Linux CPU frequency scaling is a collaboration between the scheduler, the CPUFreq governor, and a hardware-specific driver—not a direct promise that a core will run at one exact clock speed. The schedutil governor uses scheduler utilization to estimate a suitable performance request, while policy limits, driver behavior, and platform power or thermal controls influence what the processor actually delivers.
How CPUFreq and the scheduler fit together
CPUFreq is the Linux kernel framework for managing processor performance states. Its main parts have distinct roles:
- CPUFreq core: supplies common infrastructure and interfaces for policies and scaling.
- Governor: chooses a performance request based on a policy, such as scheduler utilization.
- Scaling driver: translates that request into hardware-specific control. A driver may also use platform-specific mechanisms that do not behave like a simple frequency command.
The kernel describes schedutil this way: “This governor uses CPU utilization data available from the CPU scheduler.” That connection lets the governor react to scheduler updates rather than relying only on a separate periodic load estimate. See the Linux kernel’s CPU Performance Scaling documentation.
What schedutil reads from the scheduler
CFS tasks and PELT utilization
For scheduler callbacks associated with the fair scheduling class (CFS), schedutil uses PELT utilization from the root control group. PELT—Per-Entity Load Tracking—tracks scheduler activity over time. Its utilization signal is not simply a count of runnable tasks: it reflects how much CPU time work is receiving and demanding.
#1 Best Overall
- [Brand Overview] Thermalright is a Taiwan brand with more than 20 years of development. It has a certain popularity in the domestic and foreign markets and has a pivotal influence in the player market. We have been focusing on the research and development of computer accessories. R & D product lines include: CPU air-cooled radiator, case fan, thermal silicone pad, thermal silicone grease, CPU fan controller, anti falling off mounting bracket, support mounting bracket and other commodities
- [Product specification] Thermalright PA120 SE; CPU Cooler dimensions: 125(L)x135(W)x155(H)mm (4.92x5.31x6.1 inch); heat sink material: aluminum, CPU cooler is equipped with metal fasteners of Intel & AMD platform to achieve better installation, double tower cooling is stronger（(Note:Please check your case and motherboard for compatibility with this size cooler.）
- 【2 PWM Fans】TL-C12C; Standard size PWM fan:120x120x25mm (4.72x4.72x0.98 inches); fan speed (RPM):1550rpm±10%; power port: 4pin; Voltage:12V; Air flow:66.17CFM(MAX); Noise Level≤25.6dB(A), leave room for memory-chip(RAM), so that installation of ice cooler cpu is unrestricted
- 【AGHP technique】6×6mm heat pipes apply AGHP technique, Solve the Inverse gravity effect caused by vertical / horizontal orientation, 6 pure copper sintered heat pipes & PWM fan & Pure copper base&Full electroplating reflow welding process, When CPU cooler works, match with pwm fans, aim to extreme CPU cooling performance
- 【Compatibility】The CPU cooler Socket supports: Intel:115X/1200/1700/17XX AMD:AM4;AM5; For different CPU socket platforms, corresponding mounting plate or fastener parts are provided（Note: Toinstall the AMD platform, you need to use the original motherboard's built-in backplanefor installation, which is not included with this product)
Running and runnable time are distinct. When tasks contend for a CPU, a task may receive less running time even as runnable demand grows. The scheduler’s tracking accounts for both, helping distinguish a lightly loaded CPU from one where work is waiting for service.
RT and deadline callbacks
When a real-time (RT) or deadline scheduling callback triggers an update, schedutil raises the request to the maximum allowed by the CPUFreq policy. This is a scheduling-class-specific rule; it is not the same calculation used for a CFS utilization update.
IO-wait boost
An IO-wait boost can temporarily raise the requested performance to the policy maximum. The boost then backs down toward the value calculated from utilization. It is therefore a temporary response, not evidence that the workload’s steady-state utilization requires maximum performance.
Rank #2
- Cool for R7 | i7: Four heat pipes and a copper base ensure optimal cooling performance for AMD R7 and Intel i7.
