Every degree matters. CPUs throttling at 95°C lose 15–25% performance. GPUs running hot crash mid-game. This audit walks you through every component of your thermal system — from case airflow to paste application — so nothing is left to guesswork.
Who this is for: Anyone who's built, upgraded, or inherited a desktop PC. Takes 15–25 minutes. Work through each group, check items as you verify them, and watch the progress bar fill. When you hit 100%, your cooling is dialed in.
Airflow & Case Setup 7 items
Verify case has unobstructed front intake
Remove the front panel and check for dust filters, mesh, or solid barriers. A solid front panel with only side vents starves intake fans. If yours is solid, consider removing it or switching to a mesh-front case. Cases like the Fractal Pop Air and Corsair 4000D Airflow move 30–40% more air than their solid-front counterparts.
Confirm positive or neutral pressure configuration
Count your intake fans vs exhaust fans. Positive pressure (more intake than exhaust) pushes air out of unfiltered gaps, reducing dust. Negative pressure pulls dust in through every crack. Aim for equal CFM in and out, or slightly more intake. A single 120mm fan difference is enough to tip the balance.
Check all dust filters are clean and seated
Clogged filters reduce airflow by 40–60%. Remove every filter — front, bottom, top — and clean with compressed air or warm water. Let them dry completely before reinstalling. A clogged PSU dust filter alone can raise internal temps by 5–8°C. Set a quarterly reminder to clean them.
Verify front-to-back airflow path is clear of cable obstructions
Open the side panel and trace the airflow path from front intake to rear/top exhaust. Loose cables, drive cages, and GPU sag can create dead zones. Route cables behind the motherboard tray. Remove unused drive cages. A clean airflow path can drop CPU temps by 3–7°C with zero cost.
Confirm top exhaust fans are installed (heat rises)
Hot air naturally rises. Top-mounted exhaust fans leverage convection, pulling heat up and out. If your case has top fan mounts but no fans installed, add at least one 120mm or 140mm exhaust. Even a single top fan running at low RPM drops internal ambient temps by 3–5°C under load.
Check case is elevated and not sitting on carpet
Bottom-mounted PSU intakes pull air from underneath. Carpet blocks this intake and recirculates hot air. Place the case on a hard surface, a board, or a PC stand with at least 2 inches of clearance. Carpet fibers also get sucked into filters faster, accelerating dust buildup by 2–3x.
Verify ambient room temperature is below 80°F (27°C)
Your PC can only cool to ambient + the cooler's thermal delta. A room at 30°C means your CPU idles at 45°C minimum, even with a great cooler. Every 1°C increase in room temp raises component temps by roughly the same amount. Keep the room at 22–25°C for optimal cooling performance.
Fans & Heatsinks 7 items
Verify CPU cooler is properly mounted with even pressure
Uneven mounting pressure is the #1 cause of unexpectedly high CPU temps. For tower coolers, tighten screws in an X pattern — alternating diagonals. For AMD stock coolers, ensure both clips are fully engaged. AIO coolers need the bracket screws tightened evenly. A loose corner can create a 10–15°C hotspot.
Confirm CPU fan is oriented to push air toward rear exhaust
Tower cooler fans should push air toward the rear exhaust fan, not fight against it. Check the airflow arrow on the fan frame. If the fan pushes air toward the front of the case, flip it. This common mistake raises CPU temps by 5–10°C because hot air recirculates instead of exiting.
Check all case fans are spinning and not obstructed
With the system on, visually confirm every fan is spinning. A dead or disconnected fan creates a dead zone where heat pools. Check that cables aren't touching fan blades. If a fan is making grinding or clicking noises, the bearing is failing — replace it. A 120mm case fan costs $10–15 and takes 5 minutes to swap.
Verify GPU fans spin under load and aren't sagging
Most GPUs have a zero-RPM mode at idle. Run a game or stress test and confirm the fans activate. GPU sag can misalign the heatsink from the die — use a $5 GPU support bracket if you see sag. Also check that the GPU isn't suffocating against a PSU shroud or bottom case panel. 2+ inches of clearance underneath is ideal.
Check heatsink fins are free of dust buildup
Dust-clogged heatsink fins trap heat instead of dissipating it. Remove the CPU cooler (if possible) and use compressed air to blow out the fins. Hold the fan blades still while blowing air to prevent over-spinning. Dust buildup between fins acts as insulation — a heavily clogged cooler can lose 30–40% of its thermal capacity.
Confirm VRM and chipset areas have airflow coverage
VRMs (voltage regulator modules) throttle CPU performance when they overheat. Ensure your CPU cooler or case fans move air across the VRM heatsinks near the CPU socket. If you have a top-down cooler (like AMD Wraith), it naturally cools VRMs. Tower coolers need case airflow to cover this area.
Verify fan curves are set in BIOS — not running 100% constantly
Default BIOS fan curves are often too aggressive. Set a custom curve: fans at 30% below 50°C, ramping to 60% at 70°C, and 100% above 85°C. This keeps the system quiet at idle while maintaining full cooling under load. Most modern BIOS (ASUS, MSI, Gigabyte) have graphical fan curve editors.
