan de la suite}
+Before looking  at the different variations,  we will first show  that the plain
+\besteffort{}  strategy  is valuable,  and  may be  as  good  as the  \makhoul{}
+strategy.  On  Figures~\ref{fig.results1} and~\ref{fig.resultsN},
+these strategies are respectively labeled ``b'' and ``a''.

 
 

-\subsubsection{With the virtual load extension}
+We  can  see  that  the  relative  performance of  these  strategies  is  mainly
+influenced by  the application topology.  It  is for the line  topology that the
+difference is the  more important.  In this case,  the \besteffort{} strategy is
+nearly faster than the \makhoul{} strategy.  This can  be explained by the
+fact that the \besteffort{} strategy tries to distribute the load fairly between
+all the nodes  and with the line topology,  it is easy to load  balance the load
+fairly.

 
 

-\subsubsection{The $k$ parameter}
+On the contrary, for the hypercube topology, the \besteffort{} strategy performs
+worse than the \makhoul{} strategy. In this case, the \makhoul{} strategy which
+tries to give more load to few neighbors reaches the equilibrium faster.

 
 

-\subsubsection{With an initial random repartition,  and larger platforms}
+For the torus  topology, for which the  number of links is between  the line and
+the hypercube, the \makhoul{} strategy  is slightly better but the difference is
+more nuanced when the initial load is  only on one node. The only case where the
+\makhoul{} strategy is really faster than the \besteffort{} strategy is with the
+random initial distribution when the communication are slow.

 
 

-\subsubsection{With integer load}
+Globally   the  number  of   interconnection  is   very  important.    The  more
+the interconnection links are, the  faster the \makhoul{} strategy is because
+it distributes quickly significant amount of load, even if this is unfair, between
+all the  neighbors.  In opposition,  the \besteffort{} strategy  distributes the
+load fairly so this strategy is better for low connected strategy.

 
 

-\FIXME{what about the amount of data?}
-
-\begin{itshape}
-\FIXME{remove that part}
-Dans cet ordre:
-...
-- comparer be/makhoul -> be tient la route
-        -> en réel uniquement
-- valider l'extension virtual load -> c'est 'achement bien
-- proposer le -k -> ça peut aider dans certains cas
-- conclure avec la version entière -> on n'a pas l'effet d'escalier !
-Q: comment inclure les types/tailles de platesformes ?
-Q: comment faire des moyennes ?
-Q: comment introduire les distrib 1/N ?
-...
-
-On constate quoi (vérifier avec les chiffres)?
-\begin{itemize}
-\item cluster ou grid, entier ou réel, ne font pas de grosses différences

 
 

-\item bookkeeping? améliore souvent les choses, parfois au prix d'un retard au démarrage
+\subsubsection{Virtual load}

 
 

-\item makhoul? se fait battre sur les grosses plateformes
+The influence of virtual load is most of the time really significant compared to
+the  same configuration  without  it. Sometimes  it  has no  effect  but {\bf  A
+  VERIFIER} it has never a negative effect on the load balancing we tested.

 
 

-\item taille de plateforme?
+On Figure~\ref{fig.results1}, when the load is  initially on one node, it can be
+noticed that the  average idle times are generally longer  with the virtual load
+than without  it. This  can be explained  by the  fact that, with  virtual load,
+processors  will exchange all  the load  they need  to exchange  as soon  as the
+virtual load has been balanced  between all the processors. So consequently they
+cannot  compute  at  the  beginning.  This is  especially  noticeable  when  the
+communication are slow (on the left part of Figure ~\ref{fig.results1}.

 
 

-\item ratio comp/comm?
+%Dans ce cas  légère amélioration de la cvg. max.  Temps  moyen de cvg. amélioré,
+%mais plus de temps passé en idle, surtout quand les comms coutent cher.

 
 

-\item option $k$? peut-être intéressant sur des plateformes fortement interconnectées (hypercube)
+%\subsubsection{The \besteffort{} strategy with an initial random load
+%  distribution, and larger platforms}

 
 

-\item volume de comm? souvent, besteffort/plain en fait plus. pourquoi?
+%In 
+%Mêmes conclusions pour line et hcube.
+%Sur tore, BE se fait exploser quand les comms coutent cher.

