How to Overclock a CPU

(Image credit: Intel)

Contrary to innumerable reports of its demise, overclocking is not dead — not by a long shot. Yes, overclocking headroom for core frequencies has receded as the Intel vs AMD rivalry has intensified and the chipmakers focus on squeezing out every ounce of performance, particularly in the highest-end flagship processors. However, Intel’s Alder Lake chips represent a return of generous overclocking headroom, helping the chips take key spots on our list of best CPUs for gaming. AMD’s latest chips don’t have as much headroom for manual core frequency overclocking — the company’s automated overclocking features are best for tuning — but like Intel, the company exposes a wealth of other tuning options for memory and fabrics, which can give you a nice bump in performance.

Should you overclock your CPU? Well, there are a set of best practices that you should follow when you overclock your processor, and if you take a reasonable approach, the risk is minimal. However, be advised that excess voltage can damage your chip and it voids the warranty on both Intel and AMD processors, so overclock with care. As always, you’ll be at the whims of the silicon lottery when it comes to the maximum overclock you can squeeze out of your chip — some chips simply overclock better than others, even when they are otherwise identical. 

Today, we’ll show you how to overclock your CPU and teach you how to unlock the hidden performance lurking under your heatspreader. But first, for the ‘overclocking is dead’ crowd, we have overclocking benchmarks of the performance increases we attained in different types of workloads. Before we explain how to overclock your chip, let’s take a look at our performance results.

CPU Overclocking Benchmark Results

Here are the results of our overclocking with Intel’s Alder Lake chips compared to the Ryzen 5000 lineup in Windows 11, along with DDR4 vs DDR5 benchmarks and overclocked configurations. You can find more detailed breakdowns of our overclocking with the Core i9-12900K and i5-12600K here, and the Core i7-12700K details are here. Do note that we’ve overclocked the memory with these configurations, too. 

We generated these overall measurements of gaming performance as a geometric mean of our entire test suite. We also selected the most important single- and multi-threaded tests in our suite to generate those cumulative measurements. You can see an even more expansive view in our CPU Benchmark hierarchy.

Alder Lake vs Ryzen 5000 Overclock Uplift – Stock configuration used as baseline – Windows 11
Tom’s Hardware – %age Change 1080p Gaming Single-Thread Multi-Thread
Core i9-12900K DDR4/ DDR5 +9.7% / +5.2% +1.6% / +3.2% +3.3% / +7%
Ryzen 9 5950X +5% -2.3% +5.7%
Core i7-12700K DDR4/ DDR5 +9.8% / +7.1% +2.3% / +2.1% +3.9% / +6.4%
Ryzen 9 5900X +3.7% -0.6% +2.1%
Core i5-12600K DDR4/ DDR5 +15.2% / +12.9% +4% / +4.2% +8.8% / +11.3%
Ryzen 5 5600X +6.7% +3.8% +2.7%

As you can see above, the Alder Lake chips profit more from overclocking than the AMD Ryzen models, but both lineups do experience at least some improvement. The Core i5-12600K is a standout with a massive 15% gain in 1080p gaming performance. That’s more impressive than the 9% gain you’d get from spending $110 more for the Core i7-12700K. These lower-end chips are comparatively easier to overclock, too, and don’t require the highest-end cooling to pull off impressive results.

There are plenty of gains to be had for those at higher chip tiers, too. The Core i7-12700K shows similar overclocking headroom to the more expensive Core i9-12900K. After overclocking both chips, the 12700K is within 1% in gaming to the overclocked 12900K which costs a whopping $180 more. 

To represent what we think a normal user can achieve, we chose relatively mundane and simple overclocks for these chips — yet we still pulled off comparable or better gains in gaming than we would expect from moving forward to a new generation of CPUs. Even though we can’t guarantee that your results will match ours, it’s clear that overclocking is alive and well. Here’s our guide that shows the steps we took to reach these types of overclocks. 

