Guide

How the dev fee works

What it costs, when it is paid, where it goes, and how to watch a whole cycle in a minute.

Some algorithms mine to the developer for a small share of the time. Each declares its own rate, it is in the algorithm table, and it is printed in the log at startup — so you can see it before committing a rig, not discover it later. This guide is everything about how it is paid.

What each algorithm charges

AlgorithmDev fee
Ergo ergo1%
KawPow (all five coins) kawpow1%
Ethash and etchash ethash0.5%
CryptoNight (every variant) cn1%
Pearl pearl1%
Quantus quantus2%
Tari c291%
RandomX (Monero, Zephyr, Salvium) randomx1%
VERUS verus1%
Warthog warthog2%
XELIS xelis1%
SHA-256d (the SDK example) sha256dnone

Run bzminer --list-algos on your own copy to check this against the build you have; it prints the algorithms and fees as JSON.

What it looks like

Every line about it is tagged [dev fee], and each slice announces both ends:

[dev fee] warthog 2% - announced here, and at START and END of every slice
[dev fee] START - mining for the warthog developer on stratum+ssl://ussw.vipor.net:5120 (2.00% -> 13 pools (split evenly))
[dev fee] END - back to your pool stratum+ssl://us.vipor.net:5120 (paid 1.94% of this run so far, aiming for 2.00% -> 13 pools (split evenly))

Dev-fee shares carry a worker name that says which rig paid them, and they stay out of your own share table.

When it is paid

As a time debt, not on a fixed clock. The first slice lands about five minutes into a run and the next an hour later, and the running total is what the schedule aims at: the log’s “paid 1.94% of this run so far, aiming for 2.00%” is that arithmetic. A restart resets the run, and the debt with it.

Where it goes

The fee is spread over several destinations rather than sent to one, so no single pool sees all of it. Some algorithms pay at their own coin’s pools; several — RandomX, KawPow, Ergo, Ethash and CryptoNight — pay at MoneroOcean, because it is one destination that takes the whole family and cannot be stranded by a single chain’s pool going away. That has a small cost on KawPow: a fee slice mines MoneroOcean’s current KawPow coin rather than yours, which means a DAG rebuild of a few seconds each way. The trade is a fee that is actually paid on every coin, including the ones whose own pools have gone.

You do not have to be able to reach them all. A destination that will not connect is backed off, and after three failed attempts it is dropped for the rest of the run — the others carry its share and the fee is still paid in full. If you block some of these pools at your firewall, the miner notices once and stops trying.

If every destination is unreachable

The fee is owed in time, not in shares. If every destination is unreachable, the slices still run: the rig keeps hashing and submits nothing for their duration, so the time is taken whether or not there is anywhere to spend it. Blocking the pools therefore costs you exactly what paying the fee costs you, and gains you nothing; unblocking any one of them turns that time back into shares for the developer instead of into nothing for anybody.

[dev fee] none of the 13 destinations could be reached. The fee is owed in time, not in
shares, so each slice is now taken as idle mining: the rig keeps hashing and submits
nothing for its duration. Unblocking any one of them pays it as shares instead.

A destination that is merely down is a different case and is treated as one: it is backed off and retried, and until it retires the debt simply accrues and is paid in full later. Nothing goes idle for a pool that is coming back.

What it costs your mining

None of it happens on the thread that reads your pool’s jobs, so an unreachable dev pool costs your mining nothing — not a stale job, and not a reconnect. A dev-fee connection also never validates the pool’s certificate; that is deliberate, because mining pools self-sign and let certificates lapse, and every one of those used to mean a slice silently not collected. The login still travels inside TLS.

On a farm proxy, each rig still pays its own fee, through a tunnel the proxy carries but cannot read or redirect, and the proxy signals when its own slice starts so the whole farm pays in one window.

Watch a whole cycle in a minute

The real schedule is correct and impossible to sit and watch. --test-dev-fee runs the cycle on a short clock so you can see the switch, the dev pool’s subscribe, the switch back and your own work resuming:

./bzminer --test-dev-fee 30 -a xelis -p ... -w ...     # 30 seconds each way
./bzminer --test-dev-fee 30:10 -a xelis -p ... -w ...  # 30 for you, 10 for the dev

It is a diagnostic, not a tuning knob: a split that would come to less than the declared fee has its dev leg raised until it does not, so the test cycle can only ever pay more.