Author

Topic: Python based Solo miner for CPU | Learn Basic Bitcoin Mining | Just for fun (Read 720 times)

newbie
Activity: 4
Merit: 0
It's very slow. I optimised a little, but couldn't find many ways to improve. One way way was change the libraries imported: instead of

Code:
import binascii 
binascii.hexlify(merkle_root).decode()

replaced with importing the specific function from the library and giving an assignation as below

Code:
from binascii import hexlify as h
h(merkle_root).decode()

Then do this for every function called from an imported library. Improvement was slight but noticable
member
Activity: 259
Merit: 85
So many numbers and so little time
I find this thread really quite interesting.

I've been wondering what the payloads look like going back and forth.

This simple script goes a long way to explaining it, thank you.

I may even give it a go, just to see how slow it is Wink

Thanks

G.
newbie
Activity: 1
Merit: 0
Helo,

In the script, during the loop on the nonce, merkle_root and ctx.ntime keep the same value. But, I discover that, during this loop, while continue listening socket, ctx.job_id, ctx.coinb1, ctx.coinb2, ctx.ntime, and ctx.merkle_branch change the value. So, during the loop we have an very old value of merkle_root and ctx.ntime which are not updated.

Do you think that, we have to always listening the socket during the loop and update the value of merkle_root and ctx.ntime, according to what we receive?
newbie
Activity: 4
Merit: 0

No, it should be < the target ( not equal to it).

When mining Bitcoin , the goal is to find a hash that is not equal to the target(near impossible), but less than the target value(probalistic).  The target is a 256-bit number that the network adjusts every 2016 blocks to ensure that the time between blocks remains approximately 10 minutes on average.  A lower target means a higher difficulty in finding a valid block.

The hashing process is essentially a race to find a nonce value that, when combined with the block's header data(which itself contains the Merkle root of the previous blocks tx) and passed through the SHA-256 hashing algorithm twice, produces a hash value that is less than the network-defined target, which everyone can easily verify for themselves too.

Therefore the condition, in the educational example,  if hash < target: is correct because you are looking for a hash that is numerically less than the target.

Thank you so much! I have just had to destroy my entire (wrong) understanding of how mining works, and remake it again from scratch!
jr. member
Activity: 38
Merit: 22
One more q for anyone who might be able to answer;

Code:
        if hash < target :
    #    if(hash[:10] == '0000000000'):
            print('success!!')
            print('hash: {}'.format(hash))
            payload = bytes('{"params": ["'+address+'", "'+job_id+'", "'+extranonce2 \
                +'", "'+ntime+'", "'+nonce+'"], "id": 1, "method": "mining.submit"}\n', 'utf-8')
            sock.sendall(payload)
            print(sock.recv(1024))
            input("Press Enter to continue...")

Why are we evaluating hash < target? Shouldn't that be

Code:
if hash == target:
     print('success!!')

No, it should be < the target ( not equal to it).

When mining Bitcoin , the goal is to find a hash that is not equal to the target(near impossible), but less than the target value(probalistic).  The target is a 256-bit number that the network adjusts every 2016 blocks to ensure that the time between blocks remains approximately 10 minutes on average.  A lower target means a higher difficulty in finding a valid block.

The hashing process is essentially a race to find a nonce value that, when combined with the block's header data(which itself contains the Merkle root of the previous blocks tx) and passed through the SHA-256 hashing algorithm twice, produces a hash value that is less than the network-defined target, which everyone can easily verify for themselves too.

Therefore the condition, in the educational example,  if hash < target: is correct because you are looking for a hash that is numerically less than the target.
newbie
Activity: 4
Merit: 0
One more q for anyone who might be able to answer;

Code:
        if hash < target :
    #    if(hash[:10] == '0000000000'):
            print('success!!')
            print('hash: {}'.format(hash))
            payload = bytes('{"params": ["'+address+'", "'+job_id+'", "'+extranonce2 \
                +'", "'+ntime+'", "'+nonce+'"], "id": 1, "method": "mining.submit"}\n', 'utf-8')
            sock.sendall(payload)
            print(sock.recv(1024))
            input("Press Enter to continue...")

