BitCrack - A tool for brute-forcing private keys

[BitCrack]

Hi all,

I've been working on a tool for brute-forcing Bitcoin private keys. The main purpose of this tool is to contribute to the effort of solving the Bitcoin puzzle transactions: https://blockchain.info/tx/08389f34c98c606322740c0be6a7125d9860bb8d5cb182c02f98461e5fa6cd15

Screenshot:



It is open-source under the MIT licence and requires no external dependencies other than the CUDA toolkit. It builds on Windows using Visual Studio 2015, and Linux using Make (you might have to edit the Makefile and point it towards your CUDA toolkit directory).

It can search for compressed/uncompressed keys or both.

The performance is good, but can likely be improved. On my hardware (GeForce GT 640) it gets 9.4 million keys per second compressed, 7.3 million uncompressed.

Note:
-Currently it is CUDA only.
-It can only search one target key at a time


Features I would like to add if there is enough interest for the project:

-Support for searching multiple target keys at one time Done
-OpenCL/AMD device support
-CPU with AVX/AVX2/SHA support
-Checkpoints/Stop and resume
-Vanity address generation


Source and Win32/Win64 binaries available here:
https://github.com/brichard19/BitCrack
https://github.com/brichard19/BitCrack/releases/tag/v0.0.6


Thoughts?


Thanks!

[bitmover]

Hi all,

I've been working on a tool for brute-forcing Bitcoin private keys.

.....

The performance is good, but can likely be improved. On my hardware (GeForce GT 640) it gets 9.4 million keys per second compressed, 7.3 million uncompressed.

....

Thoughts?


Thanks!

what I am going to say is purely theoretical, as I have never tried to brute force private keys.

I don´t think you will have any success doing this.
If you have enough hash power, you would probably make more money working as an honest node, mining.

Bitcoin protocol is secure enough. If it were brute force-able, bitcoin wouldn't be worth anything.

What i think you could do is to find a collision. If someone made an insecure private key, by just hashing some passphrase... maybe you could have some luck. I tried to do this once, but never found anything. I tried those terms like "wallet" "bitcoin". I found some addresses which used to have balances, but not anymore. But I was doing manually, like 1 address/3 minutes lol

If you could somehow at the same time check if any of those 9.4 million keys/sec have any balance in btc/bch/btg etc...

But I don't think any of that will work, and will be a waste of time and resources.

[HCP]

what I am going to say is purely theoretical, as I have never tried to brute force private keys.

I don´t think you will have any success doing this.

I think you misunderstand what it is that the OP is attempting to achieve...

What i think you could do is to find a collision. If someone made an insecure private key, by just hashing some passphrase... maybe you could have some luck.

That is exactly what the OP is trying to achieve:

The main purpose of this tool is to contribute to the effort of solving the Bitcoin puzzle transactions: https://blockchain.info/tx/08389f34c98c606322740c0be6a7125d9860bb8d5cb182c02f98461e5fa6cd15

The Bitcoin puzzle transactions were setup as an exercise to see how tough it is bruteforce BTC... and is essentially what the LBC is attempting to cash in on (aside from other "collateral damage" of people with weak keys).

What is interesting is that the addresses in the middle range of amounts (0.056 to 0.16 BTC), which were theoretically "easier" than the addresses with the larger amounts, have not yet been "cracked"??!?

[arulbero]

I've been working on a tool for brute-forcing Bitcoin private keys. The main purpose of this tool is to contribute to the effort of solving the Bitcoin puzzle transactions: https://blockchain.info/tx/08389f34c98c606322740c0be6a7125d9860bb8d5cb182c02f98461e5fa6cd15

Other (more or less) similar projects :

https://bitcointalk.org/index.php?topic=1877935.0

https://bitcointalk.org/index.php?topic=3102823.0

The performance is good, but can likely be improved. On my hardware (GeForce GT 640) it gets 9.4 million keys per second compressed, 7.3 million uncompressed.

Try with other software to have a comparison.

[Thirdspace]

https://bitcointalk.org/index.php?topic=1877935.0

I've been working on a tool for brute-forcing Bitcoin private keys. The main purpose of this tool is to contribute to the effort of solving the Bitcoin puzzle transactions:

that's the main group effort for this btc puzzle, however the program is closed-source (cmiiw)
OP is trying to do the same effort, but without collective efforts this will be redundant

The Bitcoin puzzle transactions were setup as an exercise to see how tough it is bruteforce BTC... and is essentially what the LBC is attempting to cash in on (aside from other "collateral damage" of people with weak keys).

What is interesting is that the addresses in the middle range of amounts (0.056 to 0.16 BTC), which were theoretically "easier" than the addresses with the larger amounts, have not yet been "cracked"??!?

I think the higher amounts that already spent was not 'cracked' but moved by the creator
and I just noticed 1LzhS3k3e9Ub8i2W1V8xQFdB8n2MYCHPCa has just been found recently! on May 29th
However LBC site still show the task for finding its key
Did someone else find it? or was LBC site just slow on updating its stats page?

Time to hit #55 of the puzzle transaction (1LzhS3k3e9Ub8i2W1V8xQFdB8n2MYCHPCa) 0 to 112 days at current speed in sub-55bit search space.

[Gerci]

Hi all,

I've been working on a tool for brute-forcing Bitcoin private keys. The main purpose of this tool is to contribute to the effort of solving the Bitcoin puzzle transactions: https://blockchain.info/tx/08389f34c98c606322740c0be6a7125d9860bb8d5cb182c02f98461e5fa6cd15

Screenshot:



It is open-source under the MIT licence and requires no external dependencies other than the CUDA toolkit. It builds on Windows using Visual Studio 2015, and Linux using Make (you might have to edit the Makefile and point it towards your CUDA toolkit directory).

It can search for compressed/uncompressed keys or both.

The performance is good, but can likely be improved. On my hardware (GeForce GT 640) it gets 9.4 million keys per second compressed, 7.3 million uncompressed.

Note:
-Currently it is CUDA only.
-It can only search one target key at a time


Features I would like to add if there is enough interest for the project:

-Support for searching multiple target keys at one time
-OpenCL/AMD device support
-CPU with AVX/AVX2/SHA support
-Checkpoints/Stop and resume
-Vanity address generation


Source and Win32/Win64 binaries available here:
https://github.com/brichard19/BitCrack
https://github.com/brichard19/BitCrack/releases/tag/v0.0.6


Thoughts?


Thanks!

A set of tools for Bitcoin private keys that destroy. Currently the project requires a CUDA GPU. The main purpose of this project is to contribute to solving Bitcoin puzzle transactions: A transaction with 32 addresses becomes increasingly difficult to solve.

Currently this project is only CUDA, but I would like to bring it to another architecture if there is enough interest in the project.

[arulbero]

I just noticed 1LzhS3k3e9Ub8i2W1V8xQFdB8n2MYCHPCa has just been found recently! on May 29th
However LBC site still show the task for finding its key
Did someone else find it? or was LBC site just slow on updating its stats page?

LBC will search any key in sub-55 bit space, its main goal is not related to the puzzle transaction.

[Thirdspace]

LBC will search any key in sub-55 bit space, its main goal is not related to the puzzle transaction.

on stats page, they put up a notice "Time to hit #55 of the puzzle transaction (1LzhS3k3e9Ub8i2W1V8xQFdB8n2MYCHPCa)"
so I thought they're aiming for those addresses listed in that puzzle transaction
https://blockchain.info/tx/08389f34c98c606322740c0be6a7125d9860bb8d5cb182c02f98461e5fa6cd15

@OP can your program work collectively in a pool?
running it individually without any means of cooperative works would be somewhat useless and waste of resources

[monkeydominicorobin]

Hi all,

I've been working on a tool for brute-forcing Bitcoin private keys. The main purpose of this tool is to contribute to the effort of solving the Bitcoin puzzle transactions: https://blockchain.info/tx/08389f34c98c606322740c0be6a7125d9860bb8d5cb182c02f98461e5fa6cd15

Screenshot:



It is open-source under the MIT licence and requires no external dependencies other than the CUDA toolkit. It builds on Windows using Visual Studio 2015, and Linux using Make (you might have to edit the Makefile and point it towards your CUDA toolkit directory).

It can search for compressed/uncompressed keys or both.

The performance is good, but can likely be improved. On my hardware (GeForce GT 640) it gets 9.4 million keys per second compressed, 7.3 million uncompressed.

Note:
-Currently it is CUDA only.
-It can only search one target key at a time


Features I would like to add if there is enough interest for the project:

-Support for searching multiple target keys at one time
-OpenCL/AMD device support
-CPU with AVX/AVX2/SHA support
-Checkpoints/Stop and resume
-Vanity address generation


Source and Win32/Win64 binaries available here:
https://github.com/brichard19/BitCrack
https://github.com/brichard19/BitCrack/releases/tag/v0.0.6


Thoughts?


Thanks!

Assuming that you are successful. What is the use of cracking keys when you are not scanning for balances. This is just another BIP39 social engineering clone tool. Yes BIP39 is a cool improvement but don't use it to fool the non-programmers.

[BitCrack]

Other (more or less) similar projects :

https://bitcointalk.org/index.php?topic=1877935.0

https://bitcointalk.org/index.php?topic=3102823.0

The performance is good, but can likely be improved. On my hardware (GeForce GT 640) it gets 9.4 million keys per second compressed, 7.3 million uncompressed.

Try with other software to have a comparison.

I'm not touching LBC, as the client source code is not available.


My machine is getting ~1 million keys per second with oclVanityGen. Not sure if it's not using my GPU or not picking optimal parameters.

Assuming that you are successful. What is the use of cracking keys when you are not scanning for balances. This is just another BIP39 social engineering clone tool. Yes BIP39 is a cool improvement but don't use it to fool the non-programmers.

I imagine people would check the balance for an address before trying to crack it. But like I said, it's main purpose is for the puzzle addresses, which have balances.

It's not social engineering anything. It's simply a brute-force tool.

[Coin-1]

My machine is getting ~1 million keys per second with oclVanityGen. Not sure if it's not using my GPU or not picking optimal parameters.

Yes, your program might be useful for creating a "beautiful" Bitcoin-addresses using VanityGen, but only for that.

It's not social engineering anything. It's simply a brute-force tool.

Please don't waste your time and resources to brute force a Bitcoin private key. Let's calculate how much time you need to crack one Bitcoin-address on your machine.

You said that your performance is 9 million BTC-addresses per second, i.e. approximately 2²³ BTC-addresses per second. Thus the brute forcing will take 2^160-23 = 2¹³⁷ seconds! I guess it is more than septillion (10²⁴) years!

Note that this calculating is very approximate, because actually the private key length is 2²⁵⁶, but I think it is enough to stop making an attempts to crack Bitcoin-addresses.

[monkeydominicorobin]

Other (more or less) similar projects :

https://bitcointalk.org/index.php?topic=1877935.0

https://bitcointalk.org/index.php?topic=3102823.0

The performance is good, but can likely be improved. On my hardware (GeForce GT 640) it gets 9.4 million keys per second compressed, 7.3 million uncompressed.

Try with other software to have a comparison.

I'm not touching LBC, as the client source code is not available.


My machine is getting ~1 million keys per second with oclVanityGen. Not sure if it's not using my GPU or not picking optimal parameters.

Assuming that you are successful. What is the use of cracking keys when you are not scanning for balances. This is just another BIP39 social engineering clone tool. Yes BIP39 is a cool improvement but don't use it to fool the non-programmers.

I imagine people would check the balance for an address before trying to crack it. But like I said, it's main purpose is for the puzzle addresses, which have balances.

It's not social engineering anything. It's simply a brute-force tool.

Once you try to get people involved in a crusade like yours. It is social engineering. Your asking them to use your tool which is wasting computing resources. How can you brute force something that has combinations that almost rivals the number of atoms in our universe which is sha256.

[MrCrank]

This is good idea.
Do you think is it possible?
For this need many years.. or create worldwide GPU online pool

[BitCrack]

Good project idea.

How to deploy this to 100 CUDA cards?

thanks

It doesn't have distributed processing right now but it's on the to-do list.

For now you'll need to run it for each GPU on each machine. It shouldn't be too difficult if you have SSH access to your machines.

[butka]

The main purpose of this tool is to contribute to the effort of solving the Bitcoin puzzle transactions: https://blockchain.info/tx/08389f34c98c606322740c0be6a7125d9860bb8d5cb182c02f98461e5fa6cd15

For all those (like me) wondering what the heck these puzzle transactions are, here is a link to an older tread trying to explain that:

https://bitcointalk.org/index.php?topic=1306983.0

Now at least you would have a clue what the OP is trying to achieve.

[Kakmakr]

Is there a way that this program can be changed to query every private key to see if it contains any bitcoins, instead of targeting specific predefined private keys?

I think the probability to find a private key that contains some coins are more likely than brute forcing a specific private key?

Looking at the odds, it might be easier to just save up the electricity cost to do this over years, than having enough luck to get 1 private key that contains some coins. 

[butka]

Hi all,

I've been working on a tool for brute-forcing Bitcoin private keys. The main purpose of this tool is to contribute to the effort of solving the Bitcoin puzzle transactions: https://blockchain.info/tx/08389f34c98c606322740c0be6a7125d9860bb8d5cb182c02f98461e5fa6cd15

Does your tool include the increasing bits of entropy with each consecutive address? This pattern was indicated before:

I see a pattern but I do not think it will help much.

My conjecture is that the sequence number (the BTC amount) simply states the number of random bits in the private key.

Check it out:

The 0.001 BTC output used a 1 bit key (already claimed obviously)

The 0.002 BTC output used a 2 bit key

The 0.003 BTC output used a 3 bit key

etc...

The 0.019 BTC output used a 19 bit key

So I would claim the next unclaimed output uses a 51 bit key.

Interesting.

Here's a thought.  What if that is the ONLY pattern.

Consider the possibility that the person that created this "puzzle" has done so as a form of "canary in the coal mine".

