Bitcoin puzzle transaction ~32 BTC prize to who solves it

[kTimesG]

Anyone know how did they solve the 130 puzzle? What they use? How hard puzzle 135 compared to 130 puzzle?

Go to RCoder thread https://bitcointalk.org/index.php?topic=5517607.0. The difficult P135 is 5 times harder than P130.

135 will, on average, is 2.5 times "harder" than 130.  Meaning it will take 2.5 times more ops, jumps, steps, hashing, computing (whatever you want to call it lol) to find a collision.

135 / 2 = 67.5
130 / 2 = 65

2**2.5 = sqrt(2**5) = sqrt(32) = 5.66 times "harder"

How many gpus needed to solve puzzle 135? Do you have a rough estimate?

probably 500 x RTX 4080 and 6 months

RTX 4080? I'd say you forgot a zero somewhere, for that GPU.

It takes somewhere around 2500 RTX 4090 for 2 months. Or 500 RTX 4090, for 10 months. With a good 50% error margin

[dastic]

Anyone know how did they solve the 130 puzzle? What they use? How hard puzzle 135 compared to 130 puzzle?

Go to RCoder thread https://bitcointalk.org/index.php?topic=5517607.0. The difficult P135 is 5 times harder than P130.

135 will, on average, is 2.5 times "harder" than 130.  Meaning it will take 2.5 times more ops, jumps, steps, hashing, computing (whatever you want to call it lol) to find a collision.

135 / 2 = 67.5
130 / 2 = 65

2**2.5 = sqrt(2**5) = sqrt(32) = 5.66 times "harder"

How many gpus needed to solve puzzle 135? Do you have a rough estimate?

probably 500 x RTX 4080 and 6 months

RTX 4080? I'd say you forgot a zero somewhere, for that GPU.

It takes somewhere around 2500 RTX 4090 for 2 months. Or 500 RTX 4090, for 10 months. With a good 50% error margin

Yeah I meant 4090. The thing is, if RetiredCoder solved 130 - he will probably solve 135 and very soon I think. He can afford a lot of gpus now.

[nomachine]

Yeah, and for puzzle 140, we'll wait for the next 10 years.

[jedi12345]

Hey Puzzlers

What would be the best offline way to test a big list of either Decimal or Binary possibilities I have.

Just CPU testing, I could buy a GPU if its like 100x faster than CPU.

Happy to give the 1-3 people 0.1btc that help me when I solve puzzle 69. (If that's allowed on this forum)

Note offline solving only, no pools or online, happy to download any requirements first. - This pc will not be connected to the internet once ready to solve the problem.

Cheers

[kTimesG]

Priorities of BTC puzzle address solvers, updated.

1. What to do with the money.
2. Showing their skills / Asking for help (case-dependent).
3. Isolating network connectivity of the unicorn CPU that is ready to write Puzzle 69 key in unsafe RAM.
4. slipstream being down, worry about bots.
5. Solving the challenge.

[madogss]

Hey Puzzlers

What would be the best offline way to test a big list of either Decimal or Binary possibilities I have.

Just CPU testing, I could buy a GPU if its like 100x faster than CPU.

Happy to give the 1-3 people 0.1btc that help me when I solve puzzle 69. (If that's allowed on this forum)

Note offline solving only, no pools or online, happy to download any requirements first. - This pc will not be connected to the internet once ready to solve the problem.

Cheers

For cpu you want to use keyhunt or keyhunt-cuda. here's the links https://github.com/WanderingPhilosopher/KeyHuntCudaClient https://github.com/albertobsd/keyhunt

You can also find pre compiled ones for keyhunt on github.

cpu averages about 3.5 million keys a second per thread if using above software or I believe according to ktimesg 10 million keys a second per thread with custom software.

Gpu's are way faster a 3080 does about 2200MK/s or 2.2 billion keys a second.

Best way to test a big list is to write a python script to read the doc and import the possibilities into a list then in a loop convert the decimal or binary into hexadecimal get rid of the 0x in the front and run keyhunt or kehunt-cuda in that range using a subprocess.