Why are we evaluating hash < target? Shouldn't that be

Code:
if hash == target:
     print('success!!')
newbie
Activity: 4
Merit: 0
I made some small improvements:
1) I called the noncework function 2**32 times, not just 10 million million times
2) every time i called the noncework function, i passed the loop value k, using that number as the nounce and not a random number

I could also have used the time parameter, another 5000 or 6000 times cycle, so that the root merkle only resets once every 5000*2**32. In that case, if we had a 100 TH/S asic instead of a CPU, it would have to recalculate the nounce every 4 or 5 seconds, if I'm not mistaken.

Code:
# omitted for brevity
   
    #  use this loop if you want to do a speed test with only 1 million rounds
    start = time.time()
    for k in range(1000000):
        noncework(k)
    end = time.time()
    print(end - start)
    # about 42.5
    '''
    # use this loop if you want to try all the nounce with 2**32 rounds
    for k in range(2**32):
    noncework(k)
    '''
    print("Finished nounce. Regaining Information.")
    sock.close()
    main()

main()


I'm interested to understand why you chose (2**32) as the target for noncework. Given the end-start times when I tried your script, the noncework function takes around 2.5 hours. In that time, it's almost certain that a block would be found and you'd be looking at coinbase etc information from a non-current block. Am I wrong here, or should you be regaining information much sooner?

Also, can you explain the advantage of noncework(k) instead of noncework() ? Thanks!

I'm trying to optimize the script as a bit of fun, given I'll never mine a block, my goal is to improve my Mh/s figure as much as possible and learn more about the technical background underpinning mining.
legendary
Activity: 3038
Merit: 4418
Crypto Swap Exchange
You say just for fun, I wonder where is the fun if we start mining with our CPU 24/7 and never mine anything, could you provide some probability estimations on a single block found/time?
Looks to be more of a fun project to learn about Bitcoin mining, rather than doing so seriously. It's an easy way to see what is being hashed and being sent over to the pool. This script isn't particularly optimized, don't think it is multithreaded?

If you want to do so, then there are more optimized CPU miners to use, but even so you shouldn't expect anything at all.
newbie
Activity: 4
Merit: 0
You say just for fun, I wonder where is the fun if we start mining with our CPU 24/7 and never mine anything, could you provide some probability estimations on a single block found/time?

I've just fired this up on one core of a i5, running through 10MM hashes in around 56 seconds. That gives me 178,000 hashes/s or 0.178M/h, which presently equates to 65 BILLION years to find a block. That's between 14 and 15 times the age of the Earth, and slightly less than 5 times the age of the universe and everything in it.

Still, gotta be in it to win it!
jr. member
Activity: 47
Merit: 18
Another small improvement, the last one I think. I didn't use hashlib to hash the nounce, but I used this 100% python implementation: https://github.com/keanemind/python-sha-256.
In this way the code is about 3.5 slower (but that's just to understand, performances aren't important) but you can understand how the hash is encoded.
It sure is useful to me, and I will use it as a starting point for my own serious mining algorithm that I am building (https://bitcointalksearch.org/topic/need-help-with-proof-of-concept-cpu-mining-5446391). I hope  it is useful to others too.

Code:
# -*- coding: utf-8 -*-

import socket
import json
import hashlib
import binascii
from pprint import pprint
import random
import time

address = 'YourBitcoinAddress'
nonce   = hex(0)[2:].zfill(8)
host    = 'solo.ckpool.org'
port    = 3333
#host    = 'pool.mainnet.bitcoin-global.io'
#port    = 9223

K = [
    0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5, 0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5,
    0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3, 0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174,
    0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc, 0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da,
    0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7, 0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967,
    0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13, 0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85,
    0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3, 0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070,
    0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5, 0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3,
    0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208, 0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2
]

def generate_hash(message: bytearray) -> bytearray:
    """Return a SHA-256 hash from the message passed.
    The argument should be a bytes, bytearray, or
    string object."""

    if isinstance(message, str):
        message = bytearray(message, 'ascii')
    elif isinstance(message, bytes):
        message = bytearray(message)
    elif not isinstance(message, bytearray):
        raise TypeError

    # Padding
    length = len(message) * 8 # len(message) is number of BYTES!!!
    message.append(0x80)
    while (len(message) * 8 + 64) % 512 != 0:
        message.append(0x00)

    message += length.to_bytes(8, 'big') # pad to 8 bytes or 64 bits

    assert (len(message) * 8) % 512 == 0, "Padding did not complete properly!"

