Commit 5a0ec933 by Francois Gygi

release 1_12_1

git-svn-id: cba15fb0-1239-40c8-b417-11db7ca47a34
parent ca2f588b
......@@ -21,9 +21,6 @@ Incorrect (uninitialized) values are printed by BOSampleStepper. Must either
include calculation of ekin_ in SDIonicStepper, or suppress printing of
ekin_ion and temp_ion if atoms_dyn = SD.
The quadrature formula for integration of semi-local potentials must be
improved to at least Simpson level.
Saving a sample with wavefunctions of size zero crashes in FourierTransform
(size zero not working)
......@@ -51,8 +48,19 @@ determine library search path, even if makefile specifies another path.
Created a libxerces-c.a using objects in $XERCESCDIR/lib and linked statically.
This will avoid the problem of needing when using qbox.
Removed Gauss-Jacobi quadrature in NonLocalPotential introduced in 1_12_0 and
reverted to trapezoidal rule. Tests on Si64 using trapezoidal, Simpson, Gauss-
Legendre and Gauss-Jacobi show that the trapezoidal rule is more efficient
than all other methods, likely because higher order methods are not in their
asymptotic range for reasonable numbers of points (4-32). A possible
improvement over the present status may be to use Richardson extrapolation,
since the trapezoidal rule has O(h^2) error. However, from the example of Si64,
little improvement is obtained for small (4-8) numbers of points.
Replaced trapezoidal integration in NonLocalPotential.C by a Gauss-Jacobi rule.
(Note: this was later found to be less efficient. Removed in 1.12.1e)
Fixed bug in PSDWavefunctionStepper.C and PSDAWavefunctionStepper.C:
use ecut if ecutprec==0.0. Caused nans when ecutprec was not set.
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment