The ABINIT Users Mailing List ( CLOSED )

[mamikon Gulian <mamikon.armen.gulian@gmail.com>]

Thanks for your rapid suggestions. I have 128 Gb of memory for this calculation, but I am willing to believe that the problem is with kssform=1. However, I thought there were downsides with using kssform=3, namely that there will be " a restricted number of states". 
Any case, its the only option here.

Once again thanks,
Mamikon

On Mon, Mar 8, 2010 at 11:47 AM, Matteo Giantomassi <Matteo.Giantomassi@uclouvain.be> wrote:

> Hi Mamikon,
>
> It says "  -  " ?
> Anyway, it seems that the code have a case in which the
> memory needed is negative. Maybe you can search
> that part on the code and know when this case is on.
>

 It's due to a wrap-around error in memkss.
 It will be solved in v6.1. It's not a serious problem and it's not
 the cause of the malloc error.
 Use kssform=3 as suggested by Fabien otherwise the KSS
 generations is unfeasible.

 Matteo

> Regards
>
> On Mon, Mar 8, 2010 at 5:13 PM, mamikon Gulian <
> mamikon.armen.gulian@gmail.com> wrote:
>
>> Strange things in the log file:
>> *
>> Total amount of disk space required by _KSS file =  -450.94 Mb.*
>>   Subdivided into :
>>   Header             =     0.82 Mb.
>>   KB elements        =   108.62 Mb.
>> *  Wavefunctions (PW) =  -560.38 Mb.*
>>   PAW projectors     =     0.00 Mb.
>>
>>
>>  Opening file for KS structure output: nt-nonscf_o_KSS
>>  number of Gamma centered plane waves  53927
>>  number of Gamma centered shells  53927
>>  number of bands  52235
>>  maximum angular momentum components      3
>>  number of symmetry operations  1 (without inversion)
>>    *   g( 1)=  0  0  0
>>    *   g( 2)=  1  0  0
>>    *   g( 3)= -1  0  0
>>    *   g( 4)=  0  1  0
>>    *   g( 5)=  0 -1  0
>>    *   g( 6)=  1  1  0
>>    *   g( 7)= -1  1  0
>>    *   g( 8)=  1 -1  0
>> *   *   g(****)=-13  2-11
>>    *   g(****)=  2 13-11
>>    *   g(****)= -2 13-11
>>    *   g(****)=  2-13-11
>>    *   g(****)= -2-13-11
>>    *   g(****)= 13 -2-11
>>    *   g(****)=-13 -2-11*
>> -P-0000
>> -P-0000  k-point   1
>> -P-0000  Calculating <G|H|G'> elements
>>
>>
>> On Mon, Mar 8, 2010 at 11:00 AM, mamikon Gulian <
>> mamikon.armen.gulian@gmail.com> wrote:
>>
>>> Dear ABINITers,
>>>
>>> I got the following unexpected error with ABINIT 5.8.4. This occurs
>>> during
>>> a non-scf KSS file calculation, after the program writes the g-vectors
>>> to
>>> the file and begins to calculate < G | H | G > elements.
>>>
>>> Running with mpi (Open MPI 1.4.1), 11 cpus:* glibc detected *** abinip:
>>> malloc(): memory corruption: 0x0000000002e0ac50 ***
>>> *
>>> I though it was a bug with MPI, so I tried a sequential run: * ***
>>> glibc
>>> detected *** abinis: malloc(): memory corruption: 0x0000000004d80260
>>> ****
>>>
>>> I never got such an error before. Anyone have an idea?
>>>
>>> Thanks
>>>
>>> Mamikon
>>>
>>
>>
>
>
> --
> Atte.
> Keitel Cervantes Salguero
>