- Quiet Cooling Fan: SickleFlow 120 Edge with Dynamic PWM control (690–2,500 RPM), designed for low noise and peak cooling performance.
- Simplify Brackets: Redesigned brackets simplify installation on AM5 and LGA 1851|1700 platforms.
- Versatile Compatibility: 152mm tall design offers performance with wide chassis compatibility.
- Easy Installation: Easy to install with included thermal paste for hassle-free setup and optimal cooling performance.
How utilization becomes a frequency request
For a frequency-invariant utilization signal, the documented schedutil relationship is:
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
f = 1.25 × f₀ × util / max
utilis the PELT utilization value.maxis the theoretical maximum for that utilization scale.f₀is the CPUFreq policy’s maximum frequency when utilization is frequency-invariant. If the signal is not frequency-invariant,f₀is the current frequency.
The formula describes the governor’s calculation, not a guarantee of sustained clock speed. The resulting request remains subject to the policy’s minimum and maximum, the frequencies or performance levels supported by the driver, and platform constraints.
Why frequency invariance matters
The same raw utilization value does not necessarily mean the same computing capacity on CPUs running at different frequencies. It also may not mean the same capacity on different CPU types. Frequency invariance adjusts utilization to account for changing work rate as frequency changes; microarchitecture invariance accounts for capacity differences across CPU types. These adjustments help the scheduler and governor interpret utilization in terms of available capacity rather than treating every CPU and clock state as equivalent. The kernel’s scheduler capacity documentation describes these invariance concepts.
Rank #3
- [Brand Overview] Thermalright is a Taiwan brand with more than 20 years of development. It has a certain popularity in the domestic and foreign markets and has a pivotal influence in the player market. We have been focusing on the research and development of computer accessories. R & D product lines include: CPU air-cooled radiator, case fan, thermal silicone pad, thermal silicone grease, CPU fan controller, anti falling off mounting bracket, support mounting bracket and other commodities
- [Product specification]AX120R SE; CPU Cooler dimensions: 125(L)x71(W)x148(H)mm (4.92x2.8x 5.83 inch); Product weight:0.645kg(1.42lb); heat sink material: aluminum, CPU cooler is equipped with metal fasteners of Intel & AMD platform to achieve better installation
- 【PWM Fans】TL-C12C; Standard size PWM fan:120x120x25mm (4.72x4.72x0.98 inches); fan speed (RPM):1550rpm±10%; power port: 4pin; Voltage:12V; Air flow:66.17CFM(MAX); Noise Level≤25.6dB(A), the fan pairs efficient cool with low-noise-level, providing you an environment with both efficient cool and true quietness
- 【AGHP technique】4×6mm heat pipes apply AGHP technique, Solve the Inverse gravity effect caused by vertical / horizontal orientation. Up to 20000 hours of industrial service life, S-FDB bearings ensure long service life of air-cooler radiators. UL class a safety insulation low-grade, industrial strength PBT + PC material to create high-quality products for you. The height is 148mm, Suitable for medium-sized computer case
- 【Compatibility】The CPU cooler Socket supports: Intel:1150/1151/1155/1156/1200/1700/17XX/1851,AMD:AM4 /AM5; For different CPU socket platforms, corresponding mounting plate or fastener parts are provided
Rate limiting updates
The rate_limit_us tunable sets the minimum interval between governor computations. The CPUFreq documentation gives a default of 1.5 times the driver transition latency, or 1 ms if no transition latency is provided. This is an update-rate setting, not a promise that the hardware will complete a transition in that time.
Other factors that change the request or the result
Utilization clamps
Utilization clamping changes the utilization signal that schedutil sees and can therefore affect frequency selection. A configured minimum clamp can keep the signal from falling below a floor; a maximum clamp can cap it. Whether that improves responsiveness, performance, or energy use depends on the clamp configuration, workload, scheduler behavior, driver, and hardware. There is no universal performance gain implied by enabling clamps. See the kernel’s schedutil and utilization-clamping documentation.
