Steps to use GPU for fusion

[sundar07]

Hello everyone,

We have been collecting a bunch of multiview light-sheet data sets, which we would like to fuse using the GPU option in BigStitcher. For setting up GPU-based analysis, so far we have done the following:

We have installed visual studio community 2022 from this site: https://learn.microsoft.com/en-us/visualstudio/releases/2022/system-requirements While installing, there was a prompt asking us to choose what we want to install. We were not sure what to choose among the given options, but chose a few C++, JAVA ones and installed. Any recommendations?

We have installed Cuda from here: https://docs.nvidia.com/cuda/cuda-installation-guide-microsoft-windows/index.html Installation was straightforward and proceeded as prompted and at the end of cuda installation, it showed that Visual studio was successfully added.

We have downloaded cmake from "Binary distributions" from here: https://cmake.org/download/ Installation was straightforward.

We have downloaded Fourier convolution cuda library from your site: https://github.com/StephanPreibisch/FourierConvolutionCUDALib

But following these, we are a bit stuck. We do not quite understand how to execute the section "Windows Build Instructions with CMake" mentioned in your github page. Can you help with this?

Both Cmake and Cuda (nvidia gpu computing toolkit folder) are present in C:\Program files. Where should the codes downloaded from your github page be placed and how to run the cmake build?

Any help will be much appreciated. Please let us know if you need any further information.

Looking forward to your reply. Thanks Sundar