How to use SDPARA-C

[Copyright (C) 2004 SDPA Project]

Here we explain the executable binary of the SDPARA-C shortly. 
The usage of the SDPARA-C is quite similar to the stand-alone 
use of the SDPA-C. Basically, instead of the command "./sdpa-c",
type "mpirun -np number_of_processors ./sdpara-c". 
Refer to "sdpa-c.x.x.x.manual.pdf" for more details on the SDPA-C.
The SDPARA-C reads SDPA format files. If you want to solve
example1.dat-s (SDPA sparse format), type

$ mpirun -np 4 ./sdpara-c example1.dat-s example1.result

Here "mpirun" is the command to execute the MPI program, 
"-np 4" indicates that we use 4 processors, "./sdpara-c" 
is the name of the SDPARA-C executable binary, "example1.dat-s" 
is the input file to solve, and "example1.result" is an 
output file which contains the solutions and other information.
Note:: Assign at least one input file and one output file. 
You will see the following messages.


-------------------------------------------------------------------------

SDPARA-C start at Fri Nov  5 09:55:10 2004
let me see your ...
data      is example1.dat-s : sparse
parameter is ./param.sdpa
out       is example1.result
aggregate sparsity pattern :                                 4 elements
extended sparsity pattern  :
        METIS4.0.1 (multilevel nested dissection)            4 elements
        Spooles2.2 (minimum degree)                          4 elements
        Spooles2.2 (generalized nested dissection)           4 elements
        Spooles2.2 (multisection)                            4 elements
        Spooles2.2 (best of ND and MS)                       4 elements
        Selecting ..... METIS4.0.1(multilevel nested dissection)
dense matrix               :                                 4 elements
----------------------------------------
   mu      thetaP  thetaD  objP      objD      alphaP  alphaD  beta 
 0 1.0e+04 1.0e+00 1.0e+00 -0.00e+00 +1.20e+03 1.0e+00 9.1e-01 3.00e-01
 1 1.6e+03 2.9e-17 9.4e-02 +8.16e+02 +7.51e+01 5.6e+00 9.6e-01 3.00e-01
 2 2.1e+02 2.3e-16 3.6e-03 +2.59e+02 -3.74e+01 2.0e+00 1.0e+00 3.00e-01
 3 3.8e+01 2.1e-16 1.5e-17 +3.35e+01 -4.19e+01 9.7e-01 9.7e-01 1.00e-01
 4 4.8e+00 2.2e-16 7.5e-18 -3.22e+01 -4.19e+01 9.9e-01 9.9e-01 1.00e-01
 5 5.3e-01 2.3e-16 1.9e-18 -4.08e+01 -4.19e+01 1.0e-00 1.0e-00 1.00e-01
 6 5.3e-02 2.0e-16 3.7e-18 -4.18e+01 -4.19e+01 1.0e-00 1.0e-00 1.00e-01
 7 5.3e-03 2.2e-16 9.3e-19 -4.19e+01 -4.19e+01 1.0e-00 9.0e+01 1.00e-01
 8 5.3e-04 2.2e-16 2.9e-15 -4.19e+01 -4.19e+01 1.0e-00 1.0e-00 1.00e-01
 9 5.3e-05 2.1e-16 1.5e-17 -4.19e+01 -4.19e+01 1.0e-00 1.0e-00 1.00e-01
10 5.3e-06 2.0e-16 7.5e-18 -4.19e+01 -4.19e+01 1.0e-00 1.0e-00 1.00e-01
11 5.3e-07 2.0e-16 2.2e-17 -4.19e+01 -4.19e+01 1.0e-00 1.0e-00 1.00e-01

phase.value = pdOPT     
   Iteration = 11
          mu = 5.3127653470891417e-07
relative gap = 2.5359263268858051e-08
         gap = 1.0625530694178283e-06
      digits = 7.5958633676245801e+00
objValPrimal = -4.1899998937446902e+01
objValDual   = -4.1900000000000020e+01
p.feas.error = 2.4646951146678475e-14
d.feas.error = 2.1316282072803006e-14
total time   = 0.094
file read time = 0.028777
main loop time = 0.061296
total     time = 0.093691

-------------------------------------------------------------------------

"phase.value = pdOPT" indicates that the SDPARA-C finds successfully
the primal and dual optimal solutions with the optimal primal 
objective function value "objValPrimal", and the optimal dual 
objective function value "objValDual". "total time" displays how 
long the SDPARA-C needs to solve example1.dat-s. Notice that the 
time reported by the SDPARA-C is the real time, and not the user 
time.