NushyPool Miner

npminer is the NushyPool GPU miner. It is a Rust-based miner with CUDA and OpenCL backends and currently supports these algorithms:

• memhash for Vecno Foundation
• cryptix for Cryptix Network
• hoohash for Hoosat Network
• xelishash for Xelis Network
• capstash for CapStash Chain
• pearlhash for Pearl Chain

It supports standard Stratum pool mining and solo gRPC/RPC mining.

Key Features

• Low CPU usage during normal GPU mining.
• 0% devfee on supported NushyPool fee-exempt pools.
• Supports solo mining directly against a blockchain node over gRPC/RPC.
• Supports Stratum mining on any compatible pool.
• Supports Linux, Windows, HiveOS, and mmpOS.
• Supports NVIDIA GPUs through CUDA and AMD GPUs through OpenCL.
• Supports multiple algorithms in one miner: memhash, hoohash, cryptix, xelishash, and capstash.
• Includes autotune with cache reuse to keep tuned launch settings across restarts.
• Supports multi-GPU rigs with simple per-GPU selection via --devices.
• Includes watchdog-based worker supervision and recovery-oriented startup checks.
• Optional signed auto-update with --enable-autoupdate.
• Local stats API which expose read-only HTTP stats for HiveOS, mmpOS, and XMRig-compatible dashboards.

Caution

• Drivers: Always install the latest GPU drivers to achieve optimal hashrate performance.
• Pearl support is experimental. Optimal Pearl performance is currently available only on NVIDIA Ampere / RTX 30-series GPUs. Support for other GPU architectures is coming soon.
• --prl-cpu-cooldown is experimental and may affect Pearl hashrate or stale/late shares.

Credits

🙏 Special thanks to @ivchobow, @galindo9856, @MrGodzillaa166, and Mike

npminer

Main executable:

• Linux (inc. HiveOS/mmpOS): npminer
• Windows: npminer.exe

Typical pool mining command:

./npminer -a memhash -o stratum+tcp://POOL:PORT -u vecno:YOUR_WALLET -w worker1

Typical solo node mining command (gRPC/RPC):

./npminer -a hoohash -o grpc://127.0.0.1:42420 -u hoosat:YOUR_WALLET
./npminer -a hoohash --url grpc://127.0.0.1:42420 --user hoosat:YOUR_WALLET
./npminer -a xelishash -o rpc://127.0.0.1:PORT -u xel:YOUR_WALLET
./npminer -a capstash -o rpc://127.0.0.1:8332 -u cap1YOUR_WALLET --rpc-user capstashrpc --rpc-pass YOUR_RPC_PASSWORD

CapStash pool mining command:

./npminer -a capstash -o stratum+tcp://eu.capspool.io:7333 -u cap1YOUR_WALLET -w worker1

Useful GPU selection commands:

./npminer --list-gpus
./npminer -a memhash -o stratum+tcp://POOL:PORT -u cryptix:YOUR_WALLET -d 0,2

-d, --devices uses the GPU Index values printed by --list-gpus.

It does not use Bus ID.

Example --list-gpus output:

========================================
 Index | Bus ID | Name
========================================
 0     | 2      | GeForce RTX 3080 Ti
 1     | 3      | GeForce RTX 3080 Ti
 2     | 7      | Radeon RX 6800 XT
========================================

In that example:

• -d 0 means the first GPU
• -d 0,2 means the first and third GPUs
• -d 2 selects the GPU with Index = 2, not the GPU with Bus ID = 2

More examples:

./npminer -a hoohash -o grpc://POOL:PORT -u hoosat:YOUR_WALLET -d 1
./npminer -a memhash -o stratum+tcp://POOL:PORT -u vecno:YOUR_WALLET -d 0
./npminer -a memhash -o stratum+tcp://POOL:PORT -u vecno:YOUR_WALLET --devices 0,2,3
./npminer -a xelishash -o rpc://POOL:PORT -u xel:YOUR_WALLET --http-enabled --http-port 42330
./npminer -a capstash -o stratum+tcp://eu.capspool.io:7333 -u cap1YOUR_WALLET -w worker1 --opencl-disable
./npminer -a capstash -o stratum+tcp://eu.capspool.io:7333 -u cap1YOUR_WALLET -w worker1 --cuda-disable --opencl-enable

If --devices is not provided, all detected GPUs are used.