    # Parsing
    blocks = [] # contains 512-bit chunks of message
    for i in range(0, len(message), 64): # 64 bytes is 512 bits
        blocks.append(message[i:i+64])

    # Setting Initial Hash Value
    h0 = 0x6a09e667
    h1 = 0xbb67ae85
    h2 = 0x3c6ef372
    h3 = 0xa54ff53a
    h5 = 0x9b05688c
    h4 = 0x510e527f
    h6 = 0x1f83d9ab
    h7 = 0x5be0cd19

    # SHA-256 Hash Computation
    for message_block in blocks:
        # Prepare message schedule
        message_schedule = []
        for t in range(0, 64):
            if t <= 15:
                # adds the t'th 32 bit word of the block,
                # starting from leftmost word
                # 4 bytes at a time
                message_schedule.append(bytes(message_block[t*4:(t*4)+4]))
            else:
                term1 = _sigma1(int.from_bytes(message_schedule[t-2], 'big'))
                term2 = int.from_bytes(message_schedule[t-7], 'big')
                term3 = _sigma0(int.from_bytes(message_schedule[t-15], 'big'))
                term4 = int.from_bytes(message_schedule[t-16], 'big')

                # append a 4-byte byte object
                schedule = ((term1 + term2 + term3 + term4) % 2**32).to_bytes(4, 'big')
                message_schedule.append(schedule)

        assert len(message_schedule) == 64

        # Initialize working variables
        a = h0
        b = h1
        c = h2
        d = h3
        e = h4
        f = h5
        g = h6
        h = h7

        # Iterate for t=0 to 63
        for t in range(64):
            t1 = ((h + _capsigma1(e) + _ch(e, f, g) + K[t] +
                   int.from_bytes(message_schedule[t], 'big')) % 2**32)

            t2 = (_capsigma0(a) + _maj(a, b, c)) % 2**32

            h = g
            g = f
            f = e
            e = (d + t1) % 2**32
            d = c
            c = b
            b = a
            a = (t1 + t2) % 2**32

        # Compute intermediate hash value
        h0 = (h0 + a) % 2**32
        h1 = (h1 + b) % 2**32
        h2 = (h2 + c) % 2**32
        h3 = (h3 + d) % 2**32
        h4 = (h4 + e) % 2**32
        h5 = (h5 + f) % 2**32
        h6 = (h6 + g) % 2**32
        h7 = (h7 + h) % 2**32

    return ((h0).to_bytes(4, 'big') + (h1).to_bytes(4, 'big') +
            (h2).to_bytes(4, 'big') + (h3).to_bytes(4, 'big') +
            (h4).to_bytes(4, 'big') + (h5).to_bytes(4, 'big') +
            (h6).to_bytes(4, 'big') + (h7).to_bytes(4, 'big'))

def _sigma0(num: int):
    """As defined in the specification."""
    num = (_rotate_right(num, 7) ^
           _rotate_right(num, 18) ^
           (num >> 3))
    return num

def _sigma1(num: int):
    """As defined in the specification."""
    num = (_rotate_right(num, 17) ^
           _rotate_right(num, 19) ^
           (num >> 10))
    return num

def _capsigma0(num: int):
    """As defined in the specification."""
    num = (_rotate_right(num, 2) ^
           _rotate_right(num, 13) ^
           _rotate_right(num, 22))
    return num

def _capsigma1(num: int):
    """As defined in the specification."""
    num = (_rotate_right(num, 6) ^
           _rotate_right(num, 11) ^
           _rotate_right(num, 25))
    return num

def _ch(x: int, y: int, z: int):
    """As defined in the specification."""
    return (x & y) ^ (~x & z)

def _maj(x: int, y: int, z: int):
    """As defined in the specification."""
    return (x & y) ^ (x & z) ^ (y & z)

def _rotate_right(num: int, shift: int, size: int = 32):
    """Rotate an integer right."""
    return (num >> shift) | (num << size - shift)



def main():
    print("address:{} nonce:{}".format(address,nonce))
    print("host:{} port:{}".format(host,port))
   
    sock    = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
    sock.connect((host,port))
   
