Assignment for the Coursera course CUDA at Scale for the Enterprise.

Apply box filters to all images in a folder, using NPP's nppiFilterBoxBorder_8u_C1R routine. It's heavily based on nvidia's sample code https://github.com/NVIDIA/cuda-samples/tree/master/Samples/4_CUDA_Libraries/boxFilterNPP

boxFilterNPP.cpp is the main implementation, it loops through all files in the input folder, reads the image, copies to device, applies the filter, and writes the result to the output folder.

Common contains the helper code for error checking and image file handling.

data contains example input images.

output contains output for the example inputs.

Makefile for compiling the code.

run.sh contains scripts for building the code and running on example inputs.

Run run.sh.

Use the -input command line flag to specify the input folder to use.

The code is tested on ubuntu 20.04 and nvcc 11.6.