Mine Pearl on the GPUs you already own.

Our miner is the fastest 0% Dev Fee, Free mining software available for Pearl. Works on AMD Radeon (RDNA 2/3/4) and every NVIDIA card with tensor cores — Volta (V100, CMP 100-210), Ampere (RTX 30 + A100), Ada (RTX 40 + L4/L40), Blackwell (RTX 50 + B100/B200), and Hopper (H100/H200).

Get Started → View blocks →

💰 0% Fee Free Miner

More shares per watt

The alpha-miner binary is a highly competitive and custom built for AlphaPool offering 0% Fees — more hashrate per kWh and more $PRL in your wallet!

💎 Lowest fees on Pearl

New Low 3% All-in Fee Active Now! Miner Loyalty Program Launching in the days ahead!

We are also getting ready to launch our AlphaPool Miner Loyalty Program which will offer all-in fees as low as 1.5%!

✅ H100 + Hopper — LIVE

~620 TH/s on H100

alpha-miner ships dedicated Hopper-arch kernels — the fastest known Pearl miner for H100/H200. Same binary covers Volta, Ampere, Ada, Blackwell.

Live pool stats

PRL Price — via SafeTrade

Hashrate & blocks

Pool Hashrate (24h)

∿

Pool hashrate timeline will populate as miners connect and submit shares.

Refreshes every 15 sec · Hourly buckets

Blocks Found (14d)

◇

We just opened the doors — be the first to find a block.

Bars show daily blocks found · Up to 14d window

Network & endpoints

Pearl Network

Stratum Endpoints

Look up your miner

No registration required. Your address is your account.

Recent blocks

Time	Height	Hash	Reward	Status
◐No blocks found yet. We just opened the doors — be the first to find one.

All blocks

Every block this pool has found, newest first. Statuses: Immature (< 100 confirmations) → Confirming (chain mature, awaiting pool payout cycle) → Confirmed → Paid. Orphans appear as Orphan.

Time	Height	Hash	Found by	Reward	Confirmations	Status
◐No blocks yet.

Top miners

Active miners ranked by 1-hour hashrate. Anonymized — each row is a PRL address; click to view its public stats.

#	Address	Workers	Hashrate (1h)	Shares (24h)	Last share
◐No miners online yet. Connect one.

Mining calculator

Estimate your PRL — and USD — earnings at the live network hashrate and PRL price. Pick your card or type a hashrate.

Estimate earnings PRL · loading…

Your hashrate

TH/s × cards

Or pick a card

Est. PRL / day

—

Est. USD / day

—

PRL / month

—

USD / month

—

Your share of the network: —. Estimated from the live network hashrate (via prlscan), the current block reward, and the 3% pool fee — mining is probabilistic, so real payouts vary with luck and network growth. PRL price via SafeTrade.

Help sharpen these estimates. Mining a card that's not listed, or seeing different numbers? Submit your measured hashrate — community data feeds the per-card figures this calculator uses.

Submit your hashrate →

Per-card hashrate

RTX 3060 Ti

~65TH/s

Ampere SM_86 · ~5× competing kernels on the same card

RTX 3070

~70TH/s

Ampere SM_86 · 8 GB VRAM

RTX 3090

~110TH/s

Ampere SM_86 · 24 GB flagship

RTX 5070 Ti

~150TH/s

Blackwell SM_120 · 16 GB

RTX 4090

~255TH/s

Ada SM_89 · 24 GB

RTX 5080

~185TH/s

Blackwell SM_120 · 16 GB

RTX 5090

~365TH/s

Blackwell SM_120 · 32 GB · flagship

H100

~620TH/s

Hopper SM_90 · 80 GB · datacenter

Measured against alpha-miner; actual hashrate varies ±10% with cooling, power limit, driver, and host CPU.

Get Started

Step 1

Get a Pearl address

Install the Pearl wallet and generate an address starting with prl1p….