    #server connection
    sock.sendall(b'{"id": 1, "method": "mining.subscribe", "params": []}\n')
    lines = sock.recv(1024).decode().split('\n')
    response = json.loads(lines[0])
    sub_details,extranonce1,extranonce2_size = response['result']
   
    #authorize workers
    sock.sendall(b'{"params": ["'+address.encode()+b'", "password"], "id": 2, "method": "mining.authorize"}\n')
   
    #we read until 'mining.notify' is reached
    response = b''
    while response.count(b'\n') < 4 and not(b'mining.notify' in response):
        response += sock.recv(1024)
   
   
    #get rid of empty lines
    responses = [json.loads(res) for res in response.decode().split('\n') if len(res.strip())>0 and 'mining.notify' in res]
    pprint(responses)
   
    job_id,prevhash,coinb1,coinb2,merkle_branch,version,nbits,ntime,clean_jobs \
        = responses[0]['params']
   
    target = (nbits[2:]+'00'*(int(nbits[:2],16) - 3)).zfill(64)

    print('nbits:{} target:{}\n'.format(nbits,target))
   
    # extranonce2 = '00'*extranonce2_size
    extranonce2 = hex(random.randint(0,2**32-1))[2:].zfill(2*extranonce2_size)      # create random
   
    coinbase = coinb1 + extranonce1 + extranonce2 + coinb2
    coinbase_hash_bin = hashlib.sha256(hashlib.sha256(binascii.unhexlify(coinbase)).digest()).digest()
   
    print('coinbase:\n{}\n\ncoinbase hash:{}\n'.format(coinbase,binascii.hexlify(coinbase_hash_bin)))
    merkle_root = coinbase_hash_bin
    for h in merkle_branch:
        merkle_root = hashlib.sha256(hashlib.sha256(merkle_root + binascii.unhexlify(h)).digest()).digest()
   
    merkle_root = binascii.hexlify(merkle_root).decode()
   
    #little endian
    merkle_root = ''.join([merkle_root[i]+merkle_root[i+1] for i in range(0,len(merkle_root),2)][::-1])
   
    print('merkle_root:{}\n'.format(merkle_root))
   
    def noncework(k):
        nonce   = hex(k)[2:].zfill(8)   #hex(int(nonce,16)+1)[2:]
        blockheader = version + prevhash + merkle_root + nbits + ntime + nonce +\
            '000000800000000000000000000000000000000000000000000000000000000000000000000000000000000080020000'
       
       
        # print('blockheader:\n{}\n'.format(blockheader))

        hash=generate_hash(generate_hash(binascii.unhexlify(blockheader))).hex()

        # hash = hashlib.sha256(hashlib.sha256(binascii.unhexlify(blockheader)).digest()).digest()
        #hash = binascii.hexlify(blockheader).decode()
        # print('hash: {}'.format(hash))
        if(hash[:5] == '00000'): print('hash: {}'.format(hash))
        if hash < target :
    #    if(hash[:10] == '0000000000'):
            print('success!!')
            print('hash: {}'.format(hash))
            payload = bytes('{"params": ["'+address+'", "'+job_id+'", "'+extranonce2 \
                +'", "'+ntime+'", "'+nonce+'"], "id": 1, "method": "mining.submit"}\n', 'utf-8')
            sock.sendall(payload)
            print(sock.recv(1024))
            input("Press Enter to continue...")
    #    else:
    #        print('failed mine, hash is greater than target')
   
    start = time.time()
    for k in range(1000000):
        noncework(k)
    end = time.time()
    print(end - start)
    '''
    for k in range(2**32):
    noncework(k)
    '''
    sock.close()
    main()

main()
jr. member
Activity: 47
Merit: 18
I made some small improvements:
1) I called the noncework function 2**32 times, not just 10 million million times
2) every time i called the noncework function, i passed the loop value k, using that number as the nounce and not a random number

I could also have used the time parameter, another 5000 or 6000 times cycle, so that the root merkle only resets once every 5000*2**32. In that case, if we had a 100 TH/S asic instead of a CPU, it would have to recalculate the nounce every 4 or 5 seconds, if I'm not mistaken.