Policy bounds and supported states
A CPUFreq policy has minimum and maximum bounds. Requests are constrained by those limits, and a driver may support only particular frequencies or performance levels. Thus, a mathematical calculation can yield a target that is subsequently constrained or mapped to something the driver can implement.
Rank #4
- Support Intel LGA 1200/1156/1155/1150/1151
- Low Profile Design. Air flow - 31.343 CFM. Noise level - 21.3 decibels
- Optimized for low power CPU's
- 7-Bladed Low Noise Fan
- Quick and Easy Installation
Driver and platform behavior
CPUFreq behavior depends on the active driver. For example, the kernel’s AMD P-State documentation lists schedutil and ondemand as supported dynamic governors and describes an adjust_perf callback for performance updates similar to CPPC. Other drivers can implement performance algorithms using hardware information or bypass the generic governor layer. The governor name alone therefore does not fully describe how a platform will control performance; consult documentation for the active driver, such as AMD P-State.
Why actual frequency can differ from the request
A governor request is an input to the system, not a guarantee of the clock frequency a processor will sustain. The kernel documentation cautions: “Though the intention may be to set an exact frequency for the policy, the actual frequency may vary depending on hardware coordination, thermal and power limits, and other factors.” Hardware may coordinate performance across cores, and thermal or power controls can constrain operation even when the requested value is higher.
This distinction matters when interpreting monitoring data. A mismatch between a requested operating point and an observed frequency does not by itself prove that schedutil miscalculated. The driver’s control model and platform limits may explain the difference.
Best Value
- Simple, High-Performance All-in-One CPU Cooling: Renowned CORSAIR engineering delivers strong, low-noise cooling that helps your CPU reach its full potential
- Efficient, Low-Noise Pump: Keeps your coolant circulating at a high flow rate while generating a whisper-quiet 20 dBA
- Convex Cold Plate with Pre-Applied Thermal Paste: The slightly convex shape ensures maximum contact with your CPU’s integrated heat spreader, with thermal paste applied in an optimised pattern to speed up installation
- RS120 ARGB Fans: RS ARGB fans create strong airflow and high static pressure, with easy ARGB control via a compatible motherboard. CORSAIR AirGuide technology and Magnetic Dome bearings ensure great cooling performance and low noise
- Easy Daisy-Chained Connections: Reduce the wiring in your system by daisy-chaining your RS ARGB fans and connecting them to just one 4-pin PWM fan header and one +5V ARGB header
How to diagnose CPUFreq behavior on a real Linux system
Before comparing two machines or drawing conclusions about governor behavior, establish the configuration and operating conditions. The exact controls available depend on kernel configuration and platform support.
- Identify the kernel and platform. Record the kernel version, processor type, and distribution or kernel build. Documentation describes general kernel behavior; it does not establish how every vendor build is configured.
- Find the active CPUFreq driver and mode. Use the CPUFreq interfaces and driver documentation for the system to determine which driver controls performance and whether it uses a conventional frequency or performance-request model.
- Check the available governor and active policy. Confirm that
schedutilis available and selected, then note the policy’s minimum and maximum bounds. Do not assume another system exposes the same governors or controls. - Account for the scheduler signal. Determine whether the workload uses CFS, RT, or deadline scheduling, and whether utilization clamps or IO-wait boosts affect the request. These cases can invoke different behavior.
- Consider invariance and CPU capacity. Check whether frequency and microarchitecture invariance are relevant to the system. Raw utilization values are not directly comparable across unlike frequencies or CPU types without that context.
- Interpret observed frequency alongside platform limits. Look for hardware coordination, thermal constraints, and power limits before treating a gap between request and actual frequency as a governor fault.
For comparisons, record the active driver and mode, governor and policy limits, utilization treatment, workload and scheduling class, and thermal or power conditions. Without those details, a governor label is not enough to make a meaningful system-to-system comparison.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
