README.md

CPUMiner-Multi

This is a multi-threaded CPU miner, fork of LucasJones' cpuminer-multi.

Table of contents

• Algorithms
• Dependencies
• Download
• Build
• Usage instructions
• Donations
• Credits
• License

Algorithms

Currently supported

• ✓ cryptonight (Bytecoin [BCN], Monero)

Dependencies

• libcurl http://curl.haxx.se/libcurl/
• jansson http://www.digip.org/jansson/ (jansson is included in-tree)

Download

• For binary releases, see Bitcointalk thread: https://bitcointalk.org/index.php?topic=632724
• Git tree: https://github.com/wolf9466/cpuminer-multi
• Clone with git clone https://github.com/wolf9466/cpuminer-multi

Build

Basic *nix build instructions:

• ./autogen.sh # only needed if building from git repo
• Optimal GCC flags are built in - you only need to use -march=native if you want it
• CFLAGS="-march=native" ./configure

  • Use -march=native if building for a single machine

• make

Architecture-specific notes:

• CryptoNight works only on x86 and x86-64.
• If you don't have AES-NI, it's slower. A lot slower, around 1/3rd the speed. This implementation is deprecated and will not be improved.

Usage instructions

Run "minerd --help" to see options.

Connecting through a proxy

Use the --proxy option.

To use a SOCKS proxy, add a socks4:// or socks5:// prefix to the proxy host
Protocols socks4a and socks5h, allowing remote name resolving, are also available since libcurl 7.18.0.

If no protocol is specified, the proxy is assumed to be a HTTP proxy.
When the --proxy option is not used, the program honors the http_proxy and all_proxy environment variables.

Donations

Donations for the work done in this fork are accepted at

• XMR: 46sSETXrZGT8bupxdc2MAbLe3PMV9nJTRTE5uaFErXFz6ymyzVdH86KDb9TNoG4ny5QLELfopynWeBSMoT1M2Ga8RBkDqTH
• BTC: 1WoLFumNUvjCgaCyjFzvFrbGfDddYrKNR

Credits

This faster CPUMiner-multi was forked from LucasJones', and has been developed by Wolf. Special thanks to Intel for helping me with the usage of the AESKEYGENASSIST instruction, which I used to replace the calls to the slow, unoptimized oaes_lib. Special thanks also to dga for his code with that loop - since it runs over 500,000 times per hash, it's extremely important.

License

GPLv2. See COPYING for details.