CPU Overclocking Prerequisites and Risks

Before we start turning up the dial on the voltages (and fans), you’ll need to make sure that your system is ready for overclocking. As always, we have to caution you that overclocking voids the warranty on any processor, and you run the risk of damaging your chip if you apply excessive voltage. Overclocking also increases power consumption and heat generation, so you’ll need to accept and accommodate those needs. Excessive voltage and heat can also result in reduced chip lifespan due to premature degradation, so you’ll want to stay within reasonable boundaries.

Keeping your CPU as cool as possible is one of the keys to attaining the highest CPU overclocks. Check out our Best CPU coolers article for recommended options, and be sure to use one of the Best Thermal Pastes to ensure your cooler is effective. Ensuring that your case has adequate ventilation is also key so be sure to ensure that you have enough airflow. 

As a general rule of thumb, more is better for cooling; a CPU cooler that can handle 40% more TDP than your CPU’s rating is preferred. However, having less cooling headroom won’t prevent you from extracting any gains at all — the chip’s peak temperatures will just limit how much you can overclock via the core voltage and frequency. The definition of sufficient cooling can vary based on your personal preference, but your overriding goal should be to prevent thermal throttling, a process that reduces the processor’s clock speeds and voltage to prevent damage (killing your chip) from excessive temperatures. We’ll dive into that shortly. 

Intel’s overclockable chips don’t come with a bundled cooler, but some of AMD’s models come with pretty competent coolers right in the box. However, you’ll need to assure that these coolers can handle overclocking, so be sure to check our reviews for each model.

(Image credit: Corsair)

Overclocking all of the CPU cores at once (‘all-core’ overclocking) is the most common and easiest method of overclocking, but it does tend to generate the most heat. As a general rule, it’s preferable to have a 240mm All-In-One (AIO) liquid cooler (or air cooler equivalent) for all-core overclocking with a modern Core i5 or Ryzen 5, and you’ll want a more powerful 280mm AIO or better to wring out the most performance possible on the higher-end Core i7, i9, Ryzen 7 and 9 SKUs. Overclocking cooling requirements can vary based on the generation of chip you’re tuning, so be aware those guidelines don’t apply to all previous-gen chips. More elegant overclocking approaches that don’t use brute-force all-core overclocking methods, like manipulating turbo ratios or only overclocking a few cores, can also extract extra performance even if you’re using a lesser cooler — you just have to pay close attention to your CPU temperature when you dial in the overclock. We’ll also cover those methods below.

Naturally, you’ll need an overclockable processor. AMD’s generous overclocking policy means that you can overclock nearly any chip (the Ryzen 7 5800X3D and a few Athlon models are the only notable exceptions). For Intel, you’ll need a K-series chip if you plan on increasing the chips’ core frequency, which is the most basic method of overclocking. That’s because K-Series chips have an unlocked multiplier that allows you to easily dial up the frequency on your chip. In addition, the graphics-less ‘KF’ models are also overclockable. If you don’t have a K-series chip, your options for overclocking will be far more limited, though you can still aim for higher memory clocks with the last few generations of Intel processors.

AMD allows overclocking on any chipset except the A-series motherboards. For Intel, if you plan on doing full core frequency overclocking you’ll need a Z-series motherboard, as Intel doesn’t allow you to change the chip’s frequency on cheaper B- and H-series motherboards. Most higher-end motherboards have robust power delivery subsystems but performance varies, so pay attention to motherboard reviews to find your best option. You can hit our list of Best Motherboards to see the best models on the market.

Listed here last, but certainly not least, you’ll also need to ensure that you have one of the best power supplies for your system, but your requirements will vary based on the other components in your system. You can see the basic guidelines with a power supply calculator, but be sure to enter the maximum overclock frequency and voltage to ensure you have plenty of room for overclocking. Having plenty of power headroom, and clean power, is critical, so don’t skimp on the power supply. 

Measuring Baseline Thermals and Performance

Now that we have the prerequisites sorted, it’s important to establish a performance and thermal baseline. You’ll use this to measure how much impact an overclock has on both CPU performance and heat, allowing you to determine the acceptable tradeoffs for the amount of performance you gain.