Code:
# -*- coding: utf-8 -*-

import socket
import json
import hashlib
import binascii
from pprint import pprint
import random
import time

address = 'YourBitcoinAddress'
nonce   = hex(0)[2:].zfill(8)
host    = 'solo.ckpool.org'
port    = 3333
#host    = 'pool.mainnet.bitcoin-global.io'
#port    = 9223

def main():
    print("address:{} nonce:{}".format(address,nonce))
    print("host:{} port:{}".format(host,port))
   
    sock    = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
    sock.connect((host,port))
   
    #server connection
    sock.sendall(b'{"id": 1, "method": "mining.subscribe", "params": []}\n')
    lines = sock.recv(1024).decode().split('\n')
    response = json.loads(lines[0])
    sub_details,extranonce1,extranonce2_size = response['result']
   
    #authorize workers
    sock.sendall(b'{"params": ["'+address.encode()+b'", "password"], "id": 2, "method": "mining.authorize"}\n')
   
    #we read until 'mining.notify' is reached
    response = b''
    while response.count(b'\n') < 4 and not(b'mining.notify' in response):
        response += sock.recv(1024)
   
   
    #get rid of empty lines
    responses = [json.loads(res) for res in response.decode().split('\n') if len(res.strip())>0 and 'mining.notify' in res]
    pprint(responses)
   
    job_id,prevhash,coinb1,coinb2,merkle_branch,version,nbits,ntime,clean_jobs \
        = responses[0]['params']
   
    target = (nbits[2:]+'00'*(int(nbits[:2],16) - 3)).zfill(64)

    print('nbits:{} target:{}\n'.format(nbits,target))
   
    # extranonce2 = '00'*extranonce2_size
    extranonce2 = hex(random.randint(0,2**32-1))[2:].zfill(2*extranonce2_size)      # create random
   
    coinbase = coinb1 + extranonce1 + extranonce2 + coinb2
    coinbase_hash_bin = hashlib.sha256(hashlib.sha256(binascii.unhexlify(coinbase)).digest()).digest()
   
    print('coinbase:\n{}\n\ncoinbase hash:{}\n'.format(coinbase,binascii.hexlify(coinbase_hash_bin)))
    merkle_root = coinbase_hash_bin
    for h in merkle_branch:
        merkle_root = hashlib.sha256(hashlib.sha256(merkle_root + binascii.unhexlify(h)).digest()).digest()
   
    merkle_root = binascii.hexlify(merkle_root).decode()
   
    #little endian
    merkle_root = ''.join([merkle_root[i]+merkle_root[i+1] for i in range(0,len(merkle_root),2)][::-1])
   
    print('merkle_root:{}\n'.format(merkle_root))
   
    def noncework(k):
        nonce   = hex(k)[2:].zfill(8)   #hex(int(nonce,16)+1)[2:]
        blockheader = version + prevhash + merkle_root + nbits + ntime + nonce +\
            '000000800000000000000000000000000000000000000000000000000000000000000000000000000000000080020000'
       
        # print('blockheader:\n{}\n'.format(blockheader))
       
        hash = hashlib.sha256(hashlib.sha256(binascii.unhexlify(blockheader)).digest()).digest()
        hash = binascii.hexlify(hash).decode()
        # print('hash: {}'.format(hash))
        if(hash[:5] == '00000'): print('hash: {}'.format(hash))
        if hash < target :
    #    if(hash[:10] == '0000000000'):
            print('success!!')
            print('hash: {}'.format(hash))
            payload = bytes('{"params": ["'+address+'", "'+job_id+'", "'+extranonce2 \
                +'", "'+ntime+'", "'+nonce+'"], "id": 1, "method": "mining.submit"}\n', 'utf-8')
            sock.sendall(payload)
            print(sock.recv(1024))
            input("Press Enter to continue...")
    #    else:
    #        print('failed mine, hash is greater than target')
   