[jedi12345]

Hey Puzzlers

What would be the best offline way to test a big list of either Decimal or Binary possibilities I have.

Just CPU testing, I could buy a GPU if its like 100x faster than CPU.

Happy to give the 1-3 people 0.1btc that help me when I solve puzzle 69. (If that's allowed on this forum)

Note offline solving only, no pools or online, happy to download any requirements first. - This pc will not be connected to the internet once ready to solve the problem.

Cheers

For cpu you want to use keyhunt or keyhunt-cuda. here's the links https://github.com/WanderingPhilosopher/KeyHuntCudaClient https://github.com/albertobsd/keyhunt

You can also find pre compiled ones for keyhunt on github.

cpu averages about 3.5 million keys a second per thread if using above software or I believe according to ktimesg 10 million keys a second per thread with custom software.

Gpu's are way faster a 3080 does about 2200MK/s or 2.2 billion keys a second.

Best way to test a big list is to write a python script to read the doc and import the possibilities into a list then in a loop convert the decimal or binary into hexadecimal get rid of the 0x in the front and run keyhunt or kehunt-cuda in that range using a subprocess.

Ok few questions before I get a GPU, because I will have around 200 trillion combinations maybe more.

Is the 4090 or 5090 way faster or not worth the $$ or is the 3090 just the best bang for your buck does it have to be Nvidia as AMD have good price to performance.

Converting into a Hexadecimal without Ox sounds easy enough. Hopefully python can output 100 trillion numbers without issues.

Could you give me a basic run though and how to set up and use keyhunt-cuda on a premade hexadecimal file (using a subprocess). Would it be txt file or a py file or a json file.

Also I take it this computation can be done offline once setup?

Cheers

[JackMazzoni]

Hey Puzzlers

What would be the best offline way to test a big list of either Decimal or Binary possibilities I have.

Just CPU testing, I could buy a GPU if its like 100x faster than CPU.

Happy to give the 1-3 people 0.1btc that help me when I solve puzzle 69. (If that's allowed on this forum)

Note offline solving only, no pools or online, happy to download any requirements first. - This pc will not be connected to the internet once ready to solve the problem.

Cheers

For cpu you want to use keyhunt or keyhunt-cuda. here's the links https://github.com/WanderingPhilosopher/KeyHuntCudaClient https://github.com/albertobsd/keyhunt

You can also find pre compiled ones for keyhunt on github.

cpu averages about 3.5 million keys a second per thread if using above software or I believe according to ktimesg 10 million keys a second per thread with custom software.

Gpu's are way faster a 3080 does about 2200MK/s or 2.2 billion keys a second.

Best way to test a big list is to write a python script to read the doc and import the possibilities into a list then in a loop convert the decimal or binary into hexadecimal get rid of the 0x in the front and run keyhunt or kehunt-cuda in that range using a subprocess.

Ok few questions before I get a GPU, because I will have around 200 trillion combinations maybe more.

Is the 4090 or 5090 way faster or not worth the $$ or is the 3090 just the best bang for your buck does it have to be Nvidia as AMD have good price to performance.

Converting into a Hexadecimal without Ox sounds easy enough. Hopefully python can output 100 trillion numbers without issues.

Could you give me a basic run though and how to set up and use keyhunt-cuda on a premade hexadecimal file (using a subprocess). Would it be txt file or a py file or a json file.

Also I take it this computation can be done offline once setup?

Cheers

Python is slow. C++ is 100 times faster than python. All computation can be done offline.  I think RTX 5070 is the best best bang for the buck. It is as fast as 4090 and almost one fourth the price.

[kTimesG]

Ok few questions before I get a GPU, because I will have around 200 trillion combinations maybe more.

Only 200 trillions in a text file? I think the least of your problems is whether you use a CPU or a GPU.

What exactly do you understand as a "combination"? Private key? Public key? What should be computed? A hash? A public key? Different problems have different solutions.

