PC Troubleshooting Guide: CPU & GPU Issues

PC issues can look dramatic—crashes, no display, random restarts, or a system that suddenly refuses to boot—but most problems come down to a handful of simple causes. Loose cables, outdated BIOS versions, driver conflicts, and RAM instability account for the majority of faults people experience after upgrading or rebuilding a system.

This guide covers the most common causes of boot problems, crashes, no‑display issues, and general instability. Work through the steps in order — most problems are solved long before you reach the end.


Before You Start: The 2‑Minute Basics

These are the checks that fix a surprising number of "dead" PCs before any parts get touched:

  • Check the power chain: confirm the PSU rocker switch is on, the cable is firmly seated at both ends, and the wall socket or extension lead is working.
  • Check the display input: make sure your monitor is on the correct input and that the cable is plugged into the graphics card, not the motherboard video output (unless you have no GPU installed).
  • Unplug everything non‑essential: disconnect USB devices, extra drives and additional monitors. Boot with just a keyboard, one display and the bare minimum.
  • Note exactly what happens: do the fans spin? Do any lights come on? Is there a beep? The answer narrows down the cause quickly (see the debug LED section below).

CPU Troubleshooting

1. Update the BIOS

If the system won’t boot, restarts repeatedly, or shows no display, an outdated BIOS is one of the most common causes. Check your motherboard manufacturer’s website for the latest version.

This matters most when fitting a newer CPU to an older board. Many boards need a BIOS update before they will recognise newer processors, so check the manufacturer’s CPU support list before you buy.

2. Clear the CMOS

Resetting the CMOS removes old settings that can conflict with a new processor. Use the motherboard jumper or remove the battery for a few minutes.

3. Reseat or test RAM

RAM problems often look like CPU problems. Try:

  • Testing one stick at a time
  • Swapping slots
  • Making sure each module is fully seated
  • Disabling XMP / EXPO in the BIOS and running at default speeds

4. Check CPU power cables

Make sure the 8‑pin (or 4+4‑pin) EPS connector is firmly connected. If this cable isn’t fully inserted, the system won’t start. Many boards also have a second EPS connector, so connect both if your PSU has the cables.

5. Reboot a few times

Some boards need several training cycles to recognise a new CPU, especially after a BIOS update or CMOS reset. On newer platforms such as AMD AM5, the first boot with new RAM can take several minutes with a black screen while the memory trains. Give it time before assuming it’s dead.

6. Inspect the socket and CPU for damage

With the cooler off and the CPU removed, check for bent pins in the socket (Intel LGA and AMD AM5 use pins in the socket) or on the CPU itself (older AMD AM4 chips). Look for debris and for any discolouration or marks around the contact area. Only ever lift or lower a CPU straight in and out, with no sliding.

For a more detailed guide on CPU testing, check out our article on How to Tell if your CPU is Failing? Signs, Symptoms, and How to Test Your Processor.

Need a CPU Upgrade?

Browse our tested AMD and Intel CPUs

View Tested CPUs →

Need a New Motherboard?

Replace a faulty or outdated board with a tested one

View Motherboards →

Decoding Debug LEDs and Beep Codes

Many modern motherboards have small diagnostic LEDs near the RAM slots or the edge of the board, and some have a two‑digit debug display. These tell you exactly which stage of startup the PC is stuck at:

  • CPU: the board can’t initialise the processor. Check the EPS power cable, BIOS version and CPU seating.
  • DRAM: a memory problem. Reseat the RAM, try one stick at a time and try different slots.
  • VGA: a graphics problem. Reseat the GPU, check the PCIe power cables and try another display cable or port.
  • BOOT: the hardware has passed and the board can’t find a bootable drive. Check the SSD connection and boot order.

If your board has no LEDs, it may use beep codes through a case speaker instead. The meaning of these differs between manufacturers, so always check your motherboard manual.


GPU Troubleshooting

1. Install fresh GPU drivers

Use the official NVIDIA or AMD installer and choose the clean‑install option to remove old driver conflicts. For stubborn problems, boot into Safe Mode and use a tool like Display Driver Uninstaller (DDU) before installing the latest driver.

2. Reseat the GPU

A slightly misaligned PCIe connection can cause:

  • No display
  • Random crashes
  • “GPU not detected” messages

Remove the card and reinstall it firmly until the latch clicks.

3. Check PCIe power cables

Make sure all connectors are fully inserted. For higher‑power cards, use separate PCIe cables, not daisy‑chained ones. On RTX 40 series and newer cards using the 16‑pin 12VHPWR connector, push the plug in fully until it is flush with no visible gap, and avoid sharp bends right next to the connector.

Need a GPU Power Cable?

