02-22-2022, 09:06 PM
i found the solution to this. The optimized kernel stops at the second last step of sha256. The final step is to add the initial values of the start of the message shedule. If you add the contant value SHA256M_A to a (SHA256M_A + a) you will get the right value.