Pool and wallet:
-a, --algo NAME algorithm: quantus, pearlhash, xelishashv3, kawpow,
nexapow, progpowz, octopus, autolykos2, qhash,
sha256d
-o, --url URL pool address; host:port is plain TCP,
stratum+ssl://host:port is encrypted. Repeat the
flag to add backup pools
-u, --user LOGIN wallet or account for the pool before it; a backup
reuses the first login unless given its own
-p, --pass PASSWORD pool password (default: x)
-w, --worker NAME worker name, sent to the pool as user.worker
Connection and failover:
--pool-retries N connection attempts per pool before the next one
(default: 1)
--pool-retry-pause N seconds between attempts (default: 5)
--pool-try-main keep testing the primary while mining on a backup
--pool-max-rejects N drop the connection after N rejects in a row
(default: 5; 0 turns it off)
--pool-send-stale always submit stale shares
--pool-timeout N connection and handshake timeout (default: 10)
--proxy ADDRESS SOCKS5 proxy: host:port, or
socks5://user:password@host:port
--tls-verify 1|0 verify TLS certificates (default: 1)
--dns-over-https NAME resolve pool hosts through google, cloudflare or
alibaba (default: off)
GPU selection:
-d, --gpu-list LIST GPUs to mine on, e.g. 0,2,3 (default: all)
--gpu-allow-igpu include integrated GPUs, skipped by default
--gpu-no-amd skip AMD cards; also --gpu-no-nvidia, --gpu-no-intel
--gpu-no-cuda disable NVIDIA mining when its driver breaks startup
--gpu-no-opencl disable AMD and Intel mining for the same reason
Performance and temperature:
-i, --gpu-intensity N work per batch, 1..31; leave unset for automatic
--gpu-limit-compute N ask for 1..100 percent of the card's compute units
--qhash-kernel 1|2 QHash mode; compare the two with a benchmark
(default: 1)
--gpu-temp-limit C pause a GPU at this temperature (default: 90)
--gpu-temp-resume C resume below this one (default: 65)
Clocks, power and fans — NVIDIA only, and some settings need administrator or root:
--gpu-core-clock N lock the core clock, MHz
--gpu-core-offset N core clock offset, MHz
--gpu-memory-clock N lock the memory clock, MHz
--gpu-memory-offset N memory clock offset, MHz
--gpu-powerlimit N power limit, watts
--gpu-fan-speed N fan speed, percent
--gpu-reset-oc unlock clocks and clear offsets at startup
--gpu-delay-oc N wait N seconds before applying the settings
Monitoring, interface and logs:
--api-port N serve JSON statistics over HTTP on this port
--api-worker-id NAME rig name the API reports (default: hostname)
--watchdog once every GPU has failed, write down the culprit
and exit, so a supervisor can restart the miner
--ui-plain plain logs instead of the full-screen dashboard
--log-file PATH append the log, and a GPU report every minute
--print-debug extra detail for troubleshooting
-h, --help print help and exit
-V, --version print the version and exit
Every option takes --name value or --name=value. The ones that accept a list — intensity, clocks, fans — apply a single value to every selected card, and a list follows the order of --gpu-list, where * skips a card.