There are a plethora of software options for stress testing and monitoring — see our how to stress test your CPU guide for additional details. Some, like AIDA64 or OCCT, have in-built stress testing and monitoring, while others, like HWInfo, are purely designed to monitor performance. The chipmakers also provide their own software: Intel has Intel’s eXtreme Tuning Utility (XTU) software, while AMD provides its Ryzen Master software. Both of these applications allow for monitoring and changing parameters, but other functions, like stress testing, vary.

(Image credit: Future)

Increasing the chips’ frequency through overclocking requires pumping more power through the chip, thus generating more heat, and higher frequencies typically result in faster aging, and thus lowered life span. As we’ve outlined in our how to check your CPU temperature article, it’s best to keep most chips below 80C during load and under 30C at idle to minimize long-term wear on the processor. However, AMD’s Ryzen 5000 processors are designed to run at up to 95C with a stock cooler, while Intel’s highest-end Core i9 Alder Lake processors can run up to 100C during normal operation. So you’ll need to do a bit of research to find the correct threshold.

Stress tests often merely serve as power viruses that will stress your system beyond what you would encounter in normal use, so it’s best to use reasonable utilities and/or intense multi-threaded applications that you would find in normal PC use. We prefer AIDA64 and OCCT for quick synthetic stress testing, but employ the HandBrake and Blender applications for extended duration testing. Now that you’re ready, kick off your stress test and let it run until temperatures stabilize, then log the final measurements.

Next, you want to set a performance baseline. The most general benchmarking rule is that the best performance benchmark is to simply measure the performance of the programs you use the most. However, those often don’t have built-in benchmarks. In that case, you can also use similar types of programs (renderers or encoders, for instance) as a proxy for your workload.

Synthetic gaming benchmarks don’t tend to translate well to real-world gaming, but given their stability and repeatability, these are great benchmarks for comparing performance before and after any changes you may make to your system. Be sure to turn off as many background tasks as possible during your benchmarks to eliminate that influence from your CPU benchmark results. Here are a few common benchmarks, but you can see a more expansive list in our CPU Benchmark article.

You can get away with just CineBench, but the more the merrier. Run your tests and log the results. You’ll use this information to compare to later after you’ve overclocked the processor. 

How to Change CPU Overclocking Settings

Overclocking requires manipulating several system parameters, like voltages and clock speeds. You can make changes via software inside Windows with utilities like Intel’s XTU or AMD’s Ryzen Master, or you can enter the values directly into the system BIOS/UEFI. It’s pretty simple to enter the BIOS to overclock, on the majority of platforms, you simply reboot the system and click delete or F2 repeatedly as it restarts.

Both approaches have their strengths and weaknesses. Software overclocking is a bit simpler because it uses a standardized nomenclature for the various settings, whereas motherboard vendors can use different names for the same settings (luckily the BIOS tends to have a short descriptor for each option). Additionally, overclocking via software allows you to make changes in real-time. In contrast, changing the values in the BIOS requires a system reboot before you see the impact.

Overclocking your CPU via the BIOS does have one big advantage, though: There are far more fine-grained options available for more advanced tuners. That means experienced tuners are better off using the BIOS if they plan to use the more advanced features. Most die-hard overclockers stick with BIOS overclocking and use software tools for monitoring.

It is important to save your BIOS settings before you proceed to make changes. Given the nature of overclocking’s trial-and-error process, you might need to restore these settings several times during the process. Most motherboards let you save your settings into a profile that you can later restore, and you can assign simple names to keep track of multiple profiles. If you reach certain solid overclocked settings but would like to push even higher, it makes sense to save a profile for that overclock so you can easily revert to a known-stable configuration if needed. 

Here you can see the BIOS options for overclocking an Intel Alder Lake CPU on an MSI Z690 board. While the names for certain settings can vary somewhat based on your motherboard vendor, the major manufacturers (Asus, ASRock, Gigabyte, and MSI) all include a wealth of options in their enthusiast-class boards. Depending on your overclocking goals, you can go as deep as you want on a top-tier motherboard, but the basics aren’t nearly as daunting as the wealth of options might suggest.

