Changes

The following table shows the timings (in milliseconds) for various image sizes for the serialized and parallelized versions of the program. The parallelized version of the code is faster than the serial version, however, '''cudaMalloc ''' has added overhead.