Supported Platforms and Hardware

Primary targets:

• NVIDIA GPUs through CUDA
• AMD GPUs through OpenCL

Current backend coverage:

Backend	Vendor	Algorithms	OS
CUDA	NVIDIA	memhash, hoohash, cryptix, xelishash, capstash, pearl	Windows, Linux, HiveOS, mmpOS
OpenCL	AMD	memhash, hoohash, cryptix, xelishash, capstash	Windows, Linux, HiveOS, mmpOS

Notes:

• OpenCL is limited to one OpenCL platform per miner process.
• --opencl-platform can be used when the default OpenCL platform is not the one you want.
• NVIDIA OpenCL is intentionally not the primary path; NVIDIA cards are expected to use CUDA.
• --devices is the recommended end-user way to select GPUs. Use --cuda-device and --opencl-device only when backend-specific selection is required.

Help Output

Print short CLI help with:

./npminer -h

Print detailed CLI help with:

./npminer --help

Use -h when you only need the option list. Use --help when you want the longer descriptions and examples for options such as --url, --user, --devices, and the HTTP API flags.

Command Reference

The current supported npminer command-line options are:

General

• -h, --help Print help.
• -v, --version Print version.
• -a, --algo <memhash|hoohash|cryptix|xelishash|capstash|pearl> Select algorithm explicitly.
• --mining-address <ADDRESS> Mining reward address.
• -u, --user <USER> Standard pool-style wallet field, for example vecno:... or vecno:....worker.
• --rpc-user <USER> JSON-RPC username for direct-node solo mining, currently used by CapStash RPC.
• --rpc-pass <PASSWORD> JSON-RPC password for direct-node solo mining, currently used by CapStash RPC.
• --list-gpus Print the miner GPU inventory and exit.
• -d, --devices <LIST> Comma-separated GPU Index values from --list-gpus, not Bus ID.
• --reset-autotune Delete cached autotune state and force a fresh retune.
• --enable-autoupdate Automatically download, verify, install, and restart into a newer npminer release when one is available.
• --prl-cpu-cooldown <ms> Reduce CPU usage during Pearl mining by adding a cooldown between GPU filter-event polls. 0 disables the cooldown and actively polls the GPU filter event. Higher values can reduce CPU usage but may reduce Pearl hashrate or increase stale/late shares. This option is Pearl-specific and experimental; use it at your own risk.

Stratum Pool and Solo Node Mining

• -o, --url <URL> Mining server URL, for example stratum+tcp://host:port for Stratum, grpc://host:port for Vecno-style solo mining, or rpc://host:port for XelisHash/CapStash direct-node solo mining.
• -w, --worker <WORKER> Standard worker name.
• -p, --pass <PASSWORD> Standard worker password.

Legacy compatibility Stratum options are also supported:

• --stratum-server <HOST>
• --stratum-port <PORT>
• --stratum-worker <WORKER>
• --stratum-password <PASSWORD>

API Reference

• --api-port <PORT> Starts the local HTTP stats API on the selected port.
• --api-bind <ADDR> Bind address for the API listener. The default is localhost.
• --api-token <TOKEN> Bearer token used to protect API responses. Required for non-local binds.
• --api-disable Disables the API, including when a mining OS package would normally enable it.
• --api-worker-id <ID> Overrides the worker ID shown in API responses.
• --api-id <ID> Overrides the miner instance ID shown in API responses.
• --http-enabled, --http-host, --http-port, --http-access-token XMRig-compatible aliases for monitoring software that already knows miner HTTP flags.
• --http-no-restricted Accepted for CLI compatibility. npminer does not expose write or control endpoints in v1.

