Electronics
Cooling
CPU's own gotchas mention a CPU reducing its own clock speed under sustained heavy load. This page is why: every watt of electrical power a component draws eventually becomes heat, and removing that heat fast enough is what lets a chip keep running at full speed instead of protecting itself by slowing down.
- What it is
- The systems (fans, heatsinks, liquid loops) that remove heat generated by electrical components fast enough to keep them within a safe operating temperature.
- Mental model
- A component under load is an engine; cooling is the radiator. If the radiator can't keep up, the engine has to slow itself down to avoid damage.
- Key numbers
- Measured in °C operating temperature and TDP (Thermal Design Power) in watts.
- Reach for it when
- Explaining why a component slows down or becomes unstable specifically under sustained heavy load.
Why heat matters
Semiconductor performance and long-term reliability both degrade as temperature rises, and every component has a maximum safe junction temperature it's rated never to exceed. Nearly all the electrical power a CPU or GPU draws under load is converted directly into heat — cooling isn't a bonus feature, it's the thing standing between rated performance and self-protective shutdown.
Air vs. liquid cooling
- Air cooling
- A metal heatsink (large surface area, usually finned) plus a fan pushing air across it. Simple, reliable, no leak risk, and sufficient for the large majority of components.
- Liquid cooling
- A pumped coolant loop carries heat away from the component to a separate radiator, more efficient at moving large amounts of heat than air alone — at real added cost, complexity, and a genuine (if now well-engineered-around) leak failure mode.
Thermal throttling
When a component approaches its maximum safe temperature, it deliberately reduces its own clock speed — directly lowering how much heat it generates per second — rather than risk damage. This is protective behavior working correctly, not a malfunction: a CPU that throttles under sustained load is doing exactly what it's designed to do when cooling can't keep pace with the heat being generated.
TDP
TDP (Thermal Design Power) is the amount of heat, in watts, a cooling solution needs to be able to continuously remove for a given component — the number cooling hardware is actually sized against, distinct from a component's peak electrical power draw in any single instant.
Real-world applications
- Every powered electronic system — a microcontroller's own datasheet specifies an operating temperature range — see the Embedded domain — the same underlying constraint at a much smaller scale.
- Data center cooling infrastructure — often one of a data center's largest ongoing operating costs, at a scale where cooling efficiency has real, direct financial impact.
- Laptop thermal design — a thin chassis is a genuine, hard cooling constraint, directly limiting how much sustained performance a laptop can offer compared to a desktop with the same chip.
Gotchas
- Thermal throttling is frequently misdiagnosed as a hardware fault when it's actually the system correctly protecting itself — checking temperatures under load is a standard first troubleshooting step before assuming anything is actually broken.
- Dust buildup inside a cooling system is one of the most common, most preventable causes of rising temperatures over a device's operating lifetime — simple, regular cleaning meaningfully extends both performance and component life.
- More or bigger fans isn't strictly better past the point of adequately removing the heat actually being generated — beyond that point, extra cooling capacity mostly buys quieter operation, not more performance.