Quote from: student4ever on June 24, 2020, 07:14:19 PM

Quote from: Jean_Luc on June 24, 2020, 02:51:27 PM

@student
If you are looking for full addresses, do not enter in the list a part of them (for instance the 15 first char)
It will be faster to enter full addresses.
That's true that there is 2^96 priv key for an address and if you perform 2^96 key you will find a collision but that would also imply that you search for the whole address range, on the blockchain there is no 2^160 address with fund and it is not possible to have a file of 2^160 address.
So if you want to find a collision between 2 priv key that give the same address, consider using BTCCollider which would require "only" 2^80 operations to reach the collision.
https://github.com/JeanLucPons/BTCCollider

ok. Thank you. I tried it with 1000 full adresses, but the gpu shows me that it is going up to 80% and down to 0%, so the speed decrease is still there, but a bit flatter.
So far I have a file of ~ 10M adresses. It is just to increase the possibility that I will find a collusion. The file has ~350 MB. And that is enough to make it possible to find something in reasonable time I think. At least if you can increase the speed and I will get more rigs to work

Can you make the speed of the software the same, so it doesnt matter if you are looking for 1 or 10M adresses? Or is that impossible?

How did you get 10M addresses? I assume there might be a script that extracts addresses from the blockchain somehow, is it?