CCL:G: Gaussian problem: What is PrismC?

From: "Roman D. Gorbunov" <rgorbuno^-^aecom.yu.edu>
Subject: CCL:G: Gaussian problem: What is PrismC?
Date: Mon, 3 Dec 2007 12:51:38 -0500
 Sent to CCL by: "Roman D. Gorbunov" [rgorbuno++aecom.yu.edu]
 Dear CCL Subscribers,
 I run a set of similar Gaussian jobs (geometry optimization + normal mode
 frequencies calculations for system consisting of 61 atoms). All this
 calculations have been finished with the "Normal Termination".
 However, I am not able to get "Normal Termination" if I slightly
 change the system. I just remove some atoms and add new ones (so that the totals
 size of the system increases to 64 atoms).
 In the "problematic" case I have no error message. Gaussian just
 stops. In the and of the output file I find the following:
  Number of processors reduced from  4 to  1 for far-field.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         4.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         4.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         4.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         4.
  Number of processors reduced from  4 to  1 for far-field.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         4.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         4.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         4.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         4.
  Number of processors reduced from  4 to  1 for far-field.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         4.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         4.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         4.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         4.
 It is repeating many times and the Gaussian stops with the following message:
      3 vectors were produced by pass224.
  Number of processors reduced from  4 to  1 for far-field.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         3.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         3.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         3.
  PrismC:  NFx=      2048 NFxT=         4 NFxU=         3.
      3 vectors were produced by pass225.
  Number of processors reduced from  4 to  1 for far-field.
 Does any body know what Gaussian try to do here and why it cannot finish that?
 Thank you in advance,
 Roman