Step 2

Run alpha-miner

One curl + one ./alpha-miner command — full setup in Quickstart below. On Windows, grab the native Windows installer.

curl -L -o alpha-miner https://pearl.alphapool.tech/downloads/alpha-miner

Step 3

Watch payouts arrive

First share in ~10 seconds. PPLNS payouts every 4 hours, 1 PRL minimum. Track live on the Dashboard by pasting your address.

Quickstart

# 1) Download the miner — source: https://github.com/AlphaMine-Tech/alpha-miner
# Linux x86_64, supports Volta/Ampere/Ada/Hopper/Blackwell auto-detect
curl -L -o alpha-miner https://pearl.alphapool.tech/downloads/alpha-miner
chmod +x alpha-miner

# 2) Replace YOUR_PRL_ADDRESS with your prl1p... wallet address, then run
# Recommended: pin a static difficulty for your card with --password "x;d=N" — it skips the
# vardiff warm-up and keeps your share rate steady from the first second. Pick your value in the Static difficulty tab.
./alpha-miner \
  --pool stratum+tcp://us2.alphapool.tech:5566 \
  --address prl1pYOUR_PRL_ADDRESS \
  --worker rig01 \
  --password "x

# AMD Radeon — RX 6000 (RDNA2), RX 7000 (RDNA3), MI300 (CDNA3). RX 9000 (RDNA4) coming soon.
# Bundles ROCm 7 user-space libs — no host ROCm install. Needs the amdgpu kernel driver +
# access to /dev/kfd and /dev/dri/renderD*. Source: https://github.com/AlphaMine-Tech/alpha-miner/releases/tag/amd-v1.0.0

# 1) Download + verify the AMD build
curl -L -o alpha-miner-amd https://github.com/AlphaMine-Tech/alpha-miner/releases/download/amd-v1.0.0/alpha-miner-amd
echo "4d240e0cecc133ee4c19d615d162ca589b1e653e0d23cf50d67922e1a1c70ef6  alpha-miner-amd" | sha256sum -c
chmod +x alpha-miner-amd

# 2) Replace YOUR_PRL_ADDRESS with your prl1p... wallet address, then run
# Recommended: pin a static difficulty with --password "x;d=N" — pick your value in the Static difficulty tab.
./alpha-miner-amd \
  --pool stratum+tcp://us2.alphapool.tech:5566 \
  --address prl1pYOUR_PRL_ADDRESS \
  --worker rig01 \
  --password "x

# Docker — works on any cloud GPU host with Docker + nvidia-container-toolkit.
# Tested on RunPod, vast.ai, Lambda Cloud, Paperspace. Image is ~600 MB, single layer cached after first pull.
# Public image: https://hub.docker.com/r/alphaminetech/pearl-miner
# IMPORTANT: use a STRATUM host (us2/eu1/sg1), not pearl.alphapool.tech for TCP mining.

# Minimal — auto-detects all GPUs, vardiff handles difficulty
docker run --gpus all \
  -e PEARL_ADDRESS=prl1pYOUR_PRL_ADDRESS \
  -e PEARL_WORKER=runpod-rig01 \
  -e PEARL_POOL_HOST=us2.alphapool.tech \
  -e PEARL_POOL_PORT=5566 \
  alphaminetech/pearl-miner

# Recommended for cloud rigs (static difficulty + run in background)
docker run -d --restart=unless-stopped --gpus all \
  --name pearl-miner \
  -e PEARL_ADDRESS=prl1pYOUR_PRL_ADDRESS \
  -e PEARL_WORKER=runpod-rig01 \
  -e PEARL_POOL_HOST=us2.alphapool.tech \
  -e PEARL_POOL_PORT=5566 \
  -e PEARL_DIFFICULTY=65536 \
  alphaminetech/pearl-miner

