There is a file of 10 million hashes (~250 MB).
The algorithm generates a sequence of symbols with a total of 2^64. Each generated sequence is passed through a hash function and compared with the list of hashes from the file. I’m having trouble loading a hash file into GPU memory because the kernel doesn’t allow double pointers (__global char** strings).
Is it possible to load this list into GPU memory outside the kernel and then access it in the kernel?
What is the best way to implement this task as a whole?