Timing in seconds for MbCO + 3830 water molecules (14026 atoms), 1000 steps 12-14 A shift on IBM-SP1.


Sorry I am still working on this, so it is still in LaTeX:


\begin{table}[ht]
\begin{tabular}{|c|r|r|r|r|r|r|r|r|c|r|} \hline
Nodes & E$_{\rm ext}^1$ & E$_{\rm int}^2$ & Wait$^3$ & Comm$^4$ & List$^5$ & Integ$^6$ & 
\multicolumn{1}{c|}{Total} & Eff.$^7$ & Speedup  \\ \hline
    1 &21349.6 & 137.6 &   0.0 &    0.1 & 2077.5 & 32.7 &23596.9 = 6.55 h &  100.0\% &    1.0  \\ \hline
    2 &10814.8 &  77.2 & 228.9 &   85.9 & 1101.6 & 17.2 &12325.6 = 3.42 h &   95.7\% &    1.9  \\ \hline
%   Commented numbers are new from ANL
%   2 & 9850.3 &  72.1 &  79.7 &   82.9 & 1214.1 & 16.2 &11315.4
    4 & 5358.0 &  47.1 & 246.9 &  133.3 &  608.8 &  9.1 & 6403.2 = 1.78 h &   92.1\% &    3.7  \\ \hline
    8 & 2658.3 &  30.8 & 252.5 &  170.8 &  354.9 &  5.4 & 3472.7 = 0.96 h &   84.9\% &    6.8  \\ \hline
   16 & 1345.2 &  21.9 & 358.2 &  223.3 &  175.0 &  4.3 & 2127.9 = 0.59 h &   69.3\% &   11.1  \\ \hline
\end{tabular}
\end{table}


E ext : External energy terms (electrostatics + Lenard-Jones) E int : Internal energy terms (bond, angle, dihedral) Wait : Load unbalance Comm : Communication time (Vector Distr. Global {Sum,Brdcst}) List : Nonbond list generation time Integ : Time needed to integrate equations of motion Total : Total elapsed time Eff : Efficiency = speedup divided by number of nodes Speedup: Time for N nodes divided by time for one node.

Comments

See IBM WWW pages for more details