Skip to content

Latest commit

 

History

History

amd

OpenCL Support in excavator

Please read this document carefully, especially if you are mining Equihash-based coins.

This version of excavator has initial support for OpenCL devices, including the AMD Radeon series graphics cards. This is still work in progress and mostly optimized for RX 470/480/570/580 for now. If you have any questions/concerns/feedback, please feel free to post them to:

Alternatively, you can directly email the author at:

It is my hope that excavator would be useful for everybody with top-notch performance, solid reliability, and exceptional flexibility, and I am grateful to NiceHash for giving me an opportunity to work on this miner. Enjoy!

zawawa

Prerequisites

Excavator should work with any OpenCL-compatible devices. For optimal performance, however, please use AMD graphics cards with the GCN2 architecture or above and AMD Crimson Software 16.9.2.

Supported Algorithms

The following algorithms are currently supported:

Name Supported devices Wcount*1 Pcount*2
equihash GCN2+ 4 2
pascal GCN1+ 4 3
sia GCN1+ 4 2
decred GCN1+ 4 4

*1 Recommended number of workers per device to reach optimal speeds.

*2 Number of supported parameters. Parameters are explained in details in section for each algorithm.

equihash

For Equihash-based coins, such as Zcash, you need to run the miner as administrator as excavator needs to access the graphics card directly. Furthermore, driver signature enforcement needs to be disabled on Windows 10 for the time being. You can deactivate it either at startup or by running the following .bat file:

turn-driver-signature-enforcement-off.bat

Please note that you need administrator privileges to do so and a restart is required after executing the .bat file. This requirement will be removed for the next stable version.

Parameter # Range Explanation
1 0 or 1 Disable/enable binary kernels
2 0 or 1 Disable/enable the Global Data Share (requires admin privileges)

pascal

Parameter # Range Explanation
1 0-∞ Global work size
2 0-256 Local work size
3 0-∞ Number of iterations

If no parameters are provided, device specific defaults are used. If provided parameter is '0' then device specific default value is used.

sia

Parameter # Range Explanation
1 0-∞ Global work size
2 0-256 Local work size

If no parameters are provided, device specific defaults are used. If provided parameter is '0' then device specific default value is used.

decred

Parameter # Range Explanation
1 0 or 1 Disable/enable binary kernels
2 0-∞ Global work size
3 0-256 Local work size
4 0-∞ Number of iterations

If no parameters are provided, device specific defaults are used. If provided parameter is '0' then device specific default value is used.