    #  use this loop if you want to do a speed test with only 1 million rounds
    start = time.time()
    for k in range(1000000):
        noncework(k)
    end = time.time()
    print(end - start)
    # about 42.5
    '''
    # use this loop if you want to try all the nounce with 2**32 rounds
    for k in range(2**32):
    noncework(k)
    '''
    print("Finished nounce. Regaining Information.")
    sock.close()
    main()

main()

jr. member
Activity: 47
Merit: 18
Quote
why are you calling sha256 twice? so you hash the hash of the header.
Because Bitcoin works in that way. For example, the latest block hash (at the moment of writing) is calculated in this way:
Code:
SHA-256(00000020ceee9013d84c76a7e5a980df9dec537f313a3259da2d02000000000000000000b90235771c6c90b9d13e4b3670b489cb0fadc6bcbf610f4d0ead0c1d1c8d068deaf22c643e020617cb4b5c51)=88b46b0817fd187eba000740d5b727fd961bcb8df91fb4a7564610a4fe2fc76c
SHA-256(88b46b0817fd187eba000740d5b727fd961bcb8df91fb4a7564610a4fe2fc76c)=ce1c179317c8c1271adf317798d8d9ba5cb0cdf93c0404000000000000000000
I did so.e investigation and....i Was wrong. Really thanks, this helped me a lot!!
copper member
Activity: 821
Merit: 1992
Quote
why are you calling sha256 twice? so you hash the hash of the header.
Because Bitcoin works in that way. For example, the latest block hash (at the moment of writing) is calculated in this way:
Code:
SHA-256(00000020ceee9013d84c76a7e5a980df9dec537f313a3259da2d02000000000000000000b90235771c6c90b9d13e4b3670b489cb0fadc6bcbf610f4d0ead0c1d1c8d068deaf22c643e020617cb4b5c51)=88b46b0817fd187eba000740d5b727fd961bcb8df91fb4a7564610a4fe2fc76c
SHA-256(88b46b0817fd187eba000740d5b727fd961bcb8df91fb4a7564610a4fe2fc76c)=ce1c179317c8c1271adf317798d8d9ba5cb0cdf93c0404000000000000000000
jr. member
Activity: 260
Merit: 6
thanks, i also want to practice some code for cpu/gpu bitcoin mining, so i very gladly looked at your work.

a question regarding this line of code: hash = hashlib.sha256(hashlib.sha256(binascii.unhexlify(blockheader)).digest()).digest()
why are you calling sha256 twice? so you hash the hash of the header.
maybe it's a mistake?

Should be able to implement PyCuda into that code. I'd love to see that myself. 

Bitcoin double hashes. That's how it works. When you hash the block and think you're done, you're not. You have to hash it a second time to see if it has the requisite zeros.
jr. member
Activity: 260
Merit: 6
You say just for fun, I wonder where is the fun if we start mining with our CPU 24/7 and never mine anything, could you provide some probability estimations on a single block found/time?

Yeah, doesn't sound fun... for educational purposes only? I'd prefer to mine some shitcoin using my CPU or GPU but with some real results/mining involved. Solo mining even on an Antminer S19 Pro is a lottery now already - 0.000065% chance of catching the block every 10 minutes (around 30 years to solo mine a block), CPU mining is not even a lottery, you're just losing money on electricity and burning out your CPU.  Grin

If you're "burning out" your CPU, you need a better cooling solution Wink.
jr. member
Activity: 47
Merit: 18
thanks, i also want to practice some code for cpu/gpu bitcoin mining, so i very gladly looked at your work.

a question regarding this line of code: hash = hashlib.sha256(hashlib.sha256(binascii.unhexlify(blockheader)).digest()).digest()
why are you calling sha256 twice? so you hash the hash of the header.
maybe it's a mistake?
legendary
Activity: 2422
Merit: 1191
Privacy Servers. Since 2009.
You say just for fun, I wonder where is the fun if we start mining with our CPU 24/7 and never mine anything, could you provide some probability estimations on a single block found/time?

Yeah, doesn't sound fun... for educational purposes only? I'd prefer to mine some shitcoin using my CPU or GPU but with some real results/mining involved. Solo mining even on an Antminer S19 Pro is a lottery now already - 0.000065% chance of catching the block every 10 minutes (around 30 years to solo mine a block), CPU mining is not even a lottery, you're just losing money on electricity and burning out your CPU.  Grin
legendary
Activity: 1568
Merit: 6660
bitcoincleanup.com / bitmixlist.org
Congratulations on learning how to mine bitcoins!

