adyson / cltem Goto Github PK

A tab to show the loaded structure so areas can be selected to perform STEM simulation on.

Dose simulation seems ~1000 times less than the input value.

Experimental image 40000 e/A^2

CUDA simulation

clTEM - 40000000 e/A^2

Set some fairly standard input values, to help user judge what are 'sensible' values.

e.g. using a simulation area of 1000x1000 on an .xyz file of approximately 30x30 will cause crashes (100x100 works).

So that STEM/TEM can easily be compared as they use the same reference coordinates

Current implementation is very basic. Improvements could be made by checking for any 'default names' used and trying to add the next one. Possible to have some default angle values too, though what to set them as would mostly be arbitrary.

The AMD FFT library does not appear to work for Nvidia at resolutions greater than 512.

numbers shown to lots of decimal places.

Currently the GUI does not indicate what units the inputs are in. Causing confusion as units are mixed and in 'non-standard' forms.

Preferably in a way that allows automatic documentation generation.

Either the image data, or the actual image should be autosaved to a temp directory to avoid potential data loss. A new simulation would overwrite the old data.

For STEM, saving images does not work outright (uses simulation resolution, not stem resolution). TEM can break by changing the resolution combo box before saving.

Zero aperture works but is pointless.
Zero voltage doesnt work and shouldn't be available anyway.

Main simulation loop doesn't not complete, all kernels compile fine, need to test output at various stages, could be problem with background threads and OpenCL device or command queue.

Because @ADyson likes his surface pro 2 too much.

Image shows blank at end of TDS run only. Simulate EW button also does not get re-enabled

Deleting multiple detectors can be a pain as the list box needs to be clicked on again.

e.g. tell the user why a certain value has not been accepted as a voltage

Crashes ensue.

Add icon for the program to use in the taskbar etc...

e.g. on surface pro 2 simulation appears to run, but nothing is output.

Using certain range of values causes the program to crash e.g.small aperture sizes. Possible from trying to normalise an array of zeros?

Be able to visualise the detectors on the diffraction images.

Some way to display the pixel coordinate and value is needed to be able to quickly analyse images. E.g. Be able to tell what features would be captured by certain values of STEM detector.

If a configuration is known to match a certain microscope, it should be savable so it can easily be reloaded later.

its in angstroms or reciprocal angstroms but it doesn't show.

GUI is currently non-resizable. Would be easier to see the images if the window could be made larger.

E.g. to allow breaking of large STEM simulations without closing and reopening the program

Either disable image after resolution change, or remember resolution that EW was simulated at and use the old values.

Currently crashes, possible due to trying to use kernels built for previous device. Temporarily disabled device switching after setting device first time.

dialog needs to take an x and y position also.

Some sort of contrast scaling is needed (e.g. log)

Selecting TDS does multiple runs, but the end image looks like the sum of the same image (i.e. no atom displacements)

Image can be dragged out of range and then cannot be retrieved.

Ideally a file that can be loaded, and the simulation then continued from, though a button would be adequate for now.

Add support for using experimental detector maps for STEM.