Skip to content

hamnaz/BitCrack

 
 

Repository files navigation

BitCrack

A tool for brute-forcing Bitcoin private keys. The main purpose of this project is to contribute to the effort of solving the Bitcoin puzzle transaction: A transaction with 32 addresses that become increasingly difficult to crack.

Using BitCrack

Usage

Use cuBitCrack.exe for CUDA devices and clBitCrack.exe for OpenCL devices.

xxBitCrack.exe [OPTIONS] [TARGETS]

Where [TARGETS] are one or more Bitcoin address

Options:

-i, --in FILE
    Read addresses from FILE, one address per line. If FILE is "-" then stdin is read

-o, --out FILE
    Append private keys to FILE, one per line

-d, --device N
    Use device with ID equal to N

-b, --blocks BLOCKS
    The number of CUDA blocks

-t, --threads THREADS
    Threads per block

-p, --per-thread NUMBER
    Each thread will process NUMBER keys at a time

-s, --start KEY
    Start the search at KEY. KEY is any valid private key in hexadecimal format

-r, --range RANGE
    Number of keys to search

-c, --compressed
    Search for compressed keys (default). Can be used with -u to also search uncompressed keys

-u, --uncompressed
    Search for uncompressed keys, can be used with -c to search compressed keys

--list-devices
    List available devices


Examples

The simplest usage, the keyspace will begin at 0, and the CUDA parameters will be chosen automatically

xxBitCrack.exe 1FshYsUh3mqgsG29XpZ23eLjWV8Ur3VwH

Multiple keys can be searched at once with minimal impact to performance. Provide the keys on the command line, or in a file with one address per line

xxBitCrack.exe 1FshYsUh3mqgsG29XpZ23eLjWV8Ur3VwH 15JhYXn6Mx3oF4Y7PcTAv2wVVAuCFFQNiP 19EEC52krRUK1RkUAEZmQdjTyHT7Gp1TYT

To start the search at a specific private key, use the -s option:

xxBitCrack.exe -s 6BBF8CCF80F8E184D1D300EF2CE45F7260E56766519C977831678F0000000000 1FshYsUh3mqgsG29XpZ23eLjWV8Ur3VwH

Use the -b, -t and -p options to specify the number of blocks, threads per block, and keys per thread.

xxBitCrack.exe -b 32 -t 256 -p 16 1FshYsUh3mqgsG29XpZ23eLjWV8Ur3VwH

Use the -r or --range option to specify how many keys to search before stopping. For instance, to search up to 1 billion keys from the starting key:

xxBitCrack.exe -s 6BBF8CCF80F8E184D1D300EF2CE45F7260E56766519C977831678F0000000000 -r 1000000000

Note:

Integer values can be specified in decimal (e.g. 123) or in hexadecimal using the 0x prefix or h suffix (e.g. 0x1234 or 1234h)

Choosing the right parameters for your device

GPUs have many cores. Work for the cores is divided into blocks. Each block contains threads.

There are 3 parameters that affect performance: blocks, threads per block, and keys per thread.

blocks: Should be a multiple of the number of compute units on the device. The default is 32.

threads: The number of threads in a block. This must be a multiple of 32. The default is 256.

Keys per thread: The number of keys each thread will process. The performance (keys per second) increases asymptotically with this value. The default is 32. Increasing this value will cause the kernel to run longer, but more keys will be processed.

Build dependencies

Visual Studio 2017 (if on Windows)

For CUDA: CUDA Toolkit 9.2

For OpenCL: An OpenCL SDK

Building in Windows

Open the Visual Studio solution.

Build the clKeyFinder project for an OpenCL build.

Build the cuKeyFinder project for a CUDA build.

Note: By default the NVIDIA OpenCL headers are used. You can set the header and library path for OpenCL in the BitCrack.props property sheet.

Building in Linux

Using make:

Build CUDA:

make BUILD_CUDA=1

Build OpenCL:

make BUILD_OPENCL=1

Or build both:

make BUILD_CUDA=1 BUILD_OPENCL=1

Supporting this project

If you find this project useful and would like to support it, consider making a donation. Your support is greatly appreciated!

BTC: 1LqJ9cHPKxPXDRia4tteTJdLXnisnfHsof

LTC: LfwqkJY7YDYQWqgR26cg2T1F38YyojD67J

ETH: 0xd28082CD48E1B279425346E8f6C651C45A9023c5

About

A tool for cracking Bitcoin private keys

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • C++ 48.1%
  • Cuda 28.7%
  • C 20.8%
  • Makefile 2.4%