AMBER Archive (2003)

Subject: [Fwd: MPI problem in modi4.ncsa.uiuc.edu]

From: Eddy (zewen_at_uci.edu)
Date: Thu Mar 13 2003 - 23:57:23 CST

Dear Sir,
I am running SANDER in modi4.ncsa.uiuc.edu, a SGI cluster. When I tried
to run MPI -4 (4 cpu parallel computing) job, I found the error message
in the output file (after running a couple of picoseconds the job
crashed). But when I run 2 cpu parallel job, it is okay.

Best,

Eddy

--------------------------------------------------------------------------------
   4. RESULTS
--------------------------------------------------------------------------------

 ---------------------------------------------------
 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 = 0.2088E-14 at 2.598900
| CHECK d/dx switch(x): max rel err = 0.7671E-11 at 2.757160
 ---------------------------------------------------
| Local SIZE OF NONBOND LIST = 1271168
| TOTAL SIZE OF NONBOND LIST = 3762323
***** Processor 0
***** System must be very inhomogeneous.
***** Readjusting recip sizes.
 In this slab, Atoms found: 7327 Allocated: 7075