How to Undervolt a Laptop to Run Modded Games Cooler (2026)

Undervolting a laptop means lowering the voltage offset applied to its CPU and GPU so they run cooler and draw less power, usually without losing performance. On a machine running modded games, it is the cleanest way to stop thermal throttling from eating your frame rate. Check that your model supports it, start conservative, and change one setting at a time.

Mods push a laptop much harder than the base game does. An ENB preset, a physics overhaul, more NPCs, more draw calls and a raised VRAM budget all mean longer sustained load with no quiet moments for the fans to catch up. A thin chassis that copes with a 40-minute vanilla session will sit pinned at its temperature ceiling for a two-hour modded run.

Once the CPU or GPU hits that ceiling, the clock drops mid-frame. That is the stutter you notice, not a lower average frame rate. Undervolting lowers the heat the chip produces for the same clock speed, which buys back the sustained performance the throttling was taking away.

Here is the whole process in short form:

1. Identify the exact laptop model and CPU, then confirm voltage or power-limit controls are actually exposed.
2. Update the BIOS, plug in the original charger, and set Windows to the Best Performance power mode.
3. Open HWiNFO64 in sensors-only mode and record CPU temperature, clock speed, throttling flag and GPU temperature.
4. Run your most demanding modded game for 20 to 30 minutes and write down the baseline numbers.
5. Open the right tool: ThrottleStop or Intel XTU for Intel, Ryzen Master or Ryzen Controller for AMD.
6. Set the core voltage offset to -100 mV, then step down by 10 mV at a time.
7. Test stability in the actual modded game, not just a desktop benchmark.
8. Save the profile, set it to load at startup, and write down how to revert it.

What You Need

Before touching anything, gather four things. Your exact laptop model number and the CPU inside it, your BIOS version, the original charger, and a monitoring tool running alongside whatever tuning utility you use.

Find the CPU first. Press Windows key plus R, type msinfo32, and read the Processor line. On AMD machines the model number tells you whether you are on a generation that still allows tuning. On Intel, the generation matters just as much, as I will explain in the first step.

For monitoring, HWiNFO64 in sensors-only mode is the standard pick because it shows the thermal throttling flag directly rather than making you infer it. HWMonitor works as a simpler alternative, and the Windows Game Bar performance overlay will give you frame rate and frametimes if you would rather not install anything extra.

Keep the charger plugged in for the whole session. Battery power puts the processor into a different, far more restricted power envelope, so any offset you test there tells you nothing about behaviour on mains power.

One warning up front: many modern laptops lock voltage control in firmware. If the BIOS exposes only performance modes and fan curves, that is a manufacturer decision, and it is not something you can safely bypass. The alternative methods below cover that case properly.

Step-by-Step: How to Undervolt a Laptop to Run Modded Games Cooler

1. Confirm Whether Your Laptop Supports Undervolting

Check the manufacturer’s manual and your BIOS setup for voltage, power-limit or performance-control options before you download anything. Boot into the BIOS, look through the advanced or power menus, and note what is actually there. A CPU offset, a wattage slider or a performance profile all count. A bare temperature readout with nothing adjustable does not.

Intel 12th, 13th and 14th generation mobile chips are the big trap here. Microcode updates including 0x129 and 0x12B added undervolt protection, so on many of those laptops the voltage offset fields are greyed out or refuse values below zero. Any guide written before 2026 that walks you through FIVR without mentioning this will send you in circles. If your offset fields are locked, skip to step 3 and use power limits or underclocking instead.

Laptop BIOSes also tend to hide voltage control behind a performance mode. Choosing Balanced or Performance changes the power envelope rather than the voltage curve directly, and it is reversible from the same menu.

2. Record a Baseline Temperature and Frame Rate

You cannot tell whether a change helped without a number to compare against. In HWiNFO64, open the sensors window and watch CPU Package temperature, Core Effective Clock and the Throttling flag alongside GPU temperature.

Now run the game that hurts the most. Use the modded build you actually play, not a benchmark, because modded loads spike in ways vanilla ones do not. Let it run for 20 to 30 minutes so the chassis reaches steady state, then note the peak CPU temperature, whether the throttling flag ever turns active, the lowest frame rate you hit and how loud the fans got.

Write it down in a text file. That log becomes your recovery target if the change goes wrong.

3. Choose the Right Way to Undervolt a Laptop

Different tools control different parts of the machine, and only one of them is right for your hardware.

• ThrottleStop works on Intel laptops and exposes core, cache and iGPU voltage offsets through the FIVR panel. It is free and the settings are gone after a reboot unless you configure it to reapply them.
• Intel XTU does the same job for Intel chips with a cleaner interface and built-in stress testing. Settings disappear when you close the program unless you apply a profile.
• Ryzen Master is AMD’s own utility, and Ryzen Controller is the community alternative for laptops the official tool refuses to touch. Current Ryzen laptops often expose per-core Curve Optimizer values instead of a flat offset.
• MSI Afterburner handles the GPU. On a laptop it controls the discrete card, usually an NVIDIA GeForce, through its voltage and frequency curve editor.
• Manufacturer performance modes and Windows power modes are the fallback when voltage is locked. They cost you some peak boost but they are stable and fully supported.

Do not run two of these at once. ThrottleStop and XTU will fight each other, and afterburner overclocking plus a CPU undervolt makes instability very hard to trace.

4. Apply a Conservative Voltage or Power Limit

Open FIVR in ThrottleStop, or the Core Voltage Offset box in XTU, and enter -100 mV. That is the standard starting point on mobile silicon. Then step down in 10 mV increments while the game runs.