# Available env vars:
#   PEARL_ADDRESS      (required)  your prl1p... wallet address
#   PEARL_WORKER       (default: docker-rig)
#   PEARL_POOL_HOST    (recommended) us2.alphapool.tech / eu1.alphapool.tech / sg1.alphapool.tech
#   PEARL_POOL_PORT    (default: 5566)
#   PEARL_DIFFICULTY   (optional)  pin static diff — see Static difficulty tab for picker
#   PEARL_DEVICES      (optional)  e.g. "0,1,2" to pin specific GPUs
#   PEARL_STATUS_INTERVAL  (default: 60)  status log frequency in seconds

# View live logs
docker logs -f pearl-miner

# Stop / restart
docker stop pearl-miner
docker start pearl-miner

Native Windows installer — recommended. Runs directly on Windows 10/11 — no WSL2, no Docker. Pick the build that matches your GPU, then follow the 3 steps below.

NVIDIA GPU

RTX 20/30/40/50, T4, V100, A100, L4/L40, H100/H200, B100/B200, CMP

⬇ AlphaMiner-Pearl-Windows.zip (~3.3 MB)

contains alpha-miner-windows.exe + start-mining.bat

AMD GPU

Radeon RX 6000 / 7000 (RDNA2 / RDNA3), MI300 (CDNA3)

⬇ AlphaMiner-Pearl-AMD.zip (~0.75 MB)

contains alpha-miner-amd-windows-fixed.exe + start-mining-amd.bat

1. Download & extract the zip for your GPU. Right-click the .zip in File Explorer → Extract All → pick a folder you can find (Desktop, Documents, whatever). You'll end up with a folder containing two files: the miner .exe and a start-mining-*.bat launcher.

2. Edit the .bat with your wallet address. Right-click start-mining-nvidia.bat (or …-amd.bat) → Edit (or open with Notepad). Find this line near the top and replace the placeholder with your real Pearl address:

set PRL_ADDRESS=prl1pYOUR_PRL_ADDRESS_HERE

Save the file (Ctrl+S) and close Notepad. Optional: also edit WORKER (rig name) and DIFFICULTY (cheat-sheet inside the .bat).

3. Double-click the .bat to mine. A command-prompt window opens, the miner starts, you see shares flowing. If Windows SmartScreen warns "Windows protected your PC", click More info → Run anyway (we're not code-signed yet).

Prefer to run the miner manually (PowerShell)?

Extract the zip, then open PowerShell or Command Prompt, cd into the extracted folder, and run the miner directly — replace prl1pYOUR_PRL_ADDRESS with your wallet address:

cd into the folder:

cd Downloads\AlphaMiner-Pearl-NVIDIA

NVIDIA start command:

.\alpha-miner-windows.exe --pool stratum+tcp://us2.alphapool.tech:5566 --address prl1pYOUR_PRL_ADDRESS --worker rig01 --password "x"

AMD start command:

.\alpha-miner-amd-windows-fixed.exe --pool stratum+tcp://us2.alphapool.tech:5566 --address prl1pYOUR_PRL_ADDRESS --worker rig01 --password "x"

The x;d=65536 sets static difficulty — works fine for most cards. Pick your card's recommended value on the Static difficulty tab, or omit --password to let the pool auto-tune.

Updating? Re-download the zip and overwrite the existing folder. NVIDIA driver: update here (R545+ recommended). AMD driver: bundles ROCm 7 user-space — needs the amdgpu driver from your distro / AMD website.

Alternative: Docker Windows installer

Prefer a guided installer that manages everything for you (auto-updates, plus desktop Start / Stop / Logs / Reconfigure shortcuts)? This wrapper runs the same Linux image via Docker Desktop's WSL2 backend.

⬇ Download AlphaMiner-Setup-1.7.7.exe (161 KB)

For Windows 10/11 + NVIDIA GPU. Wraps the same alphaminetech/pearl-miner:1.7.7 image used on Linux/HiveOS, running via Docker Desktop's WSL2 backend. One-click install — wallet, worker, and region prompts during setup. Change region anytime via the Reconfigure AlphaMiner shortcut. Start / Stop / Logs / Reconfigure shortcuts on the desktop.

