CHANGES OF VERSION 5.1.8

1.  Ports to new machines or machine specific changes

2.  BUGS

3. INPUT RELATED CHANGES

4. DEFAULTS in input and throughout the code

5. OUTPUT related changes

6. CLEANING up

7. MINOR or relatively insignificant changes

8. ALGORITHMS: additions or changes  to existing functional ability

1. PORTS to new machines, re-ports, or machine specific changes

·   added RUNPARATEC variable to sysvars.0200 and .sp3nersc -D. Roundy

2. BUGS

· various perl library fixes.

3. INPUT related changes

· added input keywords for MD structures

MD_time_step, MD_temp, MD_max_iter, ensemble_type, extrapolation_method,

MD_Q_mass

·   added occupy_levels from_input option for static and specified

occupations for magnetic insulators ! F.Mauri

· added new keywords of number_of_bands_alpha and number_of_bands_beta

· added new keywords of number_of_alpha and number_of_beta !F. Mauri

4. DEFAULTS in input and throughout the code

· default for max_iter_diag_band from 40 to 137.

5. OUTPUT related changes

· changed output of screening potential for bandstructure calc.-hjchoi

· small change to output in plot bandstructure to output fermi level

for both spins - D. Roundy

· Output a SCF_KPOINTS file every time we do an scf loop, which can be

used as a KPOINTS file to be input in another calculation. - D. Roundy

· More digits in weights of k points in OUT file. -D. Roundy

· changed output of neighbor.f90 to handle very large systems.

Now outputs neighbors with 1.2 times the n.n. length

· electron phonon output. output flags elctronphonon and hamwavefn added

·   added timing of init_diag and total time to direct routines

· output the rcut value for each PP when real space is done

· output pw_params%nbandsfft if it is greater than 1

6. CLEANING up

·   changed the main program from relax.f90p to main.f90p

relax.f90 contains relaxation routine and is called from main.f90p

· made constant num_time to give size of timing array

· cleaned up main.f90p. removed unused variables. added definitions to the

variables. moved creation of crystal variable to within main.f90p

Thus removing more artifacts of original C implementation

fewer variables are now needed for calls to relax,funeval,pwmain,md,.

·  altered direct_emin,cgemin_gcg, and cgemin to handle

magnetic insulating systems

8. ALGORITHMS: additions or changes  to existing functional ability

· added molecular_dynamics structure

· added mass,lvec, and coord to element of crystal structure

· changed inner product in lbfgs back to original method

· added Nose'-Hoover dynamics with generalized leapfrog integration

· added constant energy MD - with velocity verlet integration

created new routine this

· added subspace alignment to pwmain. - cire Arias PRB '92

· made all update_hamiltonian routines to handle mutiple band FFT

calculating charge density

· updated pwmain to handle extrapolation of wavefunctions and/or

(n-n(atomic)). extrapolaton from Arias PRB '92

· fixed (hopefully for real this time) the fermi level when you do an

scf loop followed by a band structure, so you don't need to input it

at all.

·  changed bands%min to a 2 element array. one for each spin

This caused change in many routines and hopefully I handled

everything properly in unfmailiar routines.

·  added nalpha and nbeta elements to crystal structure ! F. Mauri