There are a plethora of settings and voltages that you can manipulate for overclocking. For the scope of this article, we’ll only focus on the basic settings that you’ll need to get your overclock up and running. We’ll refer to these settings in the following sections, but we have provided a glossary of key BIOS terms at the bottom of the article. 

Removing the CPU Overclocking Power Limits

(Image credit: Shutterstock)

The first step to overclocking an Intel chip is to uncap the power limits imposed by the motherboard. For MSI motherboards with Intel’s newer processors, these settings are listed in the BIOS/UEFI as the Long Duration Power Limit, Short Duration Power Limit, and the CPU Current Limit. You should enter the first two values as 4096W, and the latter value should be set to 512A. Finally, set the Long Duration Maintained value to the longest allowed (128 seconds).

The names of these settings can vary slightly by BIOS, but you can also change these same values in XTU — they’re listed as Processor Core IccMax (set to unlimited), Turbo Boost Power Max (set to unlimited), and Turbo Boost Power Window (128 Seconds). Finally, disable “Turbo Boost Short Power Max Enable.”

You’ll never reach these levels of power usage, but removing all power caps allows you to push your silicon to the limits.

AMD has similar power limits in its PPT, TDC, and EDC settings, but these settings are most often manipulated in tandem with the company’s auto-overclocking Precision Boost Overdrive software, which we’ll cover shortly.

How to Overclock a CPU

Overclocking your system memory is a must-do item for tuners, particularly if you plan on gaming. Still, it’s best to handle memory overclocking after you find your preferred CPU core overclock frequencies. This limits the number of variables you will have to troubleshoot as you dial in your CPU overclock. Once you’ve dialed in your CPU overclock, head over to our How to Overclock RAM guide.

Before we get started, know that the all-core overclocking method described below can give you great results with Intel processors, but AMD chips tend to not overclock as well (hopefully this changes with Zen 4 Ryzen 7000 chips). You can still try your hand at Ryzen overclocking with this method, but the company’s auto-overclocking Precision Boost Overdrive feature is your best bet. Additionally, Intel’s latest chips use two different types of cores. We’ll cover both of those topics further below. We always recommend manual tuning, but Intel has its one-click auto-overclocking Intel Performance Maximizer (IPM) tool available for some chips.

Unless you’re an experienced overclocker, save yourself some time and only change one parameter between stress testing sessions. This simplifies adjustments and troubleshooting.

Here we’ll cover the simplest manual method — all-core overclocking. Look further below for instructions for other techniques, like Multi-Core Enhancement, Turbo Ratio, and Per-Core overclocking. 

1.) Change the CPU ratio multiplier to your desired frequency — You use the CPU ratio multiplier to dial in a specific core frequency. The core frequency consists of the BCLK value (which is typically at 100 MHz if you haven’t modified it) multiplied by the CPU ratio multiplier value. For instance, a 100 MHz base clock with a 50x multiplier equates to 5,000 MHz, better known as 5GHz. We typically increase this by 1 to 2x per attempt at first, then drop down to 1x increments as we begin to encounter instability. 

2.) Run a stress test — Run your stress test for a period of time to see if you stay within a safe temperature range, get a blue screen or otherwise encounter system errors. Don’t get discouraged if you get a BSOD or other errors — this is a trial-and-error process, so it is expected.

If the system is stable and temps are stable, increase the CPU ratio multiplier again and repeat the stress test. Repeat this process until you encounter errors or BSOD.

For Intel chips that support the AVX offset feature: while it isn’t a strict requirement, we recommend that you run a stress test without AVX instructions. You can easily disable AVX instructions in most common stress testing applications, like AIDA64 or OCCT. AVX instructions impart maximum heat and power consumption upon the chip, thus causing instability that requires higher voltages. However, you can tune the overclock to minimize its impact, thus yielding higher peak overclocks. We recommend that you reach your peak overclock without AVX first, and then you can dial in an AVX offset a few more steps down the list. AMD doesn’t have an AVX offset, so you can run any type of stress test you would like. 

