A sudden drop in hashrate or a rig that refuses to boot consistently points to a finite set of common hardware and software problems. My first diagnostic step is always to verify the voltage at the wall and the PSU rails; a fluctuation of just 5% can destabilise an entire rig. This isn’t abstract theory–it’s the primary cause of the random crashes many miners misdiagnose as a software configuration error. The core of effective troubleshooting is isolating the variable: does the mining software crash with one specific GPU, or do all GPUs fail simultaneously, indicating a board or power supply fault?
Beyond power, physical connections are a primary source of issues. A loosely seated PCIe riser or a slightly oxidized USB connector will cause intermittent device dropouts that cripple performance. I log temperatures rigorously; overheating VRAM on modern GPUs doesn’t always correlate with a high core temperature, leading to thermal throttling that silently reduces your hashrate by 20% without an obvious crash. This requires a methodical diagnostics process, checking each card’s memory junction temperature under load, a typical point of failure most basic monitoring tools miss.
Software stability is inextricably linked to hardware integrity. An unstable configuration in your mining OS or BIOS settings–like an overly aggressive memory clock or an insufficient page file–will manifest as rejected shares or a complete system hang. The goal isn’t just resolving the immediate crash, but engineering a system that runs for weeks. This means accepting a marginally lower hashrate for absolute stability, a trade-off that separates profitable crypto operations from a cycle of constant troubleshooting.
No Display on Monitor
Check the physical video output first. A frequent culprit is the GPU used for the display; bypass the riser and plug the monitor directly into a card seated in a primary x16 slot. Inspect all power connections to that specific GPU, ensuring both the 8-pin PCIe and the riser are secure. This basic hardware check resolves many display faults before deeper diagnostics are needed.
If connections are solid, the issue often lies with the motherboard’s BIOS configuration. Navigate to the ‘Init Display First’ setting and force it to ‘PCIe’ instead of ‘Auto’. Disable ‘CSM’ (Compatibility Support Module) and enable ‘4G Decoding’, a critical setting for multi-GPU mining rigs that directly impacts stability and the system’s ability to boot with all cards recognised.
Advanced Diagnostics for a Black Screen
Employ a systematic approach to isolate the faulty component. Test your monitor and cable with a known-working system. If the rig powers on but shows no display, listen for POST code beeps or observe diagnostic LEDs on the motherboard. Strip the rig down to a minimal configuration: one stick of RAM, a CPU, and a single GPU. If it boots, add components back one by one to identify the hardware causing the conflict.
Overheating can cause a failure to POST. Check that the CPU cooler is mounted correctly and that its fan is operational. High motherboard chipset temperatures, often overlooked, can also prevent a successful boot. Voltage irregularities from a failing or under-spec PSU are another common cause; use a PSU tester or try a different, adequately powered unit to rule this out.
Resolving a ‘no display’ problem is fundamental to restoring your mining operation. This troubleshooting process, while tedious, directly safeguards your hardware investment and is the first step to regaining optimal hashrate and long-term rig performance.
Rig Freezes or Crashes
Immediately test your power supply unit’s 12V rail under full load; a voltage drop below 11.4V is a primary cause of system instability. Use a multimeter to verify the voltage at the PCI-e connectors. A common hardware fault is an overstressed or failing PSU that cannot sustain the consistent high demand of mining, leading to abrupt crashes, especially as the rig approaches its target hashrate.
Systematic diagnostics should focus on heat management. Check each GPU’s core and memory junction temperatures; sustained operation above 80°C for core or 100°C for VRAM induces thermal throttling and eventual freezing. This overheating is not always obvious from fan speed alone. Resolving this often requires re-padding VRAM, adjusting fan curves, or improving ambient airflow directly across the rig’s frame.
For frequent software-related crashes, boot into a clean mining OS and test with a single GPU. A stable rig in this minimal configuration points to a software or driver conflict in your primary setup. Reinstall the mining client with administrator privileges and use a fixed virtual memory pagefile–set it to at least 1.5 times your total GPU memory. Many stability issues stem from an insufficient pagefile as the miner runs out of allocated memory.
Undervolting and overclocking configuration is a typical source of faults. An unstable memory clock might not cause immediate errors but can corrupt data over time, forcing a crash. Dial back all overclocks to stock settings to establish a baseline. Incrementally increase memory clocks in 25 MHz steps, monitoring for a dip in performance or rejected shares, which indicates the limit for your hardware. True stability often means sacrificing a few MH/s for a rig that runs continuously without intervention.
GPUs Not Detected
Check all physical connections before proceeding with software diagnostics. A loose PCIe riser cable or insufficient power to the GPU are typical hardware faults. Ensure the 6-pin or 8-pin PCIe power connectors are fully seated and that the riser is receiving power from both the motherboard and the PSU. A weak power supply unit that cannot deliver stable voltage under load is a common culprit for missing cards, especially during boot.
BIOS and OS-Level Configuration
Access your motherboard BIOS and set the PCIe speed to Gen2, not Auto or Gen3. Disable any ‘4G Decoding’ or ‘Above 4G Decoding’ option. For systems with multiple identical GPUs, Windows may install a generic driver that cripples performance; use Display Driver Uninstaller (DDU) in Safe Mode to wipe all GPU drivers, then reinstall the latest stable version from the manufacturer’s site. This process directly addresses driver conflicts that prevent GPU detection.
Advanced Troubleshooting for Persistent Issues
If specific GPU slots remain undetected, test each riser and GPU individually in a known-working PCIe slot. A frequent issue is a faulty riser, which can cause instability and detection problems. Monitor your power supply’s 12V rail voltage; a drop below 11.8V can cause the system to ignore GPUs to protect itself. For mining rig stability, a consistent power delivery is more critical than chasing minor hashrate gains that push your hardware to its voltage limits.
