Dear Sf kshp,
According to the output, it is caused by insufficient memory. It
means that your computer system could not run siesta in such scale.
Basically, it's true that larger box size should get smaller
effect of image box when doing molecular calculation. Actually, it
depends on users to choose the box size. I think when the basis
doesn't overlap any basis in image box, Siesta will print something
like "siesta: System type = molecule", it is the starting point to do
molecular calculation. System energy should have a convergence
behavior about box size.
Best wishes,
Kuilin Lu
On 5/26/12, sf kshp <[email protected]> wrote:
> Dear Kuilin Lu,
> Thanks a lot for your comprehensive information , actually I use 32
> bit ubuntu 11.10 ( memory :3.4 GB ,Disk: 320 GB, Processor: intel
> pentium(r) cpu G620 ) gfortran is my compiler and use serial siesta 3.1. so
> u think is that ok to do my calculation with a smaller size box(like 30 or
> 35 Ang)? or do u have any recommendation?
>
> Best, Sf kshp
>
> InitMesh: MESH = 480 x 480 x 480 = 110592000
> InitMesh: Mesh cutoff (required, used) = 300.000 314.455 Ry
> alloc_err: allocate error
> alloc_err: allocate status error 5014
> alloc_err: array CG requested by poison
> alloc_err: dim, lbound, ubound: 1 1 2
> alloc_err: dim, lbound, ubound: 2 1********
> alloc_err: allocate error
> ERROR STOP from Node: 0
> ERROR STOP from Node: 0
>
>
