Mining

ZK

RTX 5090 · ZK v5 · 148M nonces/s at 600 W

RTX 5090 / day
—
Share rate
—
Validation avg
—
Blocks found
—
Block build avg
—
Block build max
—

AI

RTX 5090 · 344 TMAC/s at 600 W

RTX 5090 / day
—
Share rate
—
Validation avg
—
Blocks found
—
Block build avg
—
Block build max
—

Pool status

Workers online

—Connecting

ZK pool hashrate

—Five-minute share estimate

AI pool hashrate

—Five-minute share estimate

Accepted shares (24h)

—Validation status

Pending payouts

—Available balance

Paid to date

Confirmed payouts

24-hour hashrate

ZK

nonce/s

AI

PMAC/s

Nockchain status

NOCK price

—Live NOCK/USD market price

Blockchain height

—Latest healthy server report

ZK network hashrate

—Current difficulty · 500 s target

AI network hashrate

—Current difficulty · 214 s target

Mining servers

Loading server heartbeats…

ServerNode heightZKAIHeartbeat
Loading mining servers…

Recent pool blocks

Loading pool block candidates…

HeightPuzzleStatusBlockWorkerSubmitted
Loading recent blocks…

Recent payouts

Loading payout batches…

BatchStatusRecipientsAmountTX FEETransactionUpdated
Loading recent payouts…

Leaderboard

ZK

ZK miners

RankMinerHashrateShare workBlocksOnline
Loading ZK miners…
AI

AI miners

RankMinerHashrateShare workBlocksOnline
Loading AI miners…
NvidiaZK: NVIDIA Pascal+
AI: NVIDIA Ampere+
Ubuntu 20.04+
AMDZK: Radeon RX 6000 / 7000 / 9000 series
AI: N/A
Ubuntu 22.04+
WalletInstall Iris Wallet
Create a NOCK payout address
02

Make it executable

chmod a+x gigahash-zk-12.9 gigahash-zk-rocm10.0 gigahash-proxy
03

Start mining

The worker name defaults to the machine hostname.

ZK
./gigahash-zk-12.9 --payout-address YOUR_NOCK_ADDRESS
AMD ZK
./gigahash-zk-rocm10.0 --payout-address YOUR_NOCK_ADDRESS
04

Use the proxy

One proxy can serve both ZK and AI workers through a shared local endpoint.

PROXY
./gigahash-proxy --payout-address YOUR_NOCK_ADDRESS
ZK VIA PROXY
./gigahash-zk-12.9 --proxy 127.0.0.1 --payout-address YOUR_NOCK_ADDRESS

--proxy uses port 9200 when no port is given. For workers on other machines, replace 127.0.0.1 with the proxy host and restrict TCP 9200 to trusted workers. Use the same payout address on the proxy and miners.

05

Optional settings

--worker-name NAME
Set the name shown on your account page
--server HOST[:PORT]
Override pool server; default port 9100 for ZK and 9101 for AI
--backup-server HOST[:PORT]
Override fallback servers; used only when no normal server is available
--proxy HOST[:PORT]
Connect through a proxy; defaults to port 9200
--devices 0,1
Mine on selected GPU devices only
--instances N
Override automatic instance count
--low-cpu
ZK only: one mining lane per GPU for the lowest host CPU usage
--listen HOST:PORT
Set the proxy endpoint exposed to workers
06

Changelog

  • v3.7 ZK: 6.5% gain for Blackwell, 4.5% for Ada, 3% for Ampere, 10% for Turing, Pascal / V100 support (6.1+)
  • v3.6 ZK: 15% gain for Blackwell, 9% for Ada, 7% for Ampere, 5% for Turing, 18% for A100/H100, 28% for RX 9000, 17% for RX 7000, 22% for RX 6000, bug fixes, reduced peak VRAM
  • v3.5 ZK: 5% gain for Nvidia, 3 to 6% gain for AMD
  • v3.4 ZK: 12% gain for Blackwell, 20% for Ampere + Ada, 90% for Turing
  • v3.3 ZK: Fixed segfault at startup (for those cases)
  • v3.2 ZK: 100% gain for Nvidia
  • v3.1 ZK: 80% gain for Nvidia
  • v3.0 ZK: update for new ZK PoW v5 algo (CUDA only), activated at height 147,500
  • v2.7 ZK: ~55% gain for Blackwell, ~65% for Ada and ~75% for Ampere
  • v2.6 ZK: fixed device telemetry mismatch for CUDA & AMD
  • v2.5 ZK: up to 40% gain for AMD, added --low-cpu mode for AMD as well
  • v2.4 ZK: another ~20% gain for Nvidia, 44k proofs/s on 5090 at 575W
  • v2.3 ZK: ~15% gain for Nvidia, 37k proofs/s on 5090 at 575W
  • v2.2 ZK: added Radeon RX 6000 support, updated to ROCm 10
  • v2.1 ZK: added AMD support for Radeon RX 7000 and 9000 series GPUs
  • v2.0 ZK: lower RAM usage, less PCIe traffic, --backup-server option
  • v1.9 ZK: added --low-cpu mode which reduces CPU load to 1/4, at the cost of hashrate
  • v1.8 ZK: reduced CPU load by 20%, 1% gain on 5090, tuned VRAM usage
  • v1.7 HiveOS support, --proxy parameter
  • v1.6 Improved AI by 10% for Blackwell, 30% for Ada, 25% for Ampere, now supports PCIe x1 mining (AI only)
  • v1.5 Improved pool connection failover and status probing
  • v1.4 Added automatic ZK/AI mining and Docker images
  • v1.3 Added GPU fault isolation
  • v1.2 Improved DNS discovery and reconnect handling

ZK Hashrate

Measured pool worker average hashrate / power limit

GPU ModelHashratePower
RTX 5090136,846,552 proof/s560 W
RTX PRO 5000 BLACKWELL87,063,129 proof/s300 W
RTX 409084,806,857 proof/s425 W
L40S82,110,345 proof/s350 W
RTX 5070 TI57,653,315 proof/s271 W
RTX 4070 TI SUPER53,857,362 proof/s294 W
RTX 507045,827,985 proof/s250 W
RTX 4070 SUPER41,175,772 proof/s197 W
RTX 3080 TI38,042,266 proof/s349 W
RTX 309036,408,956 proof/s324 W
RTX 5060 TI27,814,462 proof/s114 W
AMD RADEON RX 7900 XTX20,342,265 proof/s290 W
AMD RADEON AI PRO R970020,112,276 proof/s299 W
RTX 3060 TI17,607,326 proof/s217 W
RTX 4060 TI15,858,954 proof/s81 W
AMD RADEON RX 9060 XT14,333,634 proof/s170 W
RTX 306012,702,071 proof/s151 W
AMD RADEON PRO W68008,758,196 proof/s192 W
RTX 508037,310 proof/s324 W
RTX 4080 SUPER18,726 proof/s291 W
RTX 308010,765 proof/s288 W
RTX PRO 4000 BLACKWELL9,628 proof/s145 W
RTX 30707,396 proof/s196 W
RTX A40004,953 proof/s137 W