[Csdp] Problems with large SDP problems

[Brian Borchers]

Unfortunately, all of your SDPA files are incorrect.  Although CSDP would
ideally be able to detect this, it doesn't do so, and then the solution
process fails.  Anyway, here's an explanation of what's wrong with the file
for N=1.  The other files all appear to have the same defect:

Here's the file:

5
2
4 4
1 0 0 0
0 1 2 2 -1.000000e-02
0 1 1 2 -5.027783e-01
0 1 3 3 -1.000000e-02
0 1 1 3 4.534177e-01
0 1 4 4 -1.000000e-02
0 1 1 4 7.358455e-01
0 2 1 1 -5.000000e-01
0 2 2 2 -5.000000e-01
0 2 3 3 -5.000000e-01
0 2 4 4 -5.000000e-01
1 1 1 1 1
2 1 1 2 -1
2 2 1 2 5.000000e-01
2 2 3 4 5.000000e-01
3 1 1 3 -1
3 2 2 3 5.000000e-01
3 2 1 4 -5.000000e-01
4 1 1 4 -1
4 2 1 1 5.000000e-01
4 2 2 2 -5.000000e-01
4 2 3 3 5.000000e-01
4 2 4 4 -5.000000e-01

In the first line of the file, you say that the problem has five constraints
matrices (A1, A2, A3, A4, and A5), plus an objective matrix C.  In line 2,
you say that there are two SDP blocks in each matrix, of sizes 4x4 and 4x4
(line 3.)  In line 4, you provide the right hand sides for four
constraints.  This is incorrect, since you said that there were five
constraints.  In the following lines you give the entries for C and the
constraint matrices A1, A2, A3, and A4, but not A5.