[madogss]

Ok few questions before I get a GPU, because I will have around 200 trillion combinations maybe more.

Is the 4090 or 5090 way faster or not worth the $$ or is the 3090 just the best bang for your buck does it have to be Nvidia as AMD have good price to performance.

Converting into a Hexadecimal without Ox sounds easy enough. Hopefully python can output 100 trillion numbers without issues.

Could you give me a basic run though and how to set up and use keyhunt-cuda on a premade hexadecimal file (using a subprocess). Would it be txt file or a py file or a json file.

Also I take it this computation can be done offline once setup?

Cheers

With that many combinations it would be faster to write your own software since you would have to start and close the program every time.

What are the combinations are they ranges like 80000:fffff or are they 1 hexadecimal like d2c55?

if they are ranges then you would want to customize keyhunt-cuda so that each thread is of the gpu is searching a custom range.

if they are 1 hex number then a python script would be enough but it would take a long time and if you wanted to do it faster then you would want to write it in C or C++.

I don't know about the specific speeds for the 5090 but in this topic there should be 4090 speeds if I remember correctly it's around 4 to 6 billion keys a second.

[jedi12345]

Ok few questions before I get a GPU, because I will have around 200 trillion combinations maybe more.

Is the 4090 or 5090 way faster or not worth the $$ or is the 3090 just the best bang for your buck does it have to be Nvidia as AMD have good price to performance.

Converting into a Hexadecimal without Ox sounds easy enough. Hopefully python can output 100 trillion numbers without issues.

Could you give me a basic run though and how to set up and use keyhunt-cuda on a premade hexadecimal file (using a subprocess). Would it be txt file or a py file or a json file.

Also I take it this computation can be done offline once setup?

Cheers

With that many combinations it would be faster to write your own software since you would have to start and close the program every time.

What are the combinations are they ranges like 80000:fffff or are they 1 hexadecimal like d2c55?

if they are ranges then you would want to customize keyhunt-cuda so that each thread is of the gpu is searching a custom range.

if they are 1 hex number then a python script would be enough but it would take a long time and if you wanted to do it faster then you would want to write it in C or C++.

I don't know about the specific speeds for the 5090 but in this topic there should be 4090 speeds if I remember correctly it's around 4 to 6 billion keys a second.

Just a big list of hex/binary/decimal numbers no ranges or anything like that, I cant go into anymore detail or will be out of pocket 6.7btc A python script would probably be fine but then that is CPU only then?. But that might work. Has anyone made a basic python script to check keys from a script.

I guess GPU is a non issue, ill just get basic code working first.

I have done quite a bit of coding, but nothing on the cryptography side. I take it it would only be a few 100 lines of code?

Is it just like take 256bit private key guess, then hash it to the wallet address and check that matches the wallet address?

[madogss]

Just a big list of hex/binary/decimal numbers no ranges or anything like that, I cant go into anymore detail or will be out of pocket 6.7btc A python script would probably be fine but then that is CPU only then?. But that might work. Has anyone made a basic python script to check keys from a script.

I guess GPU is a non issue, ill just get basic code working first.

I have done quite a bit of coding, but nothing on the cryptography side. I take it it would only be a few 100 lines of code?

Is it just like take 256bit private key guess, then hash it to the wallet address and check that matches the wallet address?

with Iceland's secp256k1 library it's simple https://github.com/iceland2k14/secp256k1

Code:

import secp256k1 as ice

Target = "1BP7ByGjkekGpUPweRGjBVWWa6hagnnrnm" # change with address you want to find

n = 0
keys = []

with open("possibilities.txt", 'r') as vf:
    # here you would replace with getting each number from file and converting decimal if needed

while True:
    if ice.privatekey_to_address(0, True, keys[n]) == Target:
        with open("Found.txt", "w") as file:
            file.write(f"Found: {keys[n]}")
            file.close()
        print("FOUND")
        exit()
    n += 1

Since you have so many combinations you will either need a lot of ram or split into multiple files.