What gets installed

A configured docker-compose.yml in %USERPROFILE%\AlphaMiner\
Desktop shortcuts: Start AlphaMiner, Stop AlphaMiner, AlphaMiner Logs, Reconfigure AlphaMiner
Reconfigure lets you change wallet, worker, region, or port at any time without uninstalling
Start Menu folder + Add/Remove Programs entry for clean uninstall
Pre-pulled alphaminetech/pearl-miner:1.7.7 image (~2 GB on first install)

Requirements

Windows 10 (build 19045+) or Windows 11, 64-bit
NVIDIA GPU with CUDA driver 545+ — update here
Docker Desktop for Windows (installer offers to download if not present)
Hardware virtualization enabled in BIOS (Intel VT-x / AMD-V)

First-run note: Windows SmartScreen shows "Windows protected your PC" because the installer isn't code-signed yet — click More info → Run anyway. Code signing is on the roadmap.

SHA256: 5bae4f9fc6815da328abca633704436150955aceadaa60771e1ecf11990e29bf

# Solo mining is temporarily paused for re-integration.
# We're rebuilding it in a more fitting location.
# Pooled (PPLNS) is unaffected — use the Linux/Docker tabs above.
#
# Want a heads-up when solo is back online? Ping us in Discord.

# Static difficulty — recommended if vardiff isn't ratcheting cleanly for your rig.
# Format: --password "x;d=<DIFFICULTY>"   (open StratumX convention — see stratumx.org)
#
# Pick a value matched to your card class:
#   V100 / CMP 100-210     →  4096
#   RTX 2070 / 2080                    →  16384
#   RTX 3060 Ti / 3070                      →  131072
#   RTX 3080 / 3090 / CMP 70HX/90HX         →  262144
#   A100 / data-center Ampere               →  131072
#   RTX 4070 / 4080                     →  262144
#   RTX 4090 / 5080                     →  524288
#   RTX 5090 / H100 / H200 / B100        →  1048576
#
# Wrong difficulty just means slightly bumpier stats — you don't lose shares.
# If your rig is >100 GH/s combined and vardiff is climbing too slowly, this is the fix.
./alpha-miner \
  --pool stratum+tcp://us2.alphapool.tech:5566 \
  --address prl1pYOUR_PRL_ADDRESS \
  --worker rig01 \
  --password "x

# All CUDA devices are auto-detected by default. To pin to specific GPUs:
./alpha-miner \
  --pool stratum+tcp://us2.alphapool.tech:5566 \
  --address prl1pYOUR_PRL_ADDRESS \
  --worker rig01 \
  --devices 0,1,2 \
  --status-interval 60

# List available CUDA devices first:
./alpha-miner --list-devices

# Force a specific backend (rarely needed — auto-detect handles this):
#   --force-backend ampere|ada|blackwell|volta

# Full options:
./alpha-miner --help

# /etc/systemd/system/alpha-miner.service
[Unit]
Description=AlphaPool Pearl miner
After=network-online.target

[Service]
Type=simple
User=miner
WorkingDirectory=/home/miner
ExecStart=/home/miner/alpha-miner \
  --pool stratum+tcp://us2.alphapool.tech:5566 \
  --address prl1pYOUR_PRL_ADDRESS \
  --worker %H \
  --status-interval 60
Restart=on-failure
RestartSec=10s

[Install]
WantedBy=multi-user.target

# Enable + start:
# sudo systemctl daemon-reload && sudo systemctl enable --now alpha-miner
# Logs: sudo journalctl -fu alpha-miner

Pick the endpoint closest to you for lowest share latency.

🇺🇸N. America · Eastus1.alphapool.tech:5566

🇺🇸N. America · Westus2.alphapool.tech:5566

🇪🇺Europeeu1.alphapool.tech:5566

🇪🇺Europeeu2.alphapool.tech:5566

🇷🇺Russia · Eurasiaru1.alphapool.tech:5566

🇸🇬Asia · Singaporesg1.alphapool.tech:5566

🇮🇳Indiain1.alphapool.tech:5566

Recommended: pin a static difficulty from the first run with --password "x;d=N" — it skips the vardiff warm-up and gives steady shares immediately (and is the safe choice for multi-rig HiveOS / mixed-card setups). Pick your card's value from the table below. Prefer hands-off? Leave the password as x and vardiff auto-tunes for you instead.