At that point, proceed to the next step. 

3.) Increase the CPU voltage (vCore) — Now you begin to slowly increase the voltage to bring the CPU back to stability. First, set the CPU voltage mode to ‘override.’ As a general rule of thumb, you should start with a vCore of 1.25V or lower, and then run a stress test (if the system boots). If the system is stable, you can then increase your CPU ratio multiplier again and repeat steps 1 and 2. If you BSOD, continue to work your way up in very small 0.01V increments until you reach stability, stress testing after every increase.

Maximum voltages vary based on the generation of processor that you’re overclocking. A general rule of thumb is not to exceed 1.40V unless you’re using exotic (sub-ambient) cooling, but you should do a bit of research to find the maximum voltage for your generation of chip. Regardless, heat is the enemy here.  Higher voltages and heat will result in faster chip degradation, so you need to carefully monitor CPU temperatures during your stress tests to ensure that you aren’t exceeding the safe zone or reaching the throttle point.

There is no magic formula when it comes to overclocking. If you want to pinpoint the exact voltage for stability, use small increments of 0.01V. If you’re not the patient type, you can work with higher increments, like 0.05V.

Be aware that temperatures will rise, and frequency improvements will decline, on a non-linear basis with voltage increases. That means you’ll get less of a return in exchange for more heat as you work your way up to higher voltages. We suggest that you don’t use the borderline voltage for long-term settings. This isn’t an exact science and hardware is unpredictable, so we advise taking a ‘better-safe-than-sorry’ approach. 

4.) Configure the voltage mode — Once you’ve reached your peak overclock, you can either proceed to the next step or do a bit of experimentation with the voltage mode. Static voltages (often listed as ‘override’) provide a steady stream of power to the processor, which is the easiest method of overclocking, and best for dialing in your initial overclock. Motherboard makers have several different voltage modes on offer, and you can give those a try after you’ve solidified your overclock. Adaptive mode is popular because it can make the processor generate less heat and consume less power when it isn’t under a heavy load. You can dig into your motherboard manual to learn more about any unique voltage modes.

5.) Configure the AVX Offset — Intel — This setting reduces the CPU multiplier, and thus frequency, during workloads that use AVX instructions. Once you’ve reached your stable peak frequency, you can use a stress test with AVX enabled to see if your system remains stable during an AVX workload. You can slowly reduce the AVX offset by -1 increments, which reduces speed by 100 MHz per step, until you reach stability. It isn’t uncommon to see overclocks with -3 or -4 AVX offsets.

Optional — Adjust Load Line Calibration (LLC) — This type of setting varies based on the motherboard vendor, but setting a mid-range LLC value can help solidify an otherwise sketchy overclock. Experimentation helps here, but most newer motherboards are adept at automatically adjusting this value if left on the ‘Auto’ setting. Adjusting the LLC is optional, but can help to push a bit further in some cases. 

Optional — Disable Intel SpeedStep — You can either have the chip always run at its overclocked frequency, or have it drop to a lower clockspeed during idle or low-load conditions. If you leave SpeedStep enabled, the ‘High Performance Windows power plan will not allow the processor to shift into a lower frequency, so you’ll need to enable the ‘Balanced’ plan to enable downclocking. 

Optional – Memory Overclocking — After you’ve reached your best stable CPU overclock, feel free to proceed to overlock your memory. This pays off quite a bit, particularly if you’re into gaming. Most users can simply enter the BIOS and enable an XMP profile if the memory kit has a profile. 

If your overclocked system is unstable when you activate XMP, it might be necessary to tweak the VCCIO and VCCSA voltages. These two voltages are helpful when you want to stabilize a memory overclock. Be warned, though, VCCIO and VCCSA are touchy, meaning too much voltage can be just as bad or worse than not enough. It is best to tweak these voltages with small increments of 0.01V until your memory overclock is stable. You can also head here for more detailed memory overclocking instructions. 

