Key takeaways
  • Hashrate measures attempts per second.
  • Pool-side hashrate is estimated from submitted shares.
  • Read hashrate together with efficiency, rejects and uptime.

Hashrate is a rate of attempts

Hashrate describes how many hashing attempts a miner or the network can perform per second. Mining hardware is commonly described in terahashes per second, while large pools and the network are often discussed in petahashes or exahashes per second.

Because proof of work is probabilistic, hashrate is best understood as capacity rather than guaranteed output. More hashrate raises expected share of work over time.

Why the dashboard and pool can disagree

The miner dashboard estimates performance from the device itself. A pool estimates effective performance from valid shares received over a time window. These numbers can differ because of normal statistical variance, rejected shares, stale work or connectivity problems.

For troubleshooting, compare longer time windows before concluding that a machine is underperforming.

Accepted shares are the practical signal

High local hashrate with poor share acceptance can indicate networking, overclocking or hardware instability. Pool-side accepted work is therefore an important operational metric.

Rejected and stale shares should be investigated if they remain elevated. Common causes include unstable tuning, packet loss, high latency and incorrect pool settings.

Use hashrate with efficiency and uptime

A fleet with slightly lower hashrate but better efficiency and uptime can be economically stronger than a faster fleet that consumes more power or experiences frequent resets.

Hashrate answers only one question: how much work is being attempted. It should always be read alongside watts, joules per terahash, temperatures and availability.

Bottom line

A practical explanation of hash rate units, pool-side measurements and why short-term numbers can be noisy. Use the principles above as a repeatable operating framework, then validate decisions with current power pricing, network conditions, equipment data and local electrical requirements.