The idea would be to generate a bitcoin address completely randomly, but which only uses X bits of entropy.  Then progressively store large values of bitcoins for each increase in entropy to provide an incentive for someone to try and access it.

So:

• The only possible address with 1 bit of entropy is the address associated with the private key having binary value 1
• There are 3 possible addresses with 2 bits of entropy.  The addresses associated with the private keys having each of the following binary values (01, 10, 11), and one of them was chosen at random (11)
• There are 7 possible addresses with 3 bits of entropy, The addresses associated with the private keys having each of the following values (001, 010, 011, 100, 101, 110,  and 111), and one of them was chosen at random (111)
• There are 15 possible addresses with 4 bits of entropy, and one of them was chosen at random (1000)
• There are 31 possible addresses with 5 bits of entropy, and one of them was chosen at random (10101)
• There are 63 possible addresses with 6 bits of entropy, and one of them was chosen at random (110001)
• There are 127 possible addresses with 7 bits of entropy, and one of them was chosen at random (1001100)
• and so on

By looking at which addresses have been "claimed", the creator of the "puzzle" can get a feel for the current security of version 1 bitcoin addresses.

So, if there are 50 addresses that have been "claimed", perhaps they'll avoid using any bitcoin address that has a private key with less than 51 binary digits for storing large amounts (actually, I'd probably go a bit higher than that just to be safe), and would have a feel for how secure addresses with 160 or more bits of entropy would be.

If that is what's going on here, then there may not be a "formula" or "pattern" that can be used.  To get the 51st address, you'd just have to brute force try all addresses associated with private keys between 1125899906842624 and 2251799813685248 (which is only about 1.126 X 10¹⁵ addresses).

[holy_ship]

Hey! Still no benchmarks on good cards?

[giftseller8x]

Hey! Still no benchmarks on good cards?

I have try with my 1060

Code:

D:\C\BitCrack>BitCrack64 1FshYsUh3mqgsG29XpZ23eLjWV8Ur3VwH
[2018-09-14.17:24:58] [Info] Compression: compressed
[2018-09-14.17:24:58] [Info] Starting at: 0000000000000000000000000000000000000000000000000000000000000001
[2018-09-14.17:24:58] [Info] Initializing GeForce GTX 1060 3GB
[2018-09-14.17:25:07] [Info] Generating 1,179,648 starting points (45.0MB)
[2018-09-14.17:25:08] [Info] 10.0%
[2018-09-14.17:25:08] [Info] 20.0%
[2018-09-14.17:25:08] [Info] 30.0%
[2018-09-14.17:25:08] [Info] 40.0%
[2018-09-14.17:25:09] [Info] 50.0%
[2018-09-14.17:25:09] [Info] 60.0%
[2018-09-14.17:25:09] [Info] 70.0%
[2018-09-14.17:25:09] [Info] 80.0%
[2018-09-14.17:25:10] [Info] 90.0%
[2018-09-14.17:25:10] [Info] 100.0%
[2018-09-14.17:25:10] [Info] Done
GeForce GTX 1060 381/3072MB | 1 target 60.61 MKey/s (1,875,640,320 total) [00:00:29]^C
D:\C\BitCrack>BitCrack64 -c -u 1FshYsUh3mqgsG29XpZ23eLjWV8Ur3VwH
[2018-09-14.17:25:48] [Info] Compression: both
[2018-09-14.17:25:48] [Info] Starting at: 0000000000000000000000000000000000000000000000000000000000000001
[2018-09-14.17:25:48] [Info] Initializing GeForce GTX 1060 3GB
[2018-09-14.17:25:48] [Info] Generating 1,179,648 starting points (45.0MB)
[2018-09-14.17:25:48] [Info] 10.0%
[2018-09-14.17:25:48] [Info] 20.0%
[2018-09-14.17:25:49] [Info] 30.0%
[2018-09-14.17:25:49] [Info] 40.0%
[2018-09-14.17:25:49] [Info] 50.0%
[2018-09-14.17:25:49] [Info] 60.0%
[2018-09-14.17:25:50] [Info] 70.0%
[2018-09-14.17:25:50] [Info] 80.0%
[2018-09-14.17:25:50] [Info] 90.0%
[2018-09-14.17:25:50] [Info] 100.0%
[2018-09-14.17:25:50] [Info] Done
GeForce GTX 1060 381/3072MB | 1 target 46.93 MKey/s (685,375,488 total) [00:00:12]

This result give the same speed of Vanitigen, but it not support multi GPU at now.

https://prnt.sc/kubfus

[giftseller8x]

Have a multi GPU version running 170mKeys/s on RX570 or 150mKeys/s on GTX 1060, a rig with 6 runs just above a billion/sec.

Windows 10

Your speed is double compare with mine.1060 3G win 7, 16G ram, core i5 ivy.