How to Overclock a CPU With Multi-Core Enhancement (MCE)

Intel’s motherboard partners have infused their boards with predefined all-core boost profiles that go by many names, such as Multi-Core Enhancement (MCE) with ASUS motherboards and Enhanced Turbo with our MSI motherboard. These features are largely referred to as MCE, but the functionality remains the same: These settings essentially apply an all-core overclock to the processor that is defined by the maximum Turbo Boost bin supported by the processor. This setting modifies the CPU’s clock rate and voltage to deliver higher performance, which is basically factory-sanctioned overclocking. Performance, power consumption, and heat are all affected, naturally.

Motherboard vendors predetermine the voltage settings at the factory, meaning the settings do not take chip quality into account. Instead, the vendor bins a large number of CPUs in each respective SKU and sets the parameters based on the worst common denominator.

As such, these settings typically use a much higher voltage than required for even chips of “normal” quality, which can reduce the chips’ lifespan and result in a hotter and noisier system. We always recommend manual tuning over MCE approaches, but if you have sufficient cooling and aren’t as worried about heat generation and power efficiency, this is the quickest method. 

Intel P-Cores and E-Cores CPU Overclocking

(Image credit: Intel XTU / Tom’s Hardware)

Alder Lake has P-cores for latency-sensitive work that tends to be lightly threaded, while the E-cores step in for multi-threaded work and background tasks. The E-cores can only be overclocked in groups of four, while P-Cores can be overclocked individually or in groups. Alder Lake provides plenty of options for fine-tuning — you can disable the E-cores entirely, which often allows you to eke out a slightly higher overclock (typically a single bin) on the P-Cores.

Choosing whether to disable the E-cores will depend on your own personal preference, but leaving both the P-cores and E-cores active will offer the best blend of performance for most users. However, this means that you’ll have to overclock them separately. Head over to our How to Overclock Alder Lake CPUs guide for more detailed instructions. We also have more generalized advice for Intel processors in our How to Overclock an Intel CPU article. 

All-Core, Per-Core and Turbo Ratio CPU Overclocking

You can overclock the CPU frequency in three ways: All core, Per core, and via Turbo Ratios, with the last option only available on Intel processors. The ‘all core’ setting is what we traditionally associate with overclocking. ‘All core’ is the simplest method by far because it assigns one static frequency to all cores at once. ‘Simplest’ doesn’t always translate to ‘best,’ though.

Overclocking via the Turbo Ratios is one of the best ways to dial in a refined overclock, as this allows you to define the peak boost frequency based on how many cores are active. This feature can help you eke out a slightly higher overclock by targeting more robust cores with higher frequencies, but just as importantly, it allows the processor to drop back into its base frequency when the chip isn’t under load. This allows the chip to run cool when it isn’t busy and also reduces the amount of time the chip is at the highest frequencies, which is important for chip longevity (more on that below). 

If you overclock via the turbo ratios, you’ll need to make sure that your Windows power profile is set to ‘Balanced’ or lower (the ‘High Performance’ profile keeps the chip at its peak turbo frequency at all times). You can use our above step-by-step guide to overclock using this approach, but instead of modifying the CPU ratio multiplier in Step one, simply modify the Turbo Boost multipliers instead.

The ‘Per Core’ feature allows you to assign a unique frequency to each individual core. This can be helpful if you identify that some cores are more capable of sustaining a higher frequency than others. This setting is most useful for advanced tuners and can require a fair amount of investigative work to determine the appropriate clock speed for each core. In this case, you would cycle through each core and target it individually with a stress test as you work your way through the steps above, finding the peak for each core.

AMD Precision Boost Overdrive (PBO) CPU Overclocking

(Image credit: Tom’s Hardware)

AMD’s Ryzen chips don’t have much manual overclocking headroom available, largely because the company’s boosting algorithms automatically expose the most performance possible given the capabilities of your motherboard’s power delivery subsystem and your cooler. However, the company’s auto-overclocking Precision Boost Overdrive (PBO) feature helps increase performance in a mostly-automated fashion. 

