This hobby project maintained by Justin T. Arthur undertakes the quixotic task of maintaining a modern Bitcoin miner for programmable compute devices like GPUs. It was forked from the PyOpenCL Bitcoin Miner (poclbm), a project authored by m0mchil and contributors.
It features an OpenCL Kernel that has incorporated ideas or code from:
If your work is represented herein and I didn't give you credit, please let me know. At the moment, I reserve no rights to the mining driver or the OpenCL kernel. They were derived from public domain works.
At the time of writing, on-chip implementations of the Bitcoin mining algorithm will outperform this software in both time and joules expended. Under most conditions, mining blocks on a Bitcoin chain where on-chip implementations are competing would be at a tremendous waste of expended resources.
Bug fix release. Installation from PyPI (e.g. with pip) had some glaring issues. Additionally, PyOpenCL removed deprecated APIs we were using. I've pinned us to PyOpenCL releases prior to their pybind11 integration so that this project remains installable from pip (see inducer/pyopencl#278).
This release focused on supporting the getblocktemplate
call proposed in
BIP 22 and BIP 23 and implemented by Bitcoin Core and btcd. This finally allows
mining without a pool if you supply an address to mine coins to at the command
line.
HTTP work sources now default to using getblocktemplate
instead of getwork
.
This feature is new and might have bugs that could lead to loss of potential
mining rewards.
It looks like the work sourcing threads run into i/o issues occasionally due to using the not-thread-safe Python http lib. I don't aim to address this as most of the threaded communication ought to be completely replaced by an event runner like asyncio or trio at some point.
Thanks to @momchil for the original getwork
code, @luke-jr @sipa and @vsergeev
for helping me understand getblocktemplate.
In an environment with Python 3.5+:
pip3 install apoclypsebm
apoclypse [OPTION]... SERVER[#tag]...
SERVER
is one or more [http[s]|stratum://]user:pass@host:port (required)
[#tag] is an optional per server user-friendly name displayed in stats.
--version show program's version number and exit
-h, --help show this help message and exit
--verbose verbose output, suitable for redirection to log file
-q, --quiet suppress all output except hash rate display
--proxy=PROXY specify as
[[socks4|socks5|http://]user:pass@]host:port (default
proto is socks5)
--no-ocl don't use OpenCL
--no-bfl don't use Butterfly Labs
--stratum-proxies search for and use stratum proxies in subnet
-d DEVICE, --device=DEVICE
comma separated device IDs, by default will use all
(for OpenCL - only GPU devices)
-a ADDRESS, --address=ADDRESS
Bitcoin address to spend the block reward to if
allowed. Required for solo mining, ignored with
stratum or getwork sources.
--coinbase-msg=COINBASE_MSG
Custom text to include in the coinbase of the
generation tx if allowed, encoded as UTF-8.
default=ApoCLypse
Miner Options:
-r RATE, --rate=RATE
hash rate display interval in seconds, default=1 (60
with --verbose)
-e ESTIMATE, --estimate=ESTIMATE
estimated rate time window in seconds, default 900 (15
minutes)
-t TOLERANCE, --tolerance=TOLERANCE
use fallback pool only after N consecutive connection
errors, default 2
-b FAILBACK, --failback=FAILBACK
attempt to fail back to the primary pool after N
seconds, default 60
--cutoff-temp=CUTOFF_TEMP
AMD GPUs, BFL only. For GPUs requires
github.com/mjmvisser/adl3. Comma separated
temperatures at which to skip kernel execution, in C,
default=95
--cutoff-interval=CUTOFF_INTERVAL
how long to not execute calculations if CUTOFF_TEMP is
reached, in seconds, default=0.01
--no-server-failbacks
disable using failback hosts provided by server
OpenCL Options:
Every option except '-p' and '-v' can be specified as a
comma separated list. If there aren't enough entries specified, the
last available is used. Use --vv to specify per-device vectors usage.
-p PLATFORM, --platform=PLATFORM
use platform by id
-w WORKSIZE, --worksize=WORKSIZE
work group size, default is maximum returned by OpenCL
-f FRAMES, --frames=FRAMES
will try to bring single kernel execution to 1/frames
seconds, default=30, increase this for less desktop
lag
-s FRAMESLEEP, --sleep=FRAMESLEEP
sleep per frame in seconds, default 0
--vv=VECTORS Specifies size of SIMD vectors per selected device.
Only size 0 (no vectors) and 2 supported for now.
Comma separated for each device. e.g. 0,2,2
-v, --vectors Use 2-item vectors for all devices.