Author Topic: estimate of nrecex in trfdrs seems not very well [?]  (Read 3045 times)

resofidentity

  • Full Member
  • ***
  • Posts: 63
  • Karma: +0/-0
estimate of nrecex in trfdrs seems not very well [?]
« on: July 14, 2010, 09:09:54 am »
Hello,

when using mpshift for mp2 shieldings I get the following:

***************************************************************************
  ALL STUFF        :          4000  MB
  INFO : blocks of          4096  Bytes needed:
         gamma#1    dtdb#1      dtdb#2
      -268914      -806742       241833
            +          +
        56511       169533

 record length scale factor =        8

  WARNING: space for file intermed exhausted, i will continue anyway
 screen:  ijsym   little   screen1  screen2 |      klsym    screen1    screen2
              0.   30178.       0.       0. |     55110.         0.1798185888.
 size of the actual   halfint   file:   -12680 MB   allowed:    -9714 MB
 actual number of BF-pairs for half-trafo was :     350538
 cpu for first half-transformation                      32909.80 s
  <trf2pq>: error: irecex > nrecex             1      -264153
  the estimate of nrecex in trfdrs  seems not very well!

 MODTRACE: no modules on stack

  <trf2pq>: error: irecex > nrecex
 mpshift ended abnormally
 mpshift ended abnormally
program stopped.
*******************************************************************************

I did the statistics run  followed by dscf and then: mp2prep -c -m 4000
the halfint file is 148GB which should be no problem (although in control
mp2prep wrote):
 type=intermed unit=61       size=678414   file=halfint

I use 5.9 linux 64bit (parallel).

Anyone any idea?

Best regards
  R