AMD defines three types of power limits for its chips: PPT is the maximum power consumption allowed, TDC is the maximum sustained current, and EDC is the maximum burst current. You can override those settings either manually or with AMD’s PBO. You can access this feature via either the BIOS or Ryzen Master software. 

PBO typically doesn’t deliver huge performance gains if you adhere to the basic presets. The basic “enabled (PBO on)” preset enables significantly higher default PPT/TDC/EDC limits, but doesn’t change two important settings: PBO Scalar or Clock.

PBO Scalar overrides the AMD default settings and allows increased voltage at the maximum boost frequency and lengthens boosting duration. Changing the PBO Scalar setting unlocks the best auto-overclocking performance, so the basic preset can be lacking. You can increase this value in increments of 1X, but we typically skip right to the 10X (maximum) setting. The PBO ‘Clock’ setting also allows the CPU to exceed its standard boost by a defined variable. We usually max this setting out, but it does have a limited impact.  

You can also use the “PBO Advanced” profile that defines the limits of each motherboard based on the capabilities of the power delivery subsystem (as defined by the motherboard vendor). This setting exposes the highest PPT, TDC and EDC settings for the motherboard, but also doesn’t change the PBO Scalar and Clock settings. However, this setting does allow you to change the PBO Scalar and Clock settings manually, with the former usually unlocking much higher auto-overclocking potential. We find that using the PBO Advanced setting with adjusted PBO Scalar and Clocks values yields the best benefits. 

Ryzen also profits handsomely from memory overclocking. As such, you should pay particular attention to tuning the chips for the best possible memory frequencies with the lowest timings. We have more AMD-specific advice in our AMD Ryzen Overclocking Guide

CPU Overclocking Impact on Lifespan and Reliability

Legendary overclocker Allen “Splave” Golibersuch in action at the lab (Image credit: Tom’s Hardware)

Will overclocking kill your CPU? Not if you follow common-sense steps and take a conservative approach. There are settings and techniques that overclockers can use to minimize the impact of overclocking, and if done correctly, premature chip death from overclocking isn’t a common occurrence.

Intel’s overclocking guru Dan Ragland has given us specific advice when it comes to overclocking when we visited the company’s overclocking lab. We’ll share an excerpt of those learnings here:

Every semiconductor process has a point on its voltage/frequency curve beyond which a processor will wear out at an untenable rate. If the chip wears enough, it triggers electromigration (the process of electrons slipping through the electrical pathways), which leads to premature chip death. Some factors are known to increase the rate of wear, such as the higher current and thermal density that comes as a result of overclocking.

All this means that, like the carton of milk in your refrigerator, your chip has an expiration date. Because increasing frequency through overclocking requires pumping more power through the chip, thus generating more heat, higher frequencies typically result in faster aging, and thus lowered life span. Intel’s overclocking team recommends using adaptive voltage targets for overclocking and leaving C-States enabled. Not to mention using AVX offsets to keep temperatures in check during AVX-heavy workloads.

The amount of time a processor stays in elevated temperature and voltage states has the biggest impact on lifespan. You can control the temperature of your chip with better cooling, which then increases lifespan (assuming the voltage is kept constant). Assuming voltage remains constant, each successive drop in temperature results in a non-linear increase in life expectancy, so the ‘first drop’ in temps from 90C to 80C yields a huge increase in chip longevity. In turn, colder chips run faster at lower voltages, so dropping the temperature significantly by using a beefier cooling solution also allows you to drop the voltage further, which then helps control the voltage axis.

In the end, though, voltage is the hardest variable to contain. Ragland pointed out that voltages are really the main limiter that prevents Intel from warrantying overclocked processors, as higher voltages definitely reduce the lifespan of a processor. But Ragland has some advice: “As an overclocker, if you manage these two [voltage and temperature], but especially think about ‘time in state’ or ‘time at high voltage,’ you can make your part last quite a while if you just think about that. It’s the person that sets their system up at elevated voltages and just leaves it there 24/7 [static overclock], that’s the person that is going to burn that system out faster than someone who uses the normal turbo algorithms to do their overclocking so that when the system is idle your frequency drops and your voltage drops with it. So, There’s a reason we don’t warranty it, but there’s also a way that overclockers can manage it and be a little safer.”