Static difficulty by card class

Add --password "x;d=<DIFFICULTY>" to your command line (or set PEARL_DIFFICULTY=<N> for Docker) to pin a static share difficulty. Recommended for every setup — it skips the vardiff warm-up and keeps your share rate steady from the first second; especially important for multi-GPU rigs and HiveOS. Wrong difficulty just means bumpier stats — you don't lose shares.

Card class	--password "x;d=…"
V100 / CMP 100-210	4096
RTX 2070 / 2080	16384
RTX 3060 Ti / 3070	131072
RTX 3080 / 3090 / CMP 70HX/90HX	262144
A100 / data-center Ampere	131072
RTX 4070 / 4080	262144
RTX 4090 / 5080	524288
RTX 5090 / H100 / H200 / B100	1048576

Before you start

If your rig already mines any GPU coin, the one-liner below just works. The binary is statically linked — its only runtime dependency is the NVIDIA driver (already present on any mining rig).

NVIDIA GPU: Turing through Blackwell + Hopper — RTX 20/30/40/50, T4, V100, Titan V, A100, L4/L40, H100/H200, B100/B200, CMP. Pre-Volta Pascal (GTX 10-series) lacks tensor cores and isn't supported.
AMD GPU: Radeon RX 6000/7000 (RDNA2/RDNA3) and MI300 (CDNA3) via the dedicated AMD build — use the AMD (Radeon) tab below. RX 9000 (RDNA4) coming soon. Bundles ROCm 7 user-space (no host ROCm); needs the amdgpu driver.
OS: Linux x86_64 (Ubuntu 14.04+, Debian 8+, RHEL 7+, glibc 2.18+), or Windows 10/11 — run the native Windows installer (v1.0.0, no WSL2/Docker), the Docker Windows installer, or WSL2; see the Windows tab below. Verified on Windows 11 with NVIDIA RTX cards.
Renting cloud GPUs? Use the Docker / RunPod / vast.ai tab — one docker run command, works anywhere with NVIDIA Docker.
Driver: Any reasonably modern NVIDIA driver (R470+ for Ampere, R525+ for Ada, R550+ for Blackwell). Already installed on any mining rig.
Network: Outbound TCP to us2.alphapool.tech:5566. No inbound ports needed.
VRAM: ~3 GB per GPU. ~36 MB disk for the binary.

What to expect after you start

~30 seconds: your miner connects and starts submitting shares. Difficulty auto-tunes over the first few minutes — early numbers will look jumpy, that's normal.
First accepted share: your address appears on the Dashboard. Paste it into Look up your miner to watch your worker go live.
Healthy log line reads like accepted share … diff …. If you see that, you're mining.
First payout: typically 24–48h, once you clear the 1 PRL minimum. Pool hashrate may show "—" for the first hour while data accumulates.

Hardware support

RTX 30-series

3060–3090 Ti

Ampere · ✓ supported by alpha-miner binary

RTX 40-series

4060–4090

Ada · ✓ supported by alpha-miner binary

RTX 50-series

5060 Ti–5090

Blackwell · ✓ supported by alpha-miner binary

A100 / L4 / L40 · CMP 70HX/90HX/170HX/220HX

Ampere/Ada (consumer + datacenter)

✓ supported by alpha-miner binary (force --backend ampere)

H100 / H200 / B100 / B200

Hopper / Blackwell DC

✓ supported by alpha-miner binary (Hopper SM_90 + Blackwell kernels in current build)

V100 / Titan V / CMP 100-210

Volta (SM_70)

✓ supported by alpha-miner binary (since 2026-05-15)

RTX 20-series · T4 · CMP 30/40/50HX

Turing (SM_75)

✓ supported by alpha-miner binary

AMD Radeon · RX 6000/7000 · MI300

RDNA2 / RDNA3 / CDNA3

✓ supported via the dedicated AMD build (Get Started → AMD tab). RX 9000 (RDNA4) coming soon. Bundles ROCm 7 user-space — no host ROCm install.

