Enterprise mining - nicehash/NiceHashQuickMiner GitHub Wiki

NiceHash QuickMiner with support for NVIDIA Quadro series and Tesla series

Latest version of NiceHash QuickMiner is bringing support for Enterprise class of NVIDIA video cards. For the first time in whole cryptocurrency mining history! If you have a powerful workstation or server full of Tesla cards. Now, when your power house is idle, you can monetize idle time with just few clicks, free of worries regarding security and safeness of shady miners. Read in general why you should use QuickMiner and nothing else and what kind of security mechanisms it uses to keep you and your data safe.

We have tested and created special one click optimization profile called Enterprise for the following cards (more are coming!):

Device PCI ID Device Name Tested hashrate Tested power consumption Efficiency Remarks
20F1 Tesla A100 ~170 MH/s ~200 W ~850 kH/J Disable ECC for slightly better efficiency
2235 A40 ~83 MH/s ~190 W ~430 kH/J Disable ECC, enable TCC
1EB8 Tesla T4 ~30 MH/s ~65 W ~460 kH/J Disable ECC for extra ~5 MH/s
1B38 Tesla P40 ~37 MH/s ~160 W ~230 kH/J Disable ECC for significant boost
1E30 RTX A5000 ~91 MH/s ~200 W ~460 kH/J Disable ECC, P0 state
1E30 Quadro RTX 8000 ~57 MH/s ~160 W ~350 kH/J Need TCC and P0 state
1EB1 Quadro RTX 4000 ~37 MH/s ~83 W ~440 kH/J
1C31 Quadro P2200 ~20 MH/s ~65 W ~300 kH/J Enable TCC if getting memory issues
15F0 Quadro GP100 ~40 MH/s ~150 W ~260 kH/J Enable TCC and P0 state

What to do regarding NVIDIA Tesla?

  • Disable ECC to improve efficiency and/or hashrate,
  • can only change core clock limit - Enterprise optimization selects best settings to maximize efficiency.

What to do regarding NVIDIA Quadro?

  • If having low amount of VRAM (5GB), disable WDDM and enable TCC using nvidia-smi (without monitor plugged in),
  • update to Excavator Build 890 to fix 5GB memory limit bug,
  • can manage fan and change core clock limit - Enterprise optimization selects best settings to maximize efficiency and optimize fan profile,
  • try TCC with P0 state for CUDA workloads if possible (for now, set P0 for CUDA using this useful tool).

Do you use Linux?

No problem! Just virtualize QuickMiner. One click solutions for virtual mining are coming soon!