nnr/nnrx (CP) => nrxx ; nnrs/nnrsx (CP) => nrxxs
(the other way round would have been preferrable but it was too much work)
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@7143 c92efa57-630b-4861-b058-cf58834340f0
- dspev and zhpev parallel driver moved to independend module,
ptoolkit has grown too much!
- new subroutine in ptoolkit for parallel block-like cholesky decomposition
and triangular matrix inversion (now only for hermitian matrixes)
- new subroutine in ptoolkit for compex matrix multiplication
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@4103 c92efa57-630b-4861-b058-cf58834340f0
lambda matrixes on a square mesh PxP of processors.
The number of processors used in the run not necessarily should be equal
to a perfect square PxP, the code, in distributing lambda,
try to use an optimal (for performances) square PxP less or equal than the
number of procs used.
- the size (Np=PxP) of the processor mesh to be used in distributing lambda
and ortho, can be suggested using the namelist keyword
ortho_para = Np
in the electrons namelist
- the distribution of lambda matrixes is required to save
memory in run with an high number of bands.
In a system with 2800 bands, the memory saved is about 200Mbyte
per proc/core if a sufficient number of proc ( some hundreds )
is used.
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@3643 c92efa57-630b-4861-b058-cf58834340f0
- Many ortho auxiliary functions (tauset, rhoset, sigset, calphi, updatc)
are now in common between FPMD/CP, and moved to module ortho_base.f90
- In FPMD, three index vectors, related to real space like charge and potential
have been substituted with single index vector like in CP, for compatibility
and efficiency.
- Bug fix in pwtools/matdyn.f90 a logical variable was used in place of a
character variable
git-svn-id: http://qeforge.qe-forge.org/svn/q-e/trunk/espresso@2694 c92efa57-630b-4861-b058-cf58834340f0
ccavazzoni