Windows · Linux · HiveOS

✓ supported today

Windows users run the native Windows installer (v1.0.0 — no WSL2/Docker), the Docker Windows installer, or WSL2 — see Get Started → Windows. Verified on RTX 30/40/50 series + A100/H100. AMD Radeon now supported; macOS on the roadmap.

Pre-Volta cards (Pascal GTX 10-series and older) lack the tensor cores Pearl's NoisyGEMM PoW needs. CPU mining is not supported. Need help? Hit our Discord — real humans.

FAQ

Quick answers about Pearl mining and the AlphaPool setup.

Payout structure

Time-weighted PPLNS — payouts proportional to recent sustained contribution

Whenever the pool finds a block, the reward is distributed across miners proportionally to a time-weighted score of their recent valid shares. Every share you submit adds to your score, and that contribution decays smoothly over time. Your payout share on any given block reflects your accumulated, decay-weighted hashrate at that moment.

How the decay works

Half-life is 65 minutes
A share counts for 100% when fresh, 50% after 65 min, 25% after 2h 10m
Older contributions fade gradually — no hard cutoff

What this means for you

Steady mining is rewarded — payouts track your sustained hashrate
Brief disconnects barely affect your share; prior contribution lingers
Bursts don't get over-credited at the expense of steady miners

To audit your share: look up your address (miner search, top bar) and compare your reported hashrate to your block payout fraction. Over a long horizon they will converge — your payout share equals your hashrate share of the pool.

What is Pearl?

Pearl (PRL) is a Layer-1 blockchain that uses Proof-of-Useful-Work — mining is performed as a by-product of arbitrary matrix multiplication. The protocol is open-source under the ISC license at github.com/pearl-research-labs/pearl. Mainnet launched 2026-04-27.

What GPUs does this pool support?

Supported by the current alpha-miner binary: Turing (RTX 20-series, T4, CMP 30HX/40HX/50HX), Volta (V100, Titan V, CMP 100-210), Ampere (RTX 30-series, A100, CMP 70HX/90HX/170HX/220HX), Ada (RTX 40-series, L4/L40), Hopper (H100, H200), Blackwell (RTX 50-series, B100/B200). The miner uses ~3 GB of VRAM and auto-detects all CUDA devices.

Not supported: Pre-Volta (Pascal GTX 10-series and older) lacks the tensor cores Pearl's PoW needs. CPU mining is not supported.

Can I mine on rented cloud GPUs (RunPod, vast.ai, Lambda)?

Yes — pull our Docker image and pass your address as an env var. See the Docker / RunPod / vast.ai tab for the full command. The image (alphaminetech/pearl-miner) bundles the alpha-miner binary on top of nvidia/cuda:12.4.1-runtime, so it works on any host with Docker + the NVIDIA container toolkit. For RunPod specifically: create a Custom Template, set the image to alphaminetech/pearl-miner, leave Container Start Command empty, and add PEARL_ADDRESS + PEARL_WORKER as environment variables.

If you're running cloud arbitrage (renting cheap GPU-hours below their PRL/USD payout value), set PEARL_DIFFICULTY to skip the vardiff ramp-up time — instant steady-state share rate.

Can I mine on Windows?

Yes — natively. Download the native Windows installer (alpha-miner-windows.exe) — no WSL2 or Docker required — and run it from PowerShell or Command Prompt with your wallet address (same flags as Linux; full steps in Get Started → Windows). Prefer a guided setup? The Docker Windows installer (desktop Start / Stop / Logs shortcuts, auto-updates) and plain WSL2 both work too. Windows SmartScreen may warn that the app isn't code-signed yet — click More info → Run anyway.

