Hello,

I am trying to use HPL with LAM. I could successfully compile HPL, but I am having some run-time problems.

Whenever I execute the HPL software with a problem a little large (say, N
= 1000), one process exits abnormaly, as shown by the message below. I
tried different configurations for P and Q but the problem persists.
Executing small problems (N = 100) the program works fine.

Sample error message:

MPI_Wait: process in local group is dead (rank 0, comm 4)
Rank (0, MPI_COMM_WORLD): Call stack within LAM:
Rank (0, MPI_COMM_WORLD):  - MPI_Wait()
Rank (0, MPI_COMM_WORLD):  - main()
-----------------------------------------------------------------------------
One of the processes started by mpirun has exited with a nonzero exit
code.  This typically indicates that the process finished in error.
If your process did not finish in error, be sure to include a "return
0" or "exit(0)" in your C code before exiting the application.

PID 18713 failed on node n1 due to signal 4.
-----------------------------------------------------------------------------



My hardware configuration is a Pentium 4 (512MB of memory) as main node
and 64 Pentium 3 machines with 256MB of memory each. I am using a Fast
Ethernet 100Mbps for communication.

below is an example HPL.dat file that does not work:

HPLinpack benchmark input file
Innovative Computing Laboratory, University of Tennessee
HPL.out      output file name (if any)
8          device out (6=stdout,7=stderr,file)
1            # of problems sizes (N)
1000         Ns
1            # of NBs
200          NBs
1        # of process grids (P x Q)
2            Ps
2            Qs
16.0    threshold
3            # of panel fact
0 1 2        PFACTs (0=left, 1=Crout, 2=Right)
2            # of recursive stopping criterium
2 4      NBMINs (>= 1)
1            # of panels in recursion
2            NDIVs
3            # of recursive panel fact.
0 1 2        RFACTs (0=left, 1=Crout, 2=Right)
1            # of broadcast
0            BCASTs (0=1rg,1=1rM,2=2rg,3=2rM,4=Lng,5=LnM)
1            # of lookahead depth
0  DEPTHs (>=0)
2            SWAP (0=bin-exch,1=long,2=mix)
64  swapping threshold
0            L1 in (0=transposed,1=no-transposed) form
0            U  in (0=transposed,1=no-transposed) form
1            Equilibration (0=no,1=yes)
8            memory alignment in double (> 0)

Any ideas?

Thank you very much,
Leonardo.