How it works

Mining that does useful work.

Pearl's Proof-of-Useful-Work makes every "hash" a noised matrix multiplication — the same math that runs every LLM. That is why one GPU fleet can mine PRL, serve inference, and back a yield market.

You point a miner at the pool

Any rank-128-ready PRL miner connects over standard pearl/v1 stratum — two minutes, one command. H100/H200 operators can additionally run the official vLLM dual-use miner, which mines PRL as a by-product of serving LLM inference; upstream it is solo-only, and our share-gateway connector for it is in beta. See Connect.

Your GPU computes NoisyGEMM

The proof-of-work is a noised matrix multiply: C = (A + noise)(B + noise)ᵗ, with the noise derived deterministically from the block template. Finding a share means finding a computation whose transcript hash beats the share target — there is no separate "hash function" bolted on; the matrix math is the work.

The same math can serve inference

Because the mined GEMMs are real matrix multiplications, inference work can ride the same fleet. That is the basis of our useful-work router and the coming inference-token market: customers submit requests, the router splits them into GEMM tiles, miners compute them, customers get results — and contributors earn a revenue share on top of mined PRL.

The pool accounts every share & pays you (PPLNS)

Shares are recorded in a crash-proof, idempotent ledger — no double-pays, ever. Payouts go out on the standard PPLNS schedule to your PRL address once you pass the 1 PRL minimum.

Miner revenue = mining reward (PPLNS shares).
Pool revenue = 1.5% pool fee. Nothing hidden.

The useful-work architecture

How an inference request and a mining share flow through the same system.

Customer → POST /inference → Useful-Work Router → tiles/next → Miner (NoisyGEMM) → tiles/submit → Routercustomer gets result
In parallel: the miner stays connected to stratum :5567 for mining shares.
Miner earns: mining rewards via PPLNS (+ inference revenue share when the token market opens).
Pool earns: 1.5% fee.

August 2026 protocol changes

Two network upgrades changed what a valid Pearl proof looks like. If your miner predates them, it will look healthy locally while every share is rejected.

Dense-only certificates (block 91,630)

Since block 91,630 the network accepts only dense-model certificates; MoE (mixture-of-experts) certificates are rejected. Miners and models built for the earlier MoE fork now produce invalid work.

Rank penalty (block 96,251)

Since block 96,251 the noise rank must be at least 128, and ranks above 128 are penalized proportionally: the share/block bound is scaled by 128/rank, so rank 256 is twice as hard for the same work. In practice, rank 128 is the operating point. Miners that used higher ranks lost 50-90% of their effective hashrate overnight, which is why every pre-upgrade benchmark table is obsolete.

Consequence: use a rank-128-ready miner version (see the supported miner list), and treat any per-GPU hashrate figure published before August 2026 as unverifiable.

FAQ

What does it cost? What's the fee?
1.5% all-in pool fee — 1.0% base + 0.5% on-chain tx reserve. PPLNS payout, 1 PRL minimum payout. The dashboard shows the live pool fee straight from the pool API — nothing hidden.
Which GPUs and miners are supported?
Any GPU your miner software supports — NVIDIA RTX 2000-series and newer, AMD RDNA2, Intel Arc, and datacenter H100/H200 — mining the pearlhash algorithm over standard stratum. Use a rank-128-ready miner version; the Connect page lists the supported miners and minimum versions. We do not publish per-GPU hashrate tables: the August 2026 rank-penalty upgrade reset all published benchmarks and post-upgrade numbers are still being re-measured.
What is the official vLLM dual-use miner?
Pearl's official miner is a vLLM plugin for H100/H200: it serves real LLM inference and the mined NoisyGEMM rides inside the model's own matrix multiplications, so PRL is produced as a by-product of serving. Upstream it can only mine solo (full blocks, no shares). Our share-gateway connector, which lets it submit pool shares instead, is in beta — contact weare@linkrra.com.
How and when do I get paid?
PPLNS — you're credited per accepted share, weighted across the payout window, and paid to your PRL address once you pass the minimum-payout threshold. Every credit and payout is in the durable ledger.
Is my balance safe? What about double-pays?
The pool's money ledger is an append-only, atomic SQLite store; payouts are idempotent on txid — a crash or retry can never pay you twice or lose a balance. An independent security audit gates real-money launch.
What are "inference tokens"?
The coming second track of the marketplace: customers buy LLM inference served on the pooled GPU fleet through an OpenAI-compatible API (credits with Stripe billing are already built). GPU contributors receive a share of that inference revenue in addition to their mined PRL. Not live yet — register interest.
How do I connect?
See the Connect guide — copy the stratum snippet, drop in your PRL address, and you're earning. No registration, no login.
Connect your rig →