Hashcat slow. I am generating 400 trillions/s passwords

[Caster]

I created short CUDA program which is now generating all combination of 8 characters password (lower/upper case + digits i.e. 200 trillions of combination) with the speed 400 trillions/s on NVIDIA GeForce GTX 960M.

I am working to add PBKDF2_HMAC_SHA1 (4096 iterations - empty loop included in the program above already) and HMAC(sha1()) to crack the WiFi 22000 hash WPA password in couple of seconds.