Set the cache offset to match the core offset unless testing shows otherwise. Many mobile chips are limited by the cache domain rather than the cores, and leaving cache at zero while the core goes deep is a common reason a stable-looking offset still crashes.

Watch HWiNFO64 the whole time. If the CPU temperature is falling as the offset drops, you are in the useful range. If it stops moving, you have hit the point where the chip is power or temperature limited anyway, and going deeper buys nothing but risk.

With the voltage route locked, set the long-term power limit instead. Dropping sustained wattage by 10 to 15 W usually trims several degrees and keeps boost clocks steady for longer, which is what a modded session actually needs.

Move the offset only in small steps and keep the charger connected. A jump from zero to -150 mV tells you almost nothing about where the edge is.

5. Test Modded Games for Crashes and Performance Problems

Desktop benchmarks miss the load patterns that break undervolts. Particle-heavy scenes, dense cities and save-game recompiles are the things worth testing, because they hold the CPU at high load far longer than any Cinebench run.

Watch for the warning signs. A crash to desktop, a freeze that needs a hard reboot, a boot loop where the machine restarts several times, corrupted save files, flickering geometry or texture pop-in, and stutter that is worse than before all point at an unstable offset. So does fan noise that suddenly gets loud and uneven.

Run for at least an hour. Instability under sustained boost is common and shows up late, which is why a 15-minute test convinces people they are safe when they are not.

Then compare against your baseline. Steady frame rates and a quieter chassis mean the change worked. If the average frame rate is identical and the throttling flag never activated during your baseline, undervolting was not your problem and the cooling hardware is.

6. Save a Stable Profile and Restore the Backup if Needed

Back the profile off by one 10 mV step from whatever crashed, then confirm that value again with a full session. Test overnight if you can, since some faults only surface once the machine is properly warm.

Save that value as a named profile in your tool. In ThrottleStop, use File and Save to store it, then set the program to start with Windows and apply the profile on launch. In XTU, save a profile in the software before closing it. Without this step your settings reset to zero every reboot, which is the single most common reason people think their undervolt stopped working.

Know the exit route before you need it. Setting every offset back to zero restores default behaviour immediately. If the machine will not boot, hold the power button for about ten seconds, then boot and reset the BIOS to defaults from the firmware setup screen. For a laptop that will not power on at all, holding the power button while you reconnect the charger resets the internal battery controller on most models.

For Ryzen laptops, a bad Curve Optimizer value behaves the same way. Zero out the affected core or set the whole system back to default in Ryzen Master.

Common Mistakes

The most common failure is chasing the lowest offset instead of the most stable one. Forum threads are full of people reporting blue screens at -150 mV who would have been perfectly happy at -80 mV with identical frame rates. Find the edge, then give it margin.

Locked voltage controls are the second. If your CPU is a recent Intel mobile chip and FIVR refuses negative values, nothing is broken. The microcode is doing what it was designed to do. Move on to power limits or underclocking.

Changing core, cache, iGPU and GPU all in one sitting makes any crash impossible to diagnose. One domain at a time, and confirm stability before touching the next.

Testing on the desktop proves nothing about your game. Wallpaper and a browser tab will sit at 40 C while the same chip hits 98 C in a crowded modded town.

Running the whole session on battery changes everything, and any result from a battery test is meaningless. Plug in the original charger, not a USB-C substitute that negotiates a lower wattage.

Misreading the numbers causes just as much confusion. A CPU sitting at 90 C under load is inside its designed limit, and a GPU at 80 C is completely normal. What matters is whether the clock is dropping, not the digit on the sensor.

Finally, forgetting to revert. Some people spend days troubleshooting a crash caused by an offset they forgot they set. Reset to defaults before you debug anything else.

Tips for Running Modded Games at Lower Temperatures

Undervolting is one lever, not the whole answer. The GPU is usually the bigger heat source in a modded game, and a few settings there pay off quickly.

Cap your GPU clock in MSI Afterburner instead of leaving it on boost. Holding the card at a slightly lower clock smooths out the load spikes that ENB and physics mods produce, and it often improves the worst frametimes more than a higher average would.

Watch the VRAM side of the equation. Mod lists with high-resolution textures can push frame rate down without touching CPU or GPU temperature at all, and no amount of tuning fixes that. Lowering a texture pack’s resolution or trimming the first-person shadows is a bigger win than another 10 mV.

Run the manufacturer performance profile rather than a silent or balanced one while gaming. Silent modes cap wattage hard and trade a lot of frame rate for noise reduction you do not want during a long session.

Physical setup matters more than most people expect. Keep the laptop on a hard surface with the rear vents clear, never on a bed or your lap, and expect a measurable drop within minutes. A basic cooling pad adds more than most offsets.

Close what you do not need. Launchers, voice chat, browser tabs with 40 mod-list pages open and cloud sync all compete for the same thermal headroom during load.

Check the vents before you check the offsets. A laptop that has been open for a few years is usually full of dust, and a thick paste that has gone stiff looks identical from the outside. Plenty of people who saw zero change at an aggressive offset found the real fix was cleaning the fans or repasting.

Set one profile per game. A ThrottleStop profile for a heavy physics mod list can run a deeper offset than the one you use for a lighter title, and switching is a single click once the profiles exist.

Frequently Asked Questions

Start simple: confirm support, record a baseline in the modded game, set a -100 mV core and cache offset, and test it properly before going deeper. Most modders settle between -100 mV and -150 mV and get a laptop that holds its frame rate for hours instead of throttling after half an hour. Save the profile, keep the reset path in your notes, and everything you change stays reversible.