libbem

Presented are the times for building the panel-clustering-operator and for solving the system via GMRES. The speedup is calculated by adding both times.

An overview of the results is given at the and of this page.

Sun E6000

level time for PCop time for GMRES speedup memory
5 146.6   49.4   1.0 412  
6 707.9   241.0   1.0 1846  
7 2604.3   1162.6   1.0 8185  
8 9581.0   5522.4   1.0 36300  

Sun SunFire 6800

level time for PCop time for GMRES speedup memory
5 43.5   25.1   2.86 412  
6 215.3   126.2   2.78 1846  
7 768.9   608.1   2.74 8185  
8 3232.8   3077.3   2.39 36122  

Sun SunFire 15k

level time for PCop time for GMRES speedup memory
5 44.0   26.3   2.79 412  
6 219.3   136.6   2.67 1859  
7 784.8   688.9   2.56 8185  
8 3412.5   4717.9   1.86 36122  

HP 9000 Superdome

level time for PCop time for GMRES speedup memory
5 50.0   20.1   2.80 -  
6 210.4   99.5   3.06 -  
7 2136.4   502.7   1.43 -  
8 3665.0   2280.8   2.54 -  
  -   From 16k to 65k unknowns one can see a dramatic increase in the time to build the panel-clustering operator. In contrast to this, the floating-point performance scaled rather well, as one can see at the GMRES-times.  

IBM eServer p690

level time for PCop time for GMRES speedup memory
5 50.7   13.1   3.07 412  
6 216.3   65.1   3.37 1833  
7 885.0   351.2   3.05 8199  
8 3719.9   1770.8   2.75 36112  

AMD Athlon MP 2000

level time for PCop time for GMRES speedup memory
5 26.4   13.5   4.92 391  
6 113.3   65.4   5.31 -  

Overview

The following table shows the speedup of the systems compared to the Sun E6000. Written with a red color is the fastest system on each level.

level SF6800 SF15k HP9000 P690 Athlon
5 2.86   2.79   2.80   3.07   4.92  
6 2.78   2.67   3.06   3.37   5.31  
7 2.74   2.56   1.43   3.05   -  
8 2.39   1.86   2.54   2.75   -