X-Git-Url: https://bilbo.iut-bm.univ-fcomte.fr/and/gitweb/mpi-energy2.git/blobdiff_plain/e7807134f0d0528f0dd15e6a810316a0f631397b..137e3f73fedaab51eecf4fea642c4c32e5f68dc7:/mpi-energy2-extension/Heter_paper.tex?ds=inline

diff --git a/mpi-energy2-extension/Heter_paper.tex b/mpi-energy2-extension/Heter_paper.tex
index 56b5edd..6f52ac7 100644
--- a/mpi-energy2-extension/Heter_paper.tex
+++ b/mpi-energy2-extension/Heter_paper.tex
@@ -1094,6 +1094,74 @@ in these benchmarks.
 \subsection{The comparison of the proposed frequencies selecting algorithm }
 \label{sec.compare_EDP}
 
+The tradeoff between the energy consumption and the performance of the parallel 
+application had significant importance in the domain of the research. 
+Many researchers, \cite{EDP_for_multi_processors,Energy_aware_application_scheduling,Exploring_Energy_Performance_TradeOffs},
+are optimized the tradeoff between the energy and performance using the energy and delay product, $EDP=energy \times delay$. 
+This model is used by Spiliopoulos et al. algorithm \cite{Spiliopoulos_Green.governors.Adaptive.DVFS},
+the  objective is to selects the suitable frequencies that minimized EDP product for the multicores 
+architecture when DVFS is used. Moreover, their algorithm is applied online which synchronously optimized the energy consumption 
+and the execution time. Both energy consumption and execution time  of a processor are predicted by the their algorithm.
+In this section the proposed frequency selection algorithm, called Maxdist is compared with Spiliopoulos et al. algorithm, called EDP.
+To make both of the algorithms follow the same direction and  fairly  comparing them, the same energy model,  equation \ref{eq:energy} and
+the execution time model, equation \ref{eq:perf}, are used in the prediction process to select the best vector of the frequencies. 
+In contrast, the proposed algorithm starts the search space from the lower bound computed as in equation the  \ref{eq:Fint}. Also, the algorithm
+stops  the search process when  reaching to the lower bound as mentioned before. While, the EDP algorithm is developed to start from the 
+same upper bound until it reach to the minimum available frequencies. Finally, resulting the algorithm is an exhaustive search algorithm that
+test all possible frequencies, starting from the initial frequencies, and selecting those minimized the EDP products.
+
+Both algorithms were applied to NAS benchmarks class D over 16 nodes selected from grid'5000 clusters.
+The participating computing nodes are distributed between two sites to had two different scenarios. 
+These scenarios are two sites and one site scenarios that explained previously. 
+The experimental results of the energy saving, performance degradation and tradeoff distance are 
+presented in the figures \ref{fig:edp-eng}, \ref{fig:edp-perf} and \ref{fig:edp-dist} respectively. 
+
+In one site scenario the proposed frequencies selection algorithm outperform the EDP algorithm 
+in term of energy and performance for all of the benchmarks. While, the compassion results from the two sites scenario 
+showed that the proposed algorithm outperform EDP algorithm for all benchmarks except MG benchmark.
+In case of MG benchmark the are small communications and bigger frequencies selected in EDP algorithm 
+decreased the performance degradation more than the frequencies selected by Maxdist algorithm. 
+While the energy saving percentage are higher for Maxdist algorithm.
+
+Generally, the proposed algorithm gives better results for all benchmarks because it
+optimized the distance between the energy saving and the performance degradation. 
+Whereas, in EDP algorithm gives negative tradeoff for some benchmarks in the two sites scenarios.
+These negative tradeoffs mean the performance degradation percentage is higher than energy saving percentage.
+The higher positive value for tradeoff distance is mean the best energy and performance tradeoff is achieved synchronously, when
+the energy saving percentage is much higher than the performance degradation percentage 
+The time complexity of the proposed algorithm is $O(N \cdot M \cdot F)$, where $N$ is the number of the clusters,
+$M$ is the number of nodes and $F$ is the maximum number of available frequencies. The algorithm is selected 
+the best frequencies in small execution time, on average is equal to  0.01 $ms$ when it works over 32 nodes.
+While the EDP algorithm was slower than Maxdist algorithm by ten times, where their execution time  on average 
+takes 0.1 $ms$  to selects the suitable frequencies over 32 nodes. 
+The time complexity of this algorithm is  $O(N^2 \cdot M^2 \cdot F)$.
+
+
+  
+ 
+
+
+
+\begin{figure}
+  \centering
+  \includegraphics[scale=0.5]{fig/edp_eng}
+  \caption{Comparing of the energy saving for the proposed method with EDP method}
+  \label{fig:edp-eng}
+\end{figure}
+
+\begin{figure}
+  \centering
+  \includegraphics[scale=0.5]{fig/edp_per}
+  \caption{Comparing of the performance degradation for the proposed method with EDP method}
+  \label{fig:edp-perf}
+\end{figure}
+
+\begin{figure}
+  \centering
+  \includegraphics[scale=0.5]{fig/edp_dist}
+  \caption{Comparing of the tradeoff distance for the proposed method with EDP method}
+  \label{fig:edp-dist}
+\end{figure}
 
   
 \section{Conclusion}