VanitySearch
VanitySearch copied to clipboard
JeanLucPons
Question, about privatekey generation
Hi there! First of all, thanks for your awesome work!!! Already asked it at the bitcrack repo. But would it be possible to use the hex-privatey as Input , treat it as bytes, take one round of sha256, and take the sha-string for the privatekey generations? Keep incrementing the Byte string, and before the treads generate the addresse take one round of sha256 for every worker of the batch?
Thanks in advance
Hi,
But would it be possible to use the hex-privatey as Input , treat it as bytes, take one round of sha256, and take the sha-string for the privatekey generations?
For the first base private key, VanitySearch takes sha256(pbkdf2_hmac_sha512(seed)). You want to enter a 256bit number as input, take sha256(number) as use it as private base key ?
Keep incrementing the Byte string, and before the treads generate the addresse take one round of sha256 for every worker of the batch?
You want to sha the private key before starting each group ? If yes, It will decrease much the performance as an ec multiplication will be done for each GPU thread before launching a GPU kernel.
Yes, that's what i'm looking for. But the starting key should be used as it is, not always 64 characters. But will the group produce a consecutive array?
Lets say my GPU can generate 2M keys a second: echo -n 1 | sha256sum 6b86b273ff34fce19d6b804eff5a3f5747ada4eaa22f1d49c01e52ddb7875b4b
In general it should keep increasing the 1 and should sha256 every number up to 2M 1 sha256> 6b86b273ff34fce19d6b804eff5a3f5747ada4eaa22f1d49c01e52ddb7875b4b 2 sha256> d4735e3a265e16eee03f59718b9b5d03019c07d8b6c51f90da3a666eec13ab35 3 sha256> 4e07408562bedb8b60ce05c1decfe3ad16b72230967de01f640b7e4729b49fce ...
The group perform 1024 consecutive keys + (5*1024) keys coming from symmetries and endomorphisms. If I understand well what you would like to do is to start from a number, sha it, use it as a private key and compute the address then increment the number and retsart ?
Easy mods, unfortunately no. This will require a new GPU kernel. The present GPU code is optimized to perform what it is doing a it is not able to perform an ec multiplication, it performs only ec addition. However, such a kernel should not be too difficult to implement but don't expect a very high key rate (ec mult is slow). I'm currently working on a new kernel optimized for compressed addresses only. I'll try to add this in a future release.
Hello, Not yet, I don't have much free time at the moment. First I will increase the generator table size in order to speed up the EC multiplication. Then I can implement what you want but this will be a different program (a second target in the makefile) because the GPU kernel will be very different.
Hello Jean-Luc, I know you've been busy the last weeks, but did you make any progress?
Kind regards
Not yet, I'm ending the wildcard search, still few problems then I hope I can start a new project.
Hi Jean-Luc, great you're back! Hope you enjoyed your holidays. Do you think, you find some spare time, in the near Future?
Thx in advance
Hi Jean-Luc, any news? Maybe you found eine time to add that feature?
Kind regards