Using python or any high level language to do heavy computations is trouble .. cracking is better coded with c/cpp/golang/java .. but thank you a lot for sharing your code

If 3 RTX 3090 can finish 12 digits in an hour, such range and power against it would narrow down the entire range search time to 15 years:

16^4 = 65536 ranges to search
This is if you only pick 2 or 3 (the first significant bit)
Then if you choose to search both:
65536 x 2 = 131072 ranges to search
As each range would take one hour to finish .. we now need 131072 hours to search the entire range
131072 ÷ 24 hours = 5461 days
5461 ÷ 365 days = 14.9 years

Like the Large Bitcoin Collider .. still 16 billion keys/s seem too small for a collab work .. need many more people to join .. problem with such method is it has diminishing return:

- Too few ppl in the collab would mean collab is not worth it compared to solo.
- Many ppl in the collab would mean your cut of the prize is very minimal compared to the resources/power you've contributed to that collab search .. the more ppl joining, the more the chance of solving the key but at the same time the less each one receives as a reward.

 i think I'm gonna stick to Solo searching .. who knows

The difficulty for finding a public key that opens puzzle #66 is as difficult as trying to find the public key of a bitcoin's dead address .. or any other ordinary address  .. i e: 256 bits difficulty .. so in this case brute forcing the entire 66 bit range is more likely achievable than trying to search for an unknown public key

Your second question reminds me of the doubt we heard earlier about puzzle #64, where it took too long for ppl to find the key that some even started to doubt the range entirely is wrong or intentionally misleading by the puzzle creator. Ppl tend to forget the fact that 64 bits, and even worse, puzzle #66 are too huge that it could normally take decades to find the keys within their ranges ..To give you a feel of how big the range is, a gpu with a power of 500 million keys/seconds will finish the entire 66 bits range in 2339 Years  

 if you ask me where the key might be located, my one and only answer is: it can be anywhere within 20000000000000000  -  3ffffffffffffffff

Meaning there is no clues here .. no narrowing down options .. no predictions .. no pattern .. we have to randomly search our way into this range until one lucky bastard among us eventually finds it.



Hope that answers your questions

They're relentless already. Check out NVIDIA's "Hybrid Quantum-Classical Computing Platform" .. and "cuQuantum SDK"

I wish we would have any quantum computing students or gurus here to discuss how many Qubits we need to break as much as 66 bits .. I'm guessing around 1000-3000 at most .. Quantum computing is amazingly useful when it comes to cryptography.

Yep, and it's not so far fetched either when searching with less power .. i remember someone solving an ethereum block with a very tiny mining rig .. solo

No disagreement. But if i could offer any advice, with only 600 mil/s , it's better after dividing the big range into smaller ones, to search within those subranges "randomly". That's the advice i would tell anyone including myself. Going through subranges of 66 bits sequentially is gonna take longer to reach target address than it would take quantum computers to reach 10 million Qubits.. Meaning you'll most likely still not find anything before a teen with a tiny knowledge of Python rents a cloud quantum computer in the future and still finds it before you do

Yes and No ..

Yes, it does inspire to find solutions. After all, that's how humans evolved. Being Incredibly curious is what gave us the competitive advantage over all other creatures. However, it's very important to know what you're dealing with before raising hopes high. Puzzle #66 is 36+ million trillion private keys.. and this means you're gonna encounter tons and tons of prefixes and only one of them is the right one. Trying to make a statistical analysis out of these is like trying to find a similar pattern across the whole universe. You are looking for only one Earth among too many Goldilock zone Earths in the Cosmos. There will be no pattern leading you to it. You just have to brute force your way through planets until you find that exact one Earth. No mathematical formula will get you there. Sure it can lead you to many Earth-like planets, but it doesn't find your exact desired Earth.

That being said, we don't even have to find a pattern in order to reach the private key, we only have to keep searching long enough until we find it, just like #64 got solved by keeping at it. In fact, it blew me away that within the exact month i thought we were never gonna solve it, someone solves it and proves me wrong. Nothing is impossible when it comes to luck and randomness. But then again, Luck and randomness do not get predicted, and that's why they're the only way to beat huge numbers. I'm so grateful they exist

And the main feature of sha256 hex pvt key that makes it secure in the first place is, once you change even one letter in (input) string, the entire hex output changes! This works both way, if you change the output (the sha256 32 byte) , the input will be hugely different. And to make things even harder, that's just one hash function, now let's continue the process of generating the address: we're doing ripemd of the sha256 of the resulting pub key to get the address .. By design, this means that whatever comes out as an address will change significantly after all that hashing we went through. This means that you can have trillions and trillions of addresses starting with 13zb1hQ.... Spread around the entire 256 bit range leaving you with the fact that none of them is the exact address you're looking for. Someone then might ask then why would we search by prefix instead of addresses?! The answer is simple: it's way faster to look for part of the address. That's the only advantage. But It doesn't say anything about determining the range.

Tested this today and I just wanna thank WanderingPhilosopher for such a beautiful software .. this by all means is the best way to search within a puzzle's range .. and as a tweak of both bitcrack and vanitysearch, it has beaten both of them.. faster than all I've tried before .. also very easy to use and was the best randomization option to finally land on .. great job

32 years .. that's why i prefer random mode .. coz if you get a little bit lucky, you can squeeze this down to a few months to a few years.. you never know

I honestly don't think this is necessarily the case, i mean if you had 1000 people looking in micro random 64 bit ranges, inevitably, one of them is gonna land on the right range and solve the key.. Although no one really knows if that was what happened, but you have to know that the bigger the range, the more it's gonna be about luck than about computing power .. you need couple hundred Telsa 100 to solve puzzle #120 within two months .  If such computing power was available to a single person already, he would have sold more than just #120 .. he would have used them to crack open small range puzzles as well .. so i guess those who have huge computing power are busy doing something else making them way more money than our beloved puzzle would .. Lucky for us though, otherwise we stand no chance with our teeny tiny personal rigs

clbitcrack behaves weird on old and low gig cards .. best case scenario is it's gonna search but not find anything when testing a known key in a known range (i.e fail test)
Here is test results:

Rx 5xxx XT - work and pass test
Rx 570 8G - search but fail test
Rx 580 8G - search but fail test
Anything less - do not search

In the NVIDIA family i found anything less than 1xxx 8G to not work or behave weirdly ( yes you can use your NVIDIA card with clbitcrack .. in my case cuda has a config problem that im not so fond of spending time on fixing right now)


Your results may vary.

Could be anything .. with a range 4 times bigger than 64 bits, the only resort is random search .. range search is gonna eat up time and resources .. the only way to beat 17 random characters is by utilizing randomness itself .. but that's just me.

P65 = 1A838B13505B26867

P66 = 67 + 67 = CE

Out of range .. sorry but doesn't work.

Thanks for saving my time .. was gonna run my pub key solver to get the pvt key .. it kills me though to find out that it was in the F range .. coz i was gonna search this one solely on all cards but ditched the idea after a while coz it seemed radical .. after all, how come ignore 7 other ranges just coz you had a "hunch" that it's in the last range .. i guess trusting your instinct could pay after all

Are you sure about that, Alberto? (Cough BSGS cough) sorry i guess i got the flu 😃

Puzzle #64 just got solved .. now #66 is the next target .. very VERY hard target that puzzle #120 would look like a picnic compared to it

I would be comfortable to say it will at least be in the 254 to 256 bit range 99% of the time

You would have to intentionally create an address in the weak 160 bit range or below to be able to receive such weak address .. fun fact is though, even the strongest private key (256 bit) will generate an address that can be opened by another private key in the first 160 bit range .. good luck finding one of those copycats though