News: gpvdm - A general-purpose photovoltaic device model

News, Recent releases of gpvdm:

Faster mobile ions solving for perovskites and bugs in perovskite code fixed, much better now.
Fixes for thermal model
Rewritten doping editor to use Na/Nd
New place for thermal materials in the materials data base
json.inp now called sim.json
Position of windows now stored in json file in gpvdm_local
Layer widget rewritten so copy and paste works
Copy and paste in table widgets works
Accelerated OpenGL rendering - lots of work in this version on making the OpenGL back end better
Objects can be moved with the mouse and they go in the correct direction in 3d
Objects can be rotated with the mouse
More complex optical sources now allowed, and mixing of optical sources
Orthographic projection view added.
Light sources can be moved with the mouse
Free objects which are not attached to the device are now allowed
Back end code for optical detectors added but not fully working
Bugs fixed in oled example
New ray trace example
Micro lens demo works better

Gnuplot fitting bug in windows fixed.
Icon bug fixed.
Theta SRH from Kaienburg, Phys. Rev. App. 6, 024001 (2016), 10.1103/PhysRevApplied.6.024001 now spat out in sim_info.dat
When Voc is not reached in JV simulation, other parameters relying on it are set to -1, i.e. FF.
J_photo and I_photo added to the output of sim_info.dat.
Faster 2D plotting in the optical window.

Been through the code with valgrind (again) to pick up memory issues
Json now dumped in compressed format when fitting (faster less disk io)
Json files now have no limit on size.
Improved job list code, it's now a linked list and avoids realloc so that pointers to jobs can be passed around threads with no potential corruption issues.
Deleted lost of inp file code
Dlls now only loaded once while fitting
Fitting code moved out of dlls and into main exe
Can copy and paste fits, and also configurations as json
Optical output now stored in the snapshots directory
Dlls now only loaded once while fitting
Optical model will only become threaded when not solving a 1D structure, thus removing threading overhead.

Rewritten code to remove duplicates from the matrix to use qsort, this is much faster.
Light data is now mostly stored as floats to reduce memory footprint.
Code to generate triangular meshes from images is now in it's own python module so can be used for mass generation of shapes
Most of the input files have been moved from .inp files to .json files, this should help increase backwards compatibility.
Really strange win 10 bug fixes which was producing random crashes due to win 10 refusing to append some files.
Interface doping added.
Shape editor rewritten.
Fitting bugs fixed.

Lots of bug fixes.
More stable time domain code for starting at high light intensities.
A new backup feature which allows the user to make checkpoints of simulations, then revert to them with one click of the mouse, very useful when trying to understand how a device works.
Bug fixes in the time domain code and interface.
Fixes to the analytical DoS window.
3D ray tracing of light escaping complex 3D structures.
Calculation of XYZ, xyz and RGB color as a function of angle and wavelength for OLED simulation.
A new shapes database, so shapes are loaded from files not defined in code.
The ability to import microscope images and turn them into 3D height profiles.
Faster ray tracing.
Better backwards comparability with older versions.
Faster OpenGL rendering.
Faster OpenGL scenes are now built in memory before rendering enabling draggable objects.
Simulation of micro-lenses on the surfae of OLEDs and solar cells.
Addition of emission materials database.
Multiple light emitting layers within one device for simulation of white OLEDs.
Rewrite of file watch system for stability and speed.
gpvdm_core can now calculate the observed color of a wavelength spectrum using CIE color spectra.
Demo files for optical filters, OLEDs, micro lenses, light escape from flat surfaces, and AFM images.
Bigger new simulation window to showcase the new examples.

More warnings added to fitting code.
It will now complain if you try to feed it a non number in a numeric input - implement in core solver so should work for all values.
Doping window will complain if you do not feed it a number.
Rewrite of fitting bounds checking so bounds checking is now stored in the fit vars file.
Simplified fitting interface.
Old P3HT:PCBM results all added to fit window for the P3HT:PCBM device.
Contacts in 2D can now have different charge densities on them and types of charge - useful for OFETS and complex structures.

More warnings added to fitting code.
Potential divide by zero bug fixed in fitting code.
Automatic passivation between OFET contacts with no current escaping the boundary.

This is a really big series updates and introduces major changes to the model.
The charge density on the boundaries of the device is now associated with the contacts, so for multiple contact devices each contact can have it's own charge density.
Rewrite of the main 1D newton solver so that the RHS and LHS can both have potential applied to them when doing transients, this enables perovskite devices to be properly simulated in time domain.
The simulation model drop down box is now a series of icons to make it clear which simulation mode you are in.
Rectangular shapes can now be introduced into the electrical mesh for etching and contacts.
Significant bug fixed in the 2D OFET solver, which meant sometimes the current was applied to the wrong contact.
Re written the defect state model, so that deep trap states can now be included in the DoS.
The materials database now has over 2,000 materials within it, which are updated from the gpvdm server.
Rewrite of the opengl interface so that the xyz axis is the same projection as the material coordinates.
The device length/width limit of 10cm has been removed so that large cells can be simulated.
Added gui to enable easy editing of objects inserted into the electrical epitaxy.
Improved importing function.
Steady state PL spectra calculated from first principles.

Total rewrite of the Impedance Spectroscopy back end - it's still a work in progress but will improve over time
Fitting code added, this was quite a major undertaking as it needed a simplex downhill algorithm to be written from scratch.
Started rewriting the IMPS back end.
Bug with the progress bar for downloading updates fixed.
Fixed bug with materials database download.