Thermal Paste & Temperatures 6 items
Confirm thermal paste was applied within the last 2 years
Thermal paste dries out and pumps out over time, losing 5–15°C of effectiveness. If your paste is 2+ years old, clean and reapply. Use 99% isopropyl alcohol to remove old paste. High-quality pastes like Thermal Grizzly Kryonaut or Noctua NT-H1 maintain performance longer than generic compounds. Budget: $8–12 for a tube that lasts 10+ applications.
Verify paste coverage — pea-sized dot or full-spread thin layer
The pea-sized dot method works for most CPUs — the cooler's pressure spreads it evenly. For large CPUs (Threadripper, LGA 2066), use the X-pattern or spread method. Too much paste acts as insulation. Too little leaves dry spots. The goal is complete, thin coverage of the IHS. When you remove the cooler, you should see paste covering 90%+ of the surface.
Remove plastic protective film from cooler base (if new install)
This is the most common build mistake. New coolers ship with a clear plastic film over the base to protect the contact surface. Forgetting to peel it adds a thermal barrier that causes instant 20–30°C temperature spikes. Check both the CPU cooler base and any pre-applied paste pads. If you see a shiny plastic layer, peel it off before mounting.
Run stress test and verify CPU stays below 85°C under load
Use Cinebench R23 (free) or Prime95 for 10 minutes while monitoring temps with HWMonitor or Core Temp. Intel CPUs should stay below 85°C, AMD Ryzen below 90°C (Ryzen runs hotter by design). If temps exceed these thresholds, recheck paste application, cooler mounting, and airflow. Throttling begins at 100°C for Intel, 95°C for AMD.
Confirm GPU stays below 83°C under sustained gaming load
Run a GPU stress test (FurMark) or play a demanding game for 15+ minutes. Check temps with HWInfo or MSI Afterburner. NVIDIA GPUs throttle at 83°C by default. AMD cards throttle at 95°C (junction temp). If your GPU exceeds these, improve case airflow, repaste the GPU (advanced), or adjust the fan curve. A 5°C reduction can prevent throttling on NVIDIA cards.
Record baseline temps for future comparison
Document your idle and load temps for CPU and GPU. Screenshot HWInfo or write them down. These baselines let you detect degradation over time — rising temps at the same workload indicate dust buildup, paste drying, or fan failure. Check quarterly. If load temps rise 5+°C above baseline, it's time for maintenance.
Audit Complete — Your Cooling Is Dialed In
You've verified all 20 thermal checkpoints. Your system is running at optimal efficiency. Bookmark this page and re-run the audit quarterly to catch degradation early.
Get the Printable PDF VersionGo Deeper
Fan Testing Methodology
How we test CFM, static pressure, and noise levels. The data behind our recommendations.
Thermal Paste Comparison
12 compounds tested at 50W, 100W, and 200W TDP. Real deltas, not marketing claims.
Fan Curve Optimization Guide
Silent at idle, cool under load. Step-by-step BIOS fan curve setup for every motherboard brand.
Undervolting for Lower Temps
Drop CPU and GPU temps 10–15°C with zero performance loss. AMD and Intel walkthroughs included.
Best Airflow Cases Under $100
8 mesh-front cases tested with thermal imaging. Real airflow data, not marketing specs.
FAQ
Every 2–3 years for most pastes. High-end compounds like Thermal Grizzly Kryonaut may need reapplication every 1–2 years because they're designed for peak performance, not longevity. Budget pastes like Arctic MX-4 last 3–5 years. The real indicator is rising temps — if your CPU runs 5°C hotter than your baseline at the same workload, repaste.
For most CPUs (under 150W TDP), a $40–60 tower cooler like the Thermalright Peerless Assassin or Noctua NH-U12S matches or beats a 240mm AIO. AIOs win at 200W+ TDP (i9, Ryzen 9) where the larger radiator surface area matters. AIOs also have pump failure risk — no moving parts in a tower cooler. Budget pick: tower. High-end build: 280mm or 360mm AIO.
CFM (cubic feet per minute) measures airflow volume — best for unobstructed case intake/exhaust. Static pressure measures force through resistance — best for heatsinks and radiators. A high-CFM fan on a radiator underperforms a high-static-pressure fan by 3–8°C. Use airflow fans for case mounting, static pressure fans on coolers and radiators. Noctua NF-A12x25 excels at both.
It depends on the CPU. Intel 13th/14th gen i7/i9 chips are designed to run at 90–100°C under load — that's normal. AMD Ryzen 7000 series targets 95°C by design. However, if a mid-range chip (i5, Ryzen 5) hits 90°C, something is wrong — check paste, mounting, and airflow. Sustained temps above 95°C cause throttling and reduce lifespan.
Spinning doesn't mean performing. Sleeve-bearing fans (common in budget cases) lose efficiency after 2–3 years — they still spin but move less air. If your fans are 4+ years old or making noise, replace them. A $15 Arctic P12 PWM fan moves significantly more air than a worn stock fan. Also check RPM in BIOS — if a fan rated for 1500 RPM maxes at 800, the bearing is failing.