Mining Hardware Testing Methodology
We test mining hardware the way you would if you had unlimited time and a decent lab setup. No shortcuts, no manufacturer cherry picks. Here’s exactly what we do and why we do it.
Power Draw Testing
We use a wall meter, not software readings. Driver reported power is often optimistic by 5-15%, especially on cards with modded BIOS or aftermarket coolers. For every unit, we log:
- Idle draw (system at desktop, mining software stopped)
- Mining load draw (steady state after 30 minutes)
- Peak draw during algorithm switch or startup spike
We test at stock settings first, then with common undervolts and power limits. That way you see what a card actually pulls in a real rig, not just out of the box. For ASICs, we also measure PDU level draw when possible. The difference between nameplate and real draw can be brutal.
Hashrate Benchmarks
We run each unit on multiple algorithms where applicable. For GPUs that means Ethash (still relevant for ETC and sidechains), KawPow, Autolykos, and sometimes RandomX for CPUs. For ASICs, we test on the algorithm the machine was built for and note performance on any alternate firmware available.
Each benchmark runs for a minimum of 24 hours. A 5 minute window tells you nothing. Hashrate variance, thermal throttling, and memory error