That means manipulating the turbo boost ratio is much safer than assigning a static clock ratio via multipliers. As an additional note, you should shoot for idle temperatures below 30C, though that isn’t much of a problem if you overclock via the normal turbo algorithms as described by Ragland.

Motherboard BIOS and UEFI Overclocking Settings

It’s pretty simple to enter the BIOS to overclock, on the majority of platforms, you simply reboot the system and click delete or F2 repeatedly as it restarts. Once inside the BIOS, you’ll find settings such as these, or their equivalents. 

  • Base Clock (BCLK) – The frequency at which the processor communicates with the memory and PCIe devices. The default BCLK for Intel chips is 100 MHz, but you can adjust this for smaller incremental performance increases. Be aware that adjusting the base clock also impacts the PCIe and memory busses, so you should refrain from adjusting your BCLK until your overclock is stable. Even then, it would be best if you did so sparingly.  
  • CPU Ratio Multiplier – Dictates the ratio between the CPU and the BCLK. The formula to determine the processor’s frequency consists of multiplying the base clock by the CPU multiplier. For example, a processor with a 100 MHz BCLK with a multiplier of 50 will operate at 5,000 MHz, or 5 GHz. 
  • CPU Core Ratio – This lets you choose whether you want to set the multiplier for all the cores in a group, or individually. The latter is referred to as per-core overclocking, and it allows you to tune individual cores to their highest potential instead of the lowest common denominator. This approach can also allow you to squeeze at least some overclocking headroom out of systems with lesser coolers. 
  • Vcore – This voltage goes by many names, like Core Voltage or vCore, but it always represents the motherboard’s main input voltage to the processor. This value has the most direct input on thermals, with higher amounts of voltage generating more heat. 
  • Voltage Mode – Auto lets the motherboard decide the voltage to apply to the processor while ‘Manual’ or ‘Override’ allows you to assign a fixed Vcore. Offset mode adds a specific amount of voltage to the processor regardless of the frequency, while Adaptive voltage increases the voltage when the processor operates in turbo mode. 
  • AVX Offset – A separate multiplier that can adjust the processor frequency when it executes AVX workloads. AVX instructions yield massive speed-ups, but these instructions also generate more heat and consume more power than other types of instructions, which can lead to system instability during overclocking. Most software and games do not use AVX instructions, so dialing the AVX offset back to reduce the core frequency during these taxing workloads is critical to attaining peak performance in non-AVX applications. 
  • Load-Line Calibration (LLC) – Sometimes, typically when the processor is first placed under load, the CPU doesn’t receive the amount of voltage set by the user. This condition is caused by voltage droop (Vdroop) and it can result in either lower or higher voltages than intended. Load-line calibration basically compensates for Vdroop by assuring that voltages remain at a more even level. There are multiple LLC options in most motherboards, but Auto typically suffices for most users with higher-end (or newer) motherboards. 
  • Intel SpeedStep – Feature that increases or decreases processor speed and voltage according to system load. 
  • Intel Uncore – Regulates the frequency of the different controllers on the processor like the L3 cache, ring bus, memory controller, etc. 
  • FCLK – Intel – Controls the speed at which data is passed from the processor to the graphics card. AMD – Specifies the Infinity Fabric frequency (important for memory overclocking) 
  • VCCSA – Voltage for the System Agent. Increasing this voltage can help stability when overclocking the ring bus and cache frequency. 
  • VCCIO – Voltage for the memory controller and shared cache. 
  • Extreme Memory Profile (XMP) – Enables the XMP profile on compatible memory kits. XMP profiles apply pre-validated memory overclocks by simply toggling the feature on in either the BIOS or a software overclocking utility.