CUDA Options

• --cuda-disable Disable CUDA workers.
• --cuda-device <LIST> Backend-specific CUDA device list.
• --cuda-workload <VALUE> CUDA workload ratio.
• --cuda-workload-absolute Treat CUDA workload as an absolute nonce count.
• --cuda-no-blocking-sync Disable blocking-sync mode and actively poll.
• --cuda-nonce-gen <lean|xoshiro> CUDA nonce generator.
• --cuda-lock-core-clocks <LIST> Optional GPU clock lock values.
• --cuda-lock-mem-clocks <LIST> Optional GPU memory clock lock values.
• --cuda-power-limits <LIST> Optional GPU power limits.

OpenCL Options

• --opencl-amd-disable Disable OpenCL AMD mining.
• --opencl-device <LIST> Backend-specific OpenCL device list on the selected platform.
• --opencl-platform <INDEX> Select one OpenCL platform.
• --opencl-workload <VALUE> OpenCL workload ratio.
• --opencl-workload-absolute Treat OpenCL workload as an absolute nonce count.
• --opencl-nonce-gen <lean|xoshiro> OpenCL nonce generator.
• --opencl-no-amd-binary Disable fetching/using the precompiled AMD OpenCL binary.

Notes on Option Behavior

• --devices is the recommended end-user selector and maps to the Index column from --list-gpus.
• --list-gpus prints both Index and Bus ID; --devices always uses Index.
• Do not combine --devices with --cuda-device or --opencl-device.
• If no device selection flags are provided, all detected GPUs are used.
• --debug exists in debug builds only and is not part of the normal release-user CLI surface.

Auto-Update

npminer checks the official downloads manifest at startup. If a newer version is available and --enable-autoupdate is not passed, the miner prints a warning and continues mining with the current version.

Enable automatic updates explicitly:

./npminer --algo hoohash --url stratum+tcp://POOL:PORT --user hoosat:YOUR_WALLET --worker Hashvilly --enable-autoupdate

When enabled, npminer downloads the matching release archive, verifies the archive sha256, size, and Ed25519 signature from downloads.json, installs the update into the current miner directory, and restarts itself. If the update check, download, verification, or install step fails, npminer keeps running the current version.

HiveOS and mmpOS packages use their own package markers so auto-update selects the matching HiveOS or mmpOS archive instead of the generic Linux archive. The install directory must be writable by the miner process.

Devfee

Algorithm	Stratum	Solo gRPC/RPC	NushyPool
memhash	1%	1%	0%
hoohash	1%	1%	0%
cryptix	1%	1%	0%
xelishash	1%	1%	0%
capstash	2%	2%	0%
pearlhash	2%	n/a	0%

NushyPool fee-exempt pool users mine with no dev fee.

CapStash direct RPC node mining uses the solo RPC devfee setting.

One-Line Installation

Replace <VERSION> with the release version you want, for example 0.0.1.

Windows

Download and extract the Windows package in one line from PowerShell:

$v="<VERSION>"; Invoke-WebRequest "https://github.com/nushypool/npminer/releases/download/v$v/npminer-windows-x86_64-v$v.tar.gz" -OutFile "npminer.tar.gz"; tar -xzf .\npminer.tar.gz

Linux

Download and extract the Linux package in one line:

VERSION=<VERSION> && curl -L "https://github.com/nushypool/npminer/releases/download/v${VERSION}/npminer-linux-x86_64-v${VERSION}.tar.gz" | tar -xz

Setup for HiveOS

Example HiveOS configurations

Hoohash	Vecno	Cpay

Notes

• --list-gpus is the reference output for end-user GPU selection.
• --devices applies the same GPU selection model across enabled backends.
• Do not combine --devices with --cuda-device or --opencl-device.
• If no GPU-selection flags are provided, the miner uses all detected GPUs.