ckb-miner

A miner for CPUs, avx2 CPUs, avx512 CPUs and GPUs.

Using avx512, the hash rate can be up to 10M per thread on 2.5Ghz CPUs.

How to build

By default, GPU miner is not built, you can build GPU miner with this parameters:

# for cuda
cargo build --release --features cuda

#for opencl
cargo build --release --features opencl

CUDA runtime

By default, the cuda lib path is C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\lib\x64 for Windows and /usr/local/cuda/lib64 for other platforms, you can change this by set CUDA_LIB_DIR.

For Linux and OSX:

export CUDA_LIB_DIR=<PATH> # <PATH> is the path where you install CUDA, eg export CUDA_LIB_DIR=/usr/local/cuda/lib64

For windows:

set CUDA_LIB_DIR=<PATH> # <PATH> is the path where you install CUDA, eg set CUDA_LIB_DIR=C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v10.1\\lib\\x64

Cuda is essential to build and run miners based on Cuda.

Performance

OpenCL runtime

With OpenCL, the miner can run on any GPUs.

By default, the OpenCL include path is C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include for Windows and /usr/include for other platforms, you can change this by set OPENCL_INCLUDE_DIR; the OpenCL lib path is C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\lib\x64 for Windows and default lib path for other platforms, you can change this by set OPENCL_LIB_DIR

Performance

How to use

1. Modify ckb-miner.toml in this repo and copy it to your working directory. You can specify the CPUs and GPUs you want to use.

2. Copy ckb-miner (you can build it yourself or download it) to your working directory.

3. run the miner

./ckb-miner