LSD and MULTIPLICITY can not work together?

[Enlong Liu]

Dear CP2K Users,

I want to do spin polarized calculation for single Si atom. I check the manual, which says that the parameter of MULTIPLICITY should be an integer and for Si I think it should be 3 for triplet (http://manual.cp2k.org/cp2k-2_4-branch/CP2K_INPUT/FORCE_EVAL/DFT.html#desc_MULTIPLICITY). And I still need to put LSD (UKS) into "True" to start the spin polarized calculation (http://manual.cp2k.org/cp2k-2_4-branch/CP2K_INPUT/FORCE_EVAL/DFT.html#desc_UKS).  But the program gives error when it is submitted which is:

 ASSERTION FAILED: F


  stack:
  error in dbcsr_multiply_anytype at line   577 with error type 100
  message: Invalid last k specified (A)
    1 error in dbcsr_multiply_anytype at line   577


 libdbcsr| Abnormal program termination, stopped by process number 0
--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.


Can any one help me with it? How to fix the problem and realize spin polarized calculation? In the attachment is my input. Thanks a lot for your help!

Best regards,

[Matthias Krack]

Dear Enlong Liu,

your input works for me, i.e. shows a normal termination, using the pre-compiled binary cp2k-2.5.1-Linux-x86_64.ssmp from http://sourceforge.net/projects/cp2k/files/precompiled

Best,

Matthias

[Enlong Liu]

Dear Matthias,

Thank you for this reply and the link! Unfortunately, I use CP2K on the high-performance computer of our university, where they only have cp2k 2.4 pre-installed. So I wonder if you can tell me the final result of the energy calculation, I will check it and consider if it is necessary to install what you recommended on my PC.

Best regards,

Enlong

在 2014年6月25日星期三UTC+8下午5时13分13秒，Matthias Krack写道：

[Matthias Krack]

Dear Enlong Liu,

the output generated with cp2k-2.5.1-Linux-x86_64.ssmp is attached. There is no need for an installation. You just download the binary and run it. This should work just fine for small cases which can be run on a PC.

Matthias

[Enlong Liu]

Dear Matthias,

I tried the program you gave me. There is always an error saying that CPU time limit exceed (core dumped). Since it is really urgent for me to get the single atom energy value, so I wonder if you could do me a favor and help me calculate Ag atom. The xyz file and input file are in the attachment. Thank you very much!

Kind regards,

Enlong

在 2014年6月25日星期三UTC+8下午10时08分51秒，Matthias Krack写道：

[Matthias Krack]

Dear Enlong,

this is not a small case even if you have only one Ag atom in the cell, because your simulation cell is pretty large with 60Ax60Ax60A given a 300Ry cutoff. The CP2K ssmp binary which I proposed is not suited for such runs. Sorry, calculations of this size are beyond my service level. Well, maybe somebody else is willing to run the input for you. Alternatively, you have to reduce the cell size to something reasonable like for the Si atom (20A edge length).

Matthias