AMBER Archive (2000)

Subject: 4096wat runnning problem

From: Kouji Hauyama (hayama_at_lance.mecha.osakafu-u.ac.jp)
Date: Sat May 27 2000 - 07:41:24 CDT

Dear David and Michael and amber user

I re-try to run 4096wat the way the Michael and David say ,but I
couldn't run 4096wat .
And I got the same meesages.

---------------The messages--------------------------

| Atom division among processors:
| 0 6144 12288
| Atom division among processors for gb:
| 0 6144 12288
| Running AMBER/MPI version on 2 nodes

     Sum of charges from parm topology file = 0.00000000
     Forcing neutrality...
 ---------------------------------------------------
 APPROXIMATING switch and d/dx switch using CUBIC SPLINE INTERPOLATION
 using 5000.0 points per unit in tabled values
 TESTING RELATIVE ERROR over r ranging from 0.0 to cutoff
| CHECK switch(x): max rel err = 0.3242E-14 at 2.436720
| CHECK d/dx switch(x): max rel err = 0.8156E-11 at 2.782960
 ---------------------------------------------------
     Total number of mask terms = 12288
     Total number of mask terms = 24576
| Total Ewald setup time = 0.31999999

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

  Unit 7 Error on OPEN:

  Unit 7 Error on OPEN:
[1] MPI Abort by user Aborting program !
[1] Aborting program!
p1_24122: p4_error: : 1
rm_l_1_24123: p4_error: interrupt SIGINT: 2
bm_list_24116: p4_error: net_recv read: probable EOF on socket: 1

----------------The messages End --------------------------

And I tested the progran David said.
> mpicc test.c
> mpirun -np 2 a.out
So, I got the messages ....

Starting
Node 0 to 1: throughput 34.96 Mbytes/sec
Node 0 to 1: throughput 35.26 Mbytes/sec
Node 0 to 1: throughput 34.97 Mbytes/sec
Node 0 to 1: throughput 35.27 Mbytes/sec
Node 0 to 1: throughput 34.72 Mbytes/sec

I think that this messages are the speed of message passing.
The score is correct. Because I test the program on SMP Machine.

But I can't solve the problem of 4096wat.
Would you tell me any hint?
Thank you.