[dastic]

200 trillion combinations ? Do you have access to the NASA storage? 

[kTimesG]

Just a big list of hex/binary/decimal numbers no ranges or anything like that, I cant go into anymore detail or will be out of pocket 6.7btc A python script would probably be fine but then that is CPU only then?. But that might work. Has anyone made a basic python script to check keys from a script.

I guess GPU is a non issue, ill just get basic code working first.

I have done quite a bit of coding, but nothing on the cryptography side. I take it it would only be a few 100 lines of code?

Is it just like take 256bit private key guess, then hash it to the wallet address and check that matches the wallet address?

200 trillion keys = 200 * 10**15

Priv to Pub = bottleneck. Even if you write that in assembly code and run it on the fastest of CPUs you maybe can expect a few hundred thousands results per second (running on all cores). Doing it on a GPU involves:

- copying the private keys to device (slow)
- running point-scalar multiplication for each key in memory (slow)

so it's also really really slow. Maybe less than 100 MK/s on the best GPU that exists.

The speeds you see in KeyHunt(-CUDA) are speeds of public key additions, not of computing different "random" private keys to public keys.

Analyzing the second part (hashing the public key) is not even relevant since you already have something like a 50x slowdown just for getting to the public key.

So your text file (let's assume it's not even a text file, but you somehow produce private keys and feed them to CPU/GPU) will take many years to be processed. Something like 10.000 to 25.000 years.

[aby3er]

Hey everyone/
if someone needed for 68 :
1MVDYgVaSN odV2dkMm5CCXwDPUFYHqVXsH
(WIF): p2pkh: KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3qdCwdaSJRsEgap7BZBes
(HEX): C28A55EDE39083727
 
1MVDYgVaSN GX3x4AVSzhqLtd3PMjArLBXJ
(WIF): p2pkh: KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3qdabi5xJwfkBFCZ5Q2AY
(HEX): D2CA721767ADAD06A

you welcome 

[Akito S. M. Hosana]

so it's also really really slow. Maybe less than 100 MK/s on the best GPU that exists.

The speeds you see in KeyHunt(-CUDA) are speeds of public key additions, not of computing different "random" private keys to public keys.

There is a script here with hashing very similar to cpuminer. I have achieved results up to 2× faster than VanitySearch or KeyHunt using _mm256 and AVX on the latest AMD processors. However, it is still slow—on the order of several thousand years  

https://bitcointalk.org/index.php?topic=5532654.msg65082646#msg65082646


I bet that Frozen guy and NoMachine hiding something else. There must be something faster.  

[kTimesG]

There is a script here with hashing very similar to cpuminer. I have achieved results up to 2× faster than VanitySearch or KeyHunt using _mm256 and AVX on the latest AMD processors. However, it is still slow—on the order of several thousand years  

https://bitcointalk.org/index.php?topic=5532654.msg65082646#msg65082646


I bet that Frozen guy and NoMachine hiding something else. There must be something faster.  

Computing the public keys from private keys is magnitudes of order slower than the hashing part. So it doesn't matter if hashing is optimized since the pipeline gets stuck at EC point-scalar multiplication. Hashing will just stay at the base speed of EC computation.

[nomachine]

NoMachine hiding something else. There must be something faster.  

Are you still in the mood for someone to hide something from you? You've become boring. Go do some research and read. Everything related to hashing is publicly available on GitHub. Nothing new has been invented...