 
 

-\item répartition initiale de la charge ?
+%\FIXME{virer les 1024 ?}

 
 

-\item integer mode sur topo. line n'a jamais fini en plain? vérifier si ce n'est
-  pas à cause de l'effet d'escalier que bk est capable de gommer.
+%\subsubsection{With the virtual load extension with an initial random load
+%  distribution}

 
 

-\end{itemize}
+%Soit c'est équivalent, soit on gagne -> surtout quand les comms coutent cher et
+%qu'il y a beaucoup de voisins.
+
+\subsubsection{The $k$ parameter}
+\label{results-k}
+
+As  explained  previously when  the  communication  are  slow the  \besteffort{}
+strategy is efficient. This is due to the fact that it tries to balance the load
+fairly and consequently  a significant amount of the  load is transfered between
+processors.  In this situation, it is possible to reduce the convergence time by
+using  the leveler  parameter  (parameter  $k$).  The  advantage  of using  this
+solution is particularly efficient when the initial load is randomly distributed
+on  the nodes with  torus and  hypercube topology  and slow  communication. When
+virtual load  mechanism is used,  the effect of  this parameter is  also visible
+with the same condition.
+
+
+
+\subsubsection{With integer load}
+
+We also performed  some experiments with integer load instead  of load with real
+value.  In  this case, the  results have globally  the same behavior.   The most
+intereting  result, from  our point  of view,  is that  the virtual  mode allows
+processors in a line topology to converge to the uniform load balancing. Without
+the virtual  load, most  of the time,  processors converge  to what we  call the
+``stairway effect'', that  is to say that  there is only a difference  of one in
+the load of each processor and its neighbors (for example with 10 processors, we
+obtain 10 9 8 7 6 6 7 8 9 10 instead of 8 8 8 8 8 8 8 8 8 8).
+
+%Cas normal, ligne -> converge pas (effet d'escalier).
+%Avec vload, ça converge.
+
+%Dans les autres cas, résultats similaires au cas réel: redire que vload est
+%intéressant.
+
+\FIXME{ajouter une courbe avec l'équilibrage en entier}
+
+\FIXME{virer la metrique volume de comms}
+
+\FIXME{ajouter une courbe ou on voit l'évolution de la charge en fonction du
+  temps : avec et sans vload}
+
+% \begin{itemize}
+% \item cluster ou grid, entier ou réel, ne font pas de grosses différences
+% \item bookkeeping? améliore souvent les choses, parfois au prix d'un retard au démarrage
+% \item makhoul? se fait battre sur les grosses plateformes
+% \item taille de plateforme?
+% \item ratio comp/comm?
+% \item option $k$? peut-être intéressant sur des plateformes fortement interconnectées (hypercube)
+% \item volume de comm? souvent, besteffort/plain en fait plus. pourquoi?
+% \item répartition initiale de la charge ?
+% \item integer mode sur topo. line n'a jamais fini en plain? vérifier si ce n'est
+%   pas à cause de l'effet d'escalier que bk est capable de gommer.
+% \end{itemize}}

 
 % On veut montrer quoi ? :
 
 
 % On veut montrer quoi ? :
 


@@ -789,13 +864,12 @@ On constate quoi (vérifier avec les chiffres)?
 % Prendre un réseau hétérogène et rendre processeur homogène
 
 % Taille : 10 100 très gros
 % Prendre un réseau hétérogène et rendre processeur homogène
 
 % Taille : 10 100 très gros

-\end{itshape}

 
 \section{Conclusion and perspectives}
 
 \FIXME{conclude!}
 
 
 \section{Conclusion and perspectives}
 
 \FIXME{conclude!}
 

-\section*{Acknowledgements}
+\section*{Acknowledgments}

 
 Computations have been performed on the supercomputer facilities of the
 Mésocentre de calcul de Franche-Comté.
 
 Computations have been performed on the supercomputer facilities of the
 Mésocentre de calcul de Franche-Comté.


@@ -813,7 +887,10 @@ Mésocentre de calcul de Franche-Comté.
 %%% ispell-local-dictionary: "american"
 %%% End:
 
 %%% ispell-local-dictionary: "american"
 %%% End:
 

-% LocalWords:  Raphaël Couturier Arnaud Giersch Abderrahmane Sider Franche ij
-% LocalWords:  Bertsekas Tsitsiklis SimGrid DASUD Comté Béjaïa asynchronism ji
-% LocalWords:  ik isend irecv Cortés et al chan ctrl fifo Makhoul GFlop xml pre
-% LocalWords:  FEMTO Makhoul's fca bdee cdde Contassot Vivier underlaid
+% LocalWords:  Raphaël Couturier Arnaud Giersch Franche ij Bertsekas Tsitsiklis
+% LocalWords:  SimGrid DASUD Comté asynchronism ji ik isend irecv Cortés et al
+% LocalWords:  chan ctrl fifo Makhoul GFlop xml pre FEMTO Makhoul's fca bdee
+% LocalWords:  cdde Contassot Vivier underlaid du de Maréchal Juin cedex calcul
+% LocalWords:  biblio Institut UMR Université UFC Centre Scientifique CNRS des
+% LocalWords:  École Nationale Supérieure Mécanique Microtechniques ENSMM UTBM
+% LocalWords:  Technologie Bahi