Definition
Hashrate is the speed at which a mining device performs the SHA-256 hashing algorithm, measured in hashes per second. It is the single most important performance metric for any Bitcoin mining machine, because every hash is one attempt to find a block solution.
Also known as: hash power, computing power, hashing speed.
Common units climb in factors of 1,000: H/s, KH/s (kilohash), MH/s (megahash), GH/s (gigahash), TH/s (terahash), PH/s (petahash), and EH/s (exahash). A modern home ASIC like the Bitaxe Hex Supra produces a few TH/s, while a full data-floor machine pushes past 100 TH/s. The combined output of every miner on the network is the network hashrate, today measured in hundreds of EH/s.
Where hashrate actually comes from
On an ASIC, hashrate is the product of three physical things: the number of hash cores on each chip, the number of chips wired into series on a hashboard, and the clock frequency those chips run at. More cores or a higher frequency means more nonce attempts per second. For example, Bitmain’s BM1366 chip (used in the S19 XP, S19k Pro, and the Bitaxe Hex Supra) carries 894 small SHA cores, while the newer BM1370 in the S21 Pro carries 1,280 cores. Because the BM1370 packs more cores per chip, it can hit the same hashrate at a slightly lower clock frequency, which is exactly how newer silicon improves efficiency (J/TH).
This is also why hashrate and power draw are inseparable. Pushing a chip to a higher frequency raises its hashrate but increases voltage and heat super-linearly. Firmware autotuners exist precisely to find the best hashrate a given chip can sustain inside a fixed power budget without producing hardware errors. That tuning is calculated at runtime against each chip’s actual silicon quality, not read from a fixed table.
Nominal vs. effective hashrate
The number printed on a miner’s spec sheet is its nominal (nameplate) hashrate. The number that matters for earnings is the effective hashrate the machine actually delivers over time. The two can diverge for real, physical reasons:
- Hardware errors: Hashes that fail the chip’s internal checks are wasted work and never count toward valid output.
- Thermal throttling: A hot hashboard or a failing fan forces firmware to cut frequency to stay inside safe limits, dropping hashrate.
- Dead or degraded chips: A single bad chip in a series chain can knock an entire board offline, and a partially failing board reports lower-than-nameplate output.
- Statistical variance: Hashing is a random search, so short-term hashrate readings bounce around the long-term average even on a perfectly healthy machine.
This is why pools report a separate “pool-side” or accepted hashrate, estimated from the difficulty of the shares your miner submits over time rather than from what the machine claims locally. If your dashboard shows full nominal hashrate but your pool-side number is consistently lower, that gap is a diagnostic signal: it usually points to rejected shares, a flaky connection, or a board that is hashing but not landing valid work.
Why hashrate drives mining economics
A higher hashrate increases the probability of finding a valid block hash below the network target and earning the proof-of-work reward. But raw hashrate alone tells you nothing about whether a machine is worth running. Two miners can produce identical hashrate while one consumes far more electricity, so operators always pair hashrate with efficiency to judge real-world profitability. When you evaluate hardware, think of hashrate as the revenue side of the equation and J/TH as the cost side.
A practical consequence for anyone diagnosing a sick machine: a sudden, unexplained drop in hashrate is one of the most common early symptoms of a hashboard, PSU, or cooling fault. Reading the per-board and per-chip hashrate breakdown in your firmware is often the fastest route to isolating which board has failed.
If you are sizing a new deployment or trying to recover lost hashrate from an underperforming rig, D-Central’s ASIC troubleshooting resources walk through the hardware-level causes, and the miners catalog lists per-model nominal hashrate and efficiency so you can compare machines on the two numbers that actually decide whether a miner earns.
In Simple Terms
The speed at which a mining device performs calculations. More hashrate equals more mining power.
Hashrate is the speed at which a mining device performs the SHA-256 hashing algorithm, measured in hashes per second. It is the single most important performance metric for any Bitcoin mining machine, because every hash is one attempt to find a block solution.
Also known as: hash power, computing power, hashing speed.
Common units climb in factors of 1,000: H/s, KH/s (kilohash), MH/s (megahash), GH/s (gigahash), TH/s (terahash), PH/s (petahash), and EH/s (exahash). A modern home ASIC like the Bitaxe Hex Supra produces a few TH/s, while a full data-floor machine pushes past 100 TH/s. The combined output of every miner on the network is the network hashrate, today measured in hundreds of EH/s.
Where hashrate actually comes from
On an ASIC, hashrate is the product of three physical things: the number of hash cores on each chip, the number of chips wired into series on a hashboard, and the clock frequency those chips run at. More cores or a higher frequency means more nonce attempts per second. For example, Bitmain's BM1366 chip (used in the S19 XP, S19k Pro, and the Bitaxe Hex Supra) carries 894 small SHA cores, while the newer BM1370 in the S21 Pro carries 1,280 cores. Because the BM1370 packs more cores per chip, it can hit the same hashrate at a slightly lower clock frequency, which is exactly how newer silicon improves efficiency (J/TH).
This is also why hashrate and power draw are inseparable. Pushing a chip to a higher frequency raises its hashrate but increases voltage and heat super-linearly. Firmware autotuners exist precisely to find the best hashrate a given chip can sustain inside a fixed power budget without producing hardware errors. That tuning is calculated at runtime against each chip's actual silicon quality, not read from a fixed table.
Nominal vs. effective hashrate
The number printed on a miner's spec sheet is its nominal (nameplate) hashrate. The number that matters for earnings is the effective hashrate the machine actually delivers over time. The two can diverge for real, physical reasons:
- Hardware errors: Hashes that fail the chip's internal checks are wasted work and never count toward valid output.
- Thermal throttling: A hot hashboard or a failing fan forces firmware to cut frequency to stay inside safe limits, dropping hashrate.
- Dead or degraded chips: A single bad chip in a series chain can knock an entire board offline, and a partially failing board reports lower-than-nameplate output.
- Statistical variance: Hashing is a random search, so short-term hashrate readings bounce around the long-term average even on a perfectly healthy machine.
This is why pools report a separate "pool-side" or accepted hashrate, estimated from the difficulty of the shares your miner submits over time rather than from what the machine claims locally. If your dashboard shows full nominal hashrate but your pool-side number is consistently lower, that gap is a diagnostic signal: it usually points to rejected shares, a flaky connection, or a board that is hashing but not landing valid work.
Why hashrate drives mining economics
A higher hashrate increases the probability of finding a valid block hash below the network target and earning the proof-of-work reward. But raw hashrate alone tells you nothing about whether a machine is worth running. Two miners can produce identical hashrate while one consumes far more electricity, so operators always pair hashrate with efficiency to judge real-world profitability. When you evaluate hardware, think of hashrate as the revenue side of the equation and J/TH as the cost side.
A practical consequence for anyone diagnosing a sick machine: a sudden, unexplained drop in hashrate is one of the most common early symptoms of a hashboard, PSU, or cooling fault. Reading the per-board and per-chip hashrate breakdown in your firmware is often the fastest route to isolating which board has failed.
If you are sizing a new deployment or trying to recover lost hashrate from an underperforming rig, D-Central's ASIC troubleshooting resources walk through the hardware-level causes, and the miners catalog lists per-model nominal hashrate and efficiency so you can compare machines on the two numbers that actually decide whether a miner earns.