[bibilgin]

Hey everyone/
if someone needed for 68 :
1MVDYgVaSN odV2dkMm5CCXwDPUFYHqVXsH
(WIF): p2pkh: KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3qdCwdaSJRsEgap7BZBes
(HEX): C28A55EDE39083727
 
1MVDYgVaSN GX3x4AVSzhqLtd3PMjArLBXJ
(WIF): p2pkh: KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3qdabi5xJwfkBFCZ5Q2AY
(HEX): D2CA721767ADAD06A

you welcome  

1MVDYgVaS ZWCRp1nKLP8hZc2r5QTWx5KLo   C29B79FC2FFDF3572
1MVDYgVaS 3WFc6iwzVrv97pASwHDc5CiC2           C79D23414A3E3E030
1MVDYgVaS dMMnNZFfuLLJPPSA1VUTXQTHX   C87A398A338A21D39
1MVDYgVaS bDQwCSA5Mgk11jn2C3VHEAQqC   D06B271C2CD50BB69
1MVDYgVaS QXzMCQ5auVXoGAPAmr9sEmoMk   D12E755D842F37128
1MVDYgVaS Es5HmGKUYnnVyyyhtABqPUFwG   D35A1812864AD0268
1MVDYgVaS 1gGUUVnFiBQT4qDqK7La3NEWs   D46A77F259F08679C
1MVDYgVaS rwLPBLdfFvy1jL4XGu1io1dfH           D63B79E58EC92847E
1MVDYgVaS C2GYXqLhGhfpfcgxsskwwnJgV           D67E56765DE5C870B
1MVDYgVaS pLXUudyj1phpjGWrJ3cd4yKeC           D81B3CA26933D88E9
1MVDYgVaS bWVSg76JxytSNtyEo9DayTUik           D83C61E972501B2A0
1MVDYgVaS XAGM4pPJEWTXdaTioVbj1CLKK           D86B318CB5BCCAB0F
1MVDYgVaS 4rbgPf5UfNSttxRP2AstSaim9           D94E23D5D1DA514FC


You are welcome.

1MVDYgVaSN6ApM5LJXriH8cFZkgfCia6Dh   xxxxxxxxxx?

[dastic]

Hey everyone/
if someone needed for 68 :
1MVDYgVaSN odV2dkMm5CCXwDPUFYHqVXsH
(WIF): p2pkh: KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3qdCwdaSJRsEgap7BZBes
(HEX): C28A55EDE39083727
 
1MVDYgVaSN GX3x4AVSzhqLtd3PMjArLBXJ
(WIF): p2pkh: KwDiBf89QgGbjEhKnhXJuH7LrciVrZi3qdabi5xJwfkBFCZ5Q2AY
(HEX): D2CA721767ADAD06A

you welcome  

1MVDYgVaS ZWCRp1nKLP8hZc2r5QTWx5KLo   C29B79FC2FFDF3572
1MVDYgVaS 3WFc6iwzVrv97pASwHDc5CiC2           C79D23414A3E3E030
1MVDYgVaS dMMnNZFfuLLJPPSA1VUTXQTHX   C87A398A338A21D39
1MVDYgVaS bDQwCSA5Mgk11jn2C3VHEAQqC   D06B271C2CD50BB69
1MVDYgVaS QXzMCQ5auVXoGAPAmr9sEmoMk   D12E755D842F37128
1MVDYgVaS Es5HmGKUYnnVyyyhtABqPUFwG   D35A1812864AD0268
1MVDYgVaS 1gGUUVnFiBQT4qDqK7La3NEWs   D46A77F259F08679C
1MVDYgVaS rwLPBLdfFvy1jL4XGu1io1dfH           D63B79E58EC92847E
1MVDYgVaS C2GYXqLhGhfpfcgxsskwwnJgV           D67E56765DE5C870B
1MVDYgVaS pLXUudyj1phpjGWrJ3cd4yKeC           D81B3CA26933D88E9
1MVDYgVaS bWVSg76JxytSNtyEo9DayTUik           D83C61E972501B2A0
1MVDYgVaS XAGM4pPJEWTXdaTioVbj1CLKK           D86B318CB5BCCAB0F
1MVDYgVaS 4rbgPf5UfNSttxRP2AstSaim9           D94E23D5D1DA514FC


You are welcome.

1MVDYgVaSN6ApM5LJXriH8cFZkgfCia6Dh   xxxxxxxxxx?

This is not helpful in any way to solving the puzzle. It is just a "fun fact" thing.