feram-0.26.03

Dear current and future feram users,

I released feram-0.26.03.tar.xz from https://sourceforge.net/projects/loto/files/feram/feram-0.26.03/ . This is a stable release.

MD5 (feram-0.26.03.tar.xz) = 5f2b09aecdea0f69c5ed56dac23927dc

feram is a fast molecular dynamics (MD) simulator for bulk and thin-film ferroelectrics. Its homepage is http://loto.sourceforge.net/feram/ .

Enjoy,

Takeshi

Changes

With this feram-0.26.03, you can draw a slice of the supercell directly by feram. Output file foo123.nnnnnnnnnn.slice.eps is the visualized slice of a film at z=Lz/4, at the nnnnnnnnnn-th iteration and at evry n_coord_freq iterations. Output of this .slice.eps file can be suppressed with slice_directory = 'never' in the .feram input file.

In the feram-0.26.02, some bugs and typos are fixed. For example, src/24example-BaTiO3-new-param-heating-cooling/polarization.gp .

In the feram-0.26.01, some bugs and typos are fixed.

From the feram-0.26.00, you can use .modulation input file. You can simulate (Ba,Sr)TiO3, now. See files in feram-0.26.00/src/34example-BST/. See http://arxiv.org/abs/1608.06433 also.

In the feram-0.26.00, a bug in feram-0.24.02 of plot_dispersion is fixed.

From this feram-0.26.00, a .dipoRavg file will be written in the end.

Tips for *.slice.eps

animation

convert(1) of http://www.imagemagick.org gifsicle(1) and figview(1) of https://www.lcdf.org/gifsicle/ are useful.

for e in *.eps; do convert -density 40 -flatten $e `basename $e .eps`.gif; done
gifsicle --delay=5 e*.gif > animation.gif

color to black-and-white

Edit .eps files dilectly, or you can edit src/coord_module.F or src/feram_slicer.rb .

--- color.slice.eps 2016-10-21 21:19:02.000000000 +0900
+++ bw.slice.eps    2016-10-21 21:21:39.000000000 +0900
@@ -25,7 +25,7 @@
   0    y_height LineWidth20 sub rlineto
   x_width LineWidth20 sub neg 0 rlineto
  closepath
- 1 0 0 setrgbcolor
+ 0.5 setgray
  LineWidth setlinewidth
  stroke
 } def
@@ -42,7 +42,7 @@
   0    y_height LineWidth sub rlineto
   x_width LineWidth sub neg 0 rlineto
   closepath
-  0 0 1 setrgbcolor fill
+  0.0 setgray fill
 } def
 0 0 Down
 1 0 Down

Known bug

OpenMP-parallelization does not work efficiently over two or more NUMA chips of Intel Xeon CPUs which is released after 2011 (Sandy Bridge, Ivy Bridge, Haswell, Broadwell, ...). Execute feram within single chip by using numactl(8) command.

Intel Fortran (ifort) cannot make a fast executable, because there is no optimized fftw_plan_many_dft_r2c() nor ftw_plan_many_dft_c2r() in Intel Math Kernel Library (MKL), so far.

Hysteresis loop calculations of film goes wrong. A strange peak appears in the beginning. See src/23example-BaTiO3-new-param-hl-film-acoustic-mass/. Bulk calculations go very fine.

Version | acoustic_mass_amu | Good or N.G.
------- | ----------------- | ------------
0.20.00 | with              | strange peak
0.20.00 | without           | Good
0.21.04 | with              | strange peak
0.21.04 | without           | Good
0.22.00 | with              | strange peak
0.22.00 | without           | Good
0.22.01 | with              | strange peak
0.22.01 | without           | Good
0.22.02 | with              | strange peak
0.22.02 | without           | Good
0.22.03 | with              | paraelectric
0.22.03 | without           | paraelectric
0.22.04 | with              | paraelectric
0.22.04 | without           | paraelectric
0.22.05 | with              | strange peak
0.22.05 | without           | OK
0.22.06 | with              | strange peak
0.22.06 | without           | OK
0.23.0? | with              | strange peak
0.23.0? | without           | OK
0.24.0? | with              | strange peak
0.24.0? | without           | OK
0.26.0? | with              | strange peak
0.26.0? | without           | OK