NVIDIA RTX 40 series 12VHPWR PCIe 5.0 cable, 2x 8‑pin to 16‑pin

View Power Cable →

4. Update the BIOS and check PCIe mode

Ensure the motherboard is running the slot at x16 and is on the latest firmware.

5. Try different display cables or ports

Faulty HDMI or DisplayPort cables are extremely common and can mimic GPU failure. Try a different cable, a different port on the card, and if possible a different monitor.

6. Check airflow and GPU sag

Poor cooling or a sagging card can cause intermittent contact or thermal throttling. Heavy modern cards put real strain on the PCIe slot, so a support stand helps keep the card level.

GPU Sagging?

A GPU support stand keeps heavy cards level and stops slot strain

View GPU Support Stand →

7. Verify PSU wattage

Make sure your power supply meets the recommended wattage for your GPU and has the correct connectors. An ageing or underpowered PSU can cause black screens or restarts only when the GPU is under load.

8. Rule the card in or out with another video output

If your CPU has integrated graphics, remove the GPU and plug your monitor into the motherboard output. If the PC boots normally, the problem is likely the card, its power cables or the PCIe slot. If it behaves the same way, look at the rest of the system. If you have a spare working GPU, testing it in the same slot is the most conclusive check of all.

For a more detailed guide on GPU testing, check out our article on How to Tell If Your Graphics Card is Failing: GPU Artifacts & Stress Test Guide.

Need a Great Value GPU?

Browse our tested used graphics cards

View Tested GPUs →

Need a Reliable PSU?

Rule out power problems with a tested power supply

View PSUs →

Software Clues: Let Windows Tell You What’s Wrong

If the PC boots but crashes, Windows often records exactly what went wrong:

  • Note the BSOD code: consistent WHEA_UNCORRECTABLE_ERROR messages point towards the CPU or its settings, while a mix of codes like MEMORY_MANAGEMENT and PAGE_FAULT_IN_NONPAGED_AREA point towards RAM.
  • Check Event Viewer: open Event Viewer and look under Windows Logs > System for Critical and Error entries around the time of the crash, especially “Kernel‑Power” events, which often point to power or overheating issues.
  • Monitor temperatures: use a free tool like HWMonitor or HWiNFO to watch CPU and GPU temperatures under load. Sustained temperatures above roughly 90°C usually mean a cooling problem.
  • Test your memory: run Windows Memory Diagnostic, or for a more thorough result follow our guide on how to use MemTest86 to test your RAM.

Running Hot?

Quiet 120mm ARGB case fans to improve airflow

View Case Fan →

Quick Symptom Guide

Symptom Check First
No power, no fans, no lights Power supply, PSU switch, front panel power connector, EPS cable
Fans spin but no display RAM seating, GPU seating and power, monitor cable and input, CMOS reset
Boots, then restarts or shuts down under load Overheating, underpowered PSU, unstable RAM profile
Random blue screens with different codes RAM (test with MemTest86), XMP / EXPO settings, drivers
Artifacts, lines or flickering on screen GPU drivers, GPU temperatures, failing graphics card
Stuck on the BIOS screen or boot loop BIOS version, CMOS reset, boot drive connection

How to Tell if It’s a Hardware Fault

After troubleshooting, genuine hardware issues usually show:

  • Consistent crashes under load
  • Visual artifacts (lines, flickering, corrupted textures)
  • Failing stress tests or benchmarks
  • No POST even with known‑good components
  • Burning smell or visible damage

If you’ve ruled out BIOS, drivers, RAM, PSU, cables, and cooling, the component may need replacing.

If a part has failed, replacing only that component is almost always cheaper than replacing the whole system. It helps to test with known‑good parts where you can: swapping in a working RAM stick, PSU or GPU from another PC is the fastest way to prove which component is at fault. If you want to know how we check parts before selling them, see how we test our PC components.


Preventing Future Problems

A few simple habits help avoid repeat issues:

  • Keep BIOS and drivers updated
  • Use a reliable power supply
  • Avoid forcing connectors
  • Clean dust regularly
  • Maintain good airflow
  • Use surge protection
  • Keep your CPU and GPU temperatures in check with regular monitoring
  • Do not stack cables tightly against hot components or over fans

Upgrade Your Rig Now

High‑end specs, budget‑friendly prices. Every part we sell is tested, so you can replace the faulty component with confidence.

Build a Custom PC

Choose your own components and get an instant quote

Open the PC Builder Tool →

Need RAM or Storage?

Fully tested memory and SSDs at great prices

View Tested RAM →

0 comments

Leave a comment

Upgrade Your Rig Now. High-End Specs. Budget-Friendly Prices.

AMD CPUs
AMD CPUs

AMD CPUs

AMD GPUs
AMD GPUs

AMD GPUs

Nvidia GPUs
Nvidia GPUs

Nvidia GPUs

Intel CPUs
Intel CPUs

Intel CPUs