🔥 Thermal Stress Tests

The App Info & Tools panel (click the CoolerControl logo) includes a set of thermal stress tests. These generate sustained hardware load to help you validate that your fan curves and cooling profiles respond correctly before relying on them in real workloads.

TIP

This is a heat generator, not a benchmark. Results reflect thermal behavior, not performance scores.

How to Use

1. Click the CoolerControl logo in the top-left of the UI to open Info & Tools.
2. Scroll down to the Thermal Stress Tests section.
3. Set a duration (in seconds) for each test you want to run.
4. Click Start on the desired test type(s).
5. Monitor temperatures and fan behavior on your dashboards.
6. Tests stop automatically when the duration expires, or click Stop / Stop All at any time.

A status indicator on each test shows Active or Inactive, and updates every two seconds while any test is running.

Test Types

CPU Stress

Runs FMA and square-root compute loops across logical CPU cores. Each thread sweeps a 4 MiB buffer to stress both the L3 cache and DRAM. Uses AVX2+FMA SIMD when the CPU supports it, with a scalar fallback otherwise.

Options:

• Threads — number of threads to use (default: all logical cores)
• Duration — how long to run in seconds

GPU Stress

Runs memory-bandwidth compute shaders via wgpu (Vulkan or OpenGL ES) with interleaved FMA and sqrt operations to load both VRAM and shader cores. All detected hardware GPUs are stressed in parallel.

GPU driver requirement

GPU stress requires Vulkan or OpenGL ES drivers when using the built-in backend. If no compatible driver is found the test will fail with an error.

Options:

• Duration — how long to run in seconds

RAM Stress

Performs streaming read-modify-write passes across a large memory allocation (80% of available RAM by default). Uses AVX2 non-temporal stores to bypass the CPU cache and write directly to DRAM, stressing the DIMMs and memory controller.

Options:

• Duration — how long to run in seconds

Drive Stress

Issues random 4 KiB I/O reads against a selected block device using multiple parallel threads, to generate heat on drive controllers and NVMe/SATA controllers.

Options:

• Select drive — choose which block device to stress
• Threads — number of parallel I/O threads (default: 16)
• Duration — how long to run in seconds

stress-ng Integration

When stress-ng is installed, CoolerControl uses it instead of the built-in engine. stress-ng typically produces higher and more consistent thermal load. The active backend is shown next to each running test (built-in or stress-ng).

Install stress-ng from your distribution's package repositories:

# Fedora / RHEL
sudo dnf install stress-ng

# Debian / Ubuntu
sudo apt install stress-ng

# Arch
sudo pacman -S stress-ng

PSU Warning

High Power Warning

Running CPU and GPU stress simultaneously places significant load on your PSU. On overclocked systems or builds with an undersized PSU, this can cause system instability. CoolerControl will prompt for confirmation before starting a second test when both are running.

See Also

• Profiles — create and tune the fan curves you want to validate
• Functions — add hysteresis and response-time tuning
• Dashboards — monitor temperatures in real time while tests run