When should I set a static difficulty (x;d=N)?

The pool runs an automatic VarDiff loop that retunes your difficulty to ~1 share per 10 seconds per worker. The x;d=N static-diff override follows the open StratumX convention used by many other pools — if you've used it elsewhere, it works the same here. VarDiff works well for most setups, but a few cases benefit from pinning a static difficulty in your password field:

HiveOS rigs that restart frequently — vardiff resets on each reconnect and never reaches steady state.
Mixed-card multi-rig setups sharing one stratum connection — vardiff is per-connection and gets confused by uneven hashrate.
Very high-hashrate cards (4090, 5080, H100) — vardiff can take ~7 minutes to ratchet up to the right value; a static diff is instant.

Format: --password "x. Pick a value from the table in the Static difficulty tab. Wrong picks aren't dangerous — slightly bumpier stats, no lost shares.

What's the pool fee?

3% on PPLNS. One of the lowest fees on the Pearl network. Fees subsidize operating cost (US East server, 24/7 ops); any surplus goes to upstream protocol development.

How does PPLNS work here?

AlphaPool uses time-weighted PPLNS with a 65-minute half-life — see Payout structure at the top of this page for the full breakdown. In short: each block reward is split in proportion to every miner's recent valid shares, weighted so fresh shares count fully and older ones decay smoothly. Steady mining is rewarded and brief disconnects barely dent your share.

What's the minimum payout?

1 PRL. Payouts run every 4 hours; balances below the floor roll over until cleared.

How long until I see my first payout?

First share usually arrives within 10 seconds of starting your miner. First payout typically lands within 24–48 hours of starting if you have meaningful hashrate. Smaller miners may take longer to accumulate the 1 PRL minimum.

Why is the pool hashrate sometimes shown as "—"?

AlphaPool opened 2026-05-14 (Pearl mainnet launched 2026-04-27). Hashrate displays go live once we have enough share data to compute a stable estimate — usually within the first hour of mining.

Is the source code open?

The pool server is internal for now. The Pearl protocol itself is fully open-source. The integration spec for connecting a custom miner will be published once stable.

I have a question or issue — who do I talk to?

Hit the AlphaPool Discord / Telegram (links on the main site). Real humans, no bots.

← Back to pool

Miner dashboard

—

All pool stats →

Statistics overview

Miner hashrate (24h)

Activity

Hide idle workers (offline or 0 H/s) · tap any column header to sort (name uses natural order: rig2 < rig10)

Worker	Hashrate (3 min)	Hashrate (1h)	Hashrate (24h)	Last share	Difficulty	Status
⚙No workers reporting for this address yet.

Time	Amount	TxID	Status
⌑No payments sent yet.

Height	Time	Hash	Reward	Your share	Mode	Status
◐No blocks contributed to yet.

Payment history (14d)

GPU Performance

Hashrates (TH/s), recommended settings, and version-over-version improvements for Pearl mining.

Last updated: — Entries: — Source: AlphaMine

Running a card we don't list — or seeing different numbers? Add yours to sharpen the calculator's estimates. + Submit your hashrate

Click any column header to sort ↕

GPU	Arch	VRAM	v1.6.x	v1.7.2	v1.7.3	v1.7.4	Δ vs v1.6	Power	Diff	Core / Mem OC	Notes
Loading…

Static difficulty values are recommended starting points. Small rigs without shares in the first minute should try the next lower value from the Static Difficulty tab. too-easy share rejects = go higher.

Recommended clocks are conservative starting points. Your specific silicon will vary — adjust in 50 MHz steps and watch reject rate. Mem OC matters more than core OC for Pearl's NoisyGEMM PoW.

Versions: v1.6 added Blackwell tensor paths; v1.7.2 introduced power-tuning; v1.7.3 fixed a duplicate-share corner case; v1.7.4 (current) is the revised production build with measurable per-arch gains.