I also appreciate the fact that you're not connecting to some regtest node, but rather, you are using a live mining pool (ckpool, in this case).

Have you considered making a guide to connecting a USB miner such as the Compac F or AntMiner U1/2 to a mining pool?
legendary
Activity: 2464
Merit: 4415
🔐BitcoinMessage.Tools🔑
I'm curious, why does it randomly go through nonces instead of just counting through them by counting up one by one?
It may be that I am missing something but I think it is just an implementation detail that doesn't affect the chances of finding a correct hash but makes the function a little bit easier to design and reason about. As you can see from the code, the algorithm performs a predetermined number of operations and then starts over by invoking the main function. If it were incrementing nonce instead of taking a random value, it would need to keep track of the actual value, pass it back to the main function, and reset it after each found block. Using random numbers saves you from having to choose the right moment to relaunch the nonce management algorithm. The only problem I see is that the range of random values is tiny compared to all possible values.
jr. member
Activity: 260
Merit: 6
This code is great! Its a good start to what I was looking for. Any intentions on updating it? I'm curious, why does it randomly go through nonces instead of just counting through them by counting up one by one?
copper member
Activity: 1330
Merit: 899
🖤😏
You say just for fun, I wonder where is the fun if we start mining with our CPU 24/7 and never mine anything, could you provide some probability estimations on a single block found/time?
full member
Activity: 994
Merit: 117
Quote
My quick search your code might be based on one of these repository,
https://github.com/iceland2k14/solominer
https://github.com/Pymmdrza/SoloMiner


Yeah I guess its from iceland with few modifications.

Quote
This only applies to Python 2, your code is in Python 3.
Yes but this code was initially written for Python 2.7 so I described that method. For python3 users can directly import as suggested by @witcher_sense. And usually people don't get missing dependecy errors in Python3 as it comes with multiple pre installed libraries.
For python3 pip3 install works fine in many cases.
legendary
Activity: 2464
Merit: 4415
🔐BitcoinMessage.Tools🔑
pip install  error
/home/user/Downloads# pip install hashlib

/home/user/Downloads# pip install binascii


You don't need to install from pip, these modules are from standard python library, which means they come with interpreter.

Just import them like this:

Code:
import binascii
import hashlib
newbie
Activity: 13
Merit: 0
pip install  error
/home/user/Downloads# pip install hashlib
Code:
Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/
Collecting hashlib
  Using cached hashlib-20081119.zip (42 kB)
  error: subprocess-exited-with-error
 
  × python setup.py egg_info did not run successfully.
  │ exit code: 1
  ╰─> See above for output.
 
  note: This error originates from a subprocess, and is likely not a problem with pip.
  Preparing metadata (setup.py) ... error
error: metadata-generation-failed

× Encountered error while generating package metadata.
╰─> See above for output.

note: This is an issue with the package mentioned above, not pip.
hint: See above for details.

/home/user/Downloads# pip install binascii
Code:
Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/
ERROR: Could not find a version that satisfies the requirement binascii (from versions: none)
ERROR: No matching distribution found for binascii

newbie
Activity: 13
Merit: 0
full member
Activity: 994
Merit: 117
Hello Bitcoiners

I want to share a python based solo bitcoin miner which uses ckpool. You can use other pools as well if you want. It is basically like a lottery which has extremly low chances to win but it can be used as a proof of concept. Maybe it has already been shared somewhere and I do not remember its true origin so cannot give proper reference.
You can make changes as you want. This code can help users understand true basics of mining bitcoins.
Installing Dependencies
You must have python. Try with old versions first.
Install dependencies with
Code:
pip install hashlib
pip install binascii
If any dependency is missing you can use pip install DependencyName.

