~
can somebody implement this please i have millions of private keys so it cannot be converted quickly to addresses with cpu because cpu is not powerful enough can somebody make small modification to the software where it read single hex private key to read those values from file this way we can convert millions of private keys using gpu thanks again for anyone out there who can help.
How exactly many millions are we talking about? Because GPUs cannot do disk I/O, so only the actual conversion of hex to WIF can be accelerated with CUDA. The reads and writes still have to be done on CPU, and then all that has to be copied to the GPU memory and then back again, because CUDA cannot read memory in RAM (unless you use their new feature called Unified Memory but honestly I don't know how fast that is), so what I'm trying to tell you is it might be better implementing this on CPU using multiple threads and the AVX instruction set and that should the job. The CPU can already do much more complicated address generation at millions of keys per second anyway.
Because none have made it as you said the conversation is possible with gpu and at the time of saving you can use cpu i think the program already does it can convert millions keys per second and cpu can save it in a minute this way is better then all the way to cpu