Code
Code:
# -*- coding: utf-8 -*-

import socket
import json
import hashlib
import binascii
from pprint import pprint
import random


address = 'YourBitcoinAddress'
nonce   = hex(random.randint(0,2**32-1))[2:].zfill(8)
host    = 'solo.ckpool.org'
port    = 3333
#host    = 'pool.mainnet.bitcoin-global.io'
#port    = 9223

def main():
    print("address:{} nonce:{}".format(address,nonce))
    print("host:{} port:{}".format(host,port))
   
    sock    = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
    sock.connect((host,port))
   
    #server connection
    sock.sendall(b'{"id": 1, "method": "mining.subscribe", "params": []}\n')
    lines = sock.recv(1024).decode().split('\n')
    response = json.loads(lines[0])
    sub_details,extranonce1,extranonce2_size = response['result']
   
    #authorize workers
    sock.sendall(b'{"params": ["'+address.encode()+b'", "password"], "id": 2, "method": "mining.authorize"}\n')
   
    #we read until 'mining.notify' is reached
    response = b''
    while response.count(b'\n') < 4 and not(b'mining.notify' in response):
        response += sock.recv(1024)
   
   
    #get rid of empty lines
    responses = [json.loads(res) for res in response.decode().split('\n') if len(res.strip())>0 and 'mining.notify' in res]
    pprint(responses)
   
    job_id,prevhash,coinb1,coinb2,merkle_branch,version,nbits,ntime,clean_jobs \
        = responses[0]['params']
   
    target = (nbits[2:]+'00'*(int(nbits[:2],16) - 3)).zfill(64)
    print('nbits:{} target:{}\n'.format(nbits,target))
   
    # extranonce2 = '00'*extranonce2_size
    extranonce2 = hex(random.randint(0,2**32-1))[2:].zfill(2*extranonce2_size)      # create random
   
    coinbase = coinb1 + extranonce1 + extranonce2 + coinb2
    coinbase_hash_bin = hashlib.sha256(hashlib.sha256(binascii.unhexlify(coinbase)).digest()).digest()
   
    print('coinbase:\n{}\n\ncoinbase hash:{}\n'.format(coinbase,binascii.hexlify(coinbase_hash_bin)))
    merkle_root = coinbase_hash_bin
    for h in merkle_branch:
        merkle_root = hashlib.sha256(hashlib.sha256(merkle_root + binascii.unhexlify(h)).digest()).digest()
   
    merkle_root = binascii.hexlify(merkle_root).decode()
   
    #little endian
    merkle_root = ''.join([merkle_root[i]+merkle_root[i+1] for i in range(0,len(merkle_root),2)][::-1])
   
    print('merkle_root:{}\n'.format(merkle_root))
   
    def noncework():
        nonce   = hex(random.randint(0,2**32-1))[2:].zfill(8)   #hex(int(nonce,16)+1)[2:]
        blockheader = version + prevhash + merkle_root + nbits + ntime + nonce +\
            '000000800000000000000000000000000000000000000000000000000000000000000000000000000000000080020000'
       
    #    print('blockheader:\n{}\n'.format(blockheader))
       
        hash = hashlib.sha256(hashlib.sha256(binascii.unhexlify(blockheader)).digest()).digest()
        hash = binascii.hexlify(hash).decode()
        # print('hash: {}'.format(hash))
        if(hash[:5] == '00000'): print('hash: {}'.format(hash))
        if hash < target :
    #    if(hash[:10] == '0000000000'):
            print('success!!')
            print('hash: {}'.format(hash))
            payload = bytes('{"params": ["'+address+'", "'+job_id+'", "'+extranonce2 \
                +'", "'+ntime+'", "'+nonce+'"], "id": 1, "method": "mining.submit"}\n', 'utf-8')
            sock.sendall(payload)
            print(sock.recv(1024))
            input("Press Enter to continue...")
    #    else:
    #        print('failed mine, hash is greater than target')
   
    for k in range(10000000):
        noncework()
    print("Finished 10M Search. Regaining Information.")
    sock.close()
    main()

main()



Edit Code
You can edit code. Set your Bitcoin address to receive your mining rewards. Set pool host and port. Save code in a .py file.
One of popular solo pool is solo.ckpool.org which has low fee of like 2%. You get 98% of your mining reward.

Run Code
You can run code by opening command prompt in same folder of file and run command
Code:
python FileName.py


Missing Dependencies and errors
As decribed above you can use pip install command to install any missing dependency. You can google errors and still if you did not find solution maybe your python version is not compatible. Try changing python version.

Future Work
If you improve this code then share here so others can benefit as well and I will update this code as well.


Honor List for Code Improvers
  • 1.
  • 2.
  • 3.


Tribute
I want to pay tribute to original author of this code. I am not original author.
Jump to: