Understanding and maximizing your PC's performance

How to determine the actual root bottleneck? You have two main methods.

1. Specifications. If you know how to read a computer specification sheet, it would not be difficult to realize the approximate power of a component, their interactions with other components, mainly the performance output - understanding what components are faster or slower than others can quickly lead you to the culprit of a performance bottleneck.

2. Usage, system loads. Tools such as “Task Manager”, “KSysGuard”, “htop” among others use services integrated into your operating system to semi-accurately display the status and usage statistics of your components. They can tell you how much of your total GPU or CPU processing capacity is used, even for memory or your storage. If your CPU is sitting at 40% load on one core where the game is running, yet your graphics card is sitting at 100% usage constantly - from what you've learned so far, it indicates that the GPU is the one overwhelmed by your task, while other system components have no trouble handling the load, and are even forced to hold back due to the GPU not being able to catch up.

IMPORTANT: It's important to understand concepts such as multithreading and how logical and physical processing works. In all modern processors, there is more than 1 core (4-16 in modern computers), and with multithreading enabled, there are twice as many logical processors. The cores in a processor are, well, actual separate processors, when you load a game or program, what happens is that it's loaded to run on 1 of the cores - if you have 4 cores then the program is technically going to use only 25% of your total CPU power. The reason that most games are single-threaded is that having a program take advantage of multiple logical CPUs (threads) or physical CPUs (cores) is quite difficult. In the most basic form, what needs to be done is that the developer has to have a separate program or “worker” in another thread or core which can then be used to offload certain tasks to, for example, character AI. This is extremely difficult in games, but is often not an issue in professional video applications and rendering software such as Blender, as the logic there only involves plain tasks that can be easily parallelized, a concept known as embarrassingly parallel tasks.

Due to the nature of single-threading in games, what can happen is that both your CPU is sitting at not full usage, and so is your GPU, even though not that many frames are being generated. In a lot of cases, this indicates a software bottleneck. This is when the software or game you are running physically cannot take advantage of your system components, this is mostly seen in multicore processors that sacrifice single-core performance for just a larger amount of cores, which is beneficial to editing and professional environments, but not gaming.

How are bottlenecks resolved? Might be a bit disappointing to low-end and cashless individuals.

Bottlenecks cannot be easily resolved. Just thinking about it should tell you why - it is a problem with your hardware, the only solution would be to somehow increase the performance of said hardware, either by replacing/upgrading the hardware, overclocking it, or ensuring that it is at its full potential by providing good voltage and good cooling to it.

(2) EVALUATING SOFTWARE RESOURCE HOGGING

Software refers to interchangeable programs/instructions running usually under the hood of your operating system. While in a computer you may be focusing on only one task - for example, a game, there's much additional software working in the background. Anything from anti-malware “antivirus” software, to chat apps (Discord, Skype, Element, Keybase), to operating system processes to track system events, perform updates, cloud synchronization, system monitoring, desktop management, and other countless reasons. This makes the operating system operate as it should, but when those are way too plentiful or arrogant with their system resource usage (cough cough entire Windows operating system cough cough), oftentimes the processes combined may be using a substantial amount of your CPU cycles (total usage) or taking quite a sizeable chunk of your memory. This can and often does decrease performance in games, as less processing power or memory is available to the game, and more so to random system processes and background apps.

It is a useful and easy to learn skill to understand the impact of certain applications and processes running on your system. For example, it's worth it to understand that having a web browser open will impact your memory, even more so if you have video content playback loaded somewhere in it (Ex. YouTube), for higher resolution video - there's going to be a higher amount of CPU power being used due to video decoders needing a substantial amount of CPU cycles, so if you want to listen to music - either lower the resolution, or better yet, use music services like music.youtube.com instead, or download it all locally. Applications like Discord have memory leaks (referring to bad memory management practices that end in unused data not being freed from memory), which oftentimes ends up using upwards of 600-1000 MBs of memory.

The best practice here is to limit what you have in the background ranking them in a sort of “is this necessary enough to run while I get this amount of a performance hit?” manner.

Your toolkit here is the same as was specified before - for Windows, you will mainly be using “Task Manager”, which you can access from the start menu, the hotkey CTRL+SHIFT+ESC or the CTRL+ALT+DEL menu. You can sort running processes by their hit on system memory, CPU, disk, GPU. Keep in mind not to go and start killing random processes, that can cause your system to become unstable and crash. If you don't know what a process does, don't mess with it until you've read about it and can confirm that it's not system-critical.

(3) SETTINGS, PERFORMANCE TINKERING

The section is in a pre-release state, the information provided here may have not been fact-checked fully yet, or is incomplete / not ready for proper usage.

Awaiting contributions or help from individuals experienced in game configuration, driver-level game optimizations, and OS-specific performance options.

The default settings are almost always bad, and this applies to a lot of scenarios - whether you're running an operating system, random software, or a video game. There is always room for optimization, tinkering, and other options that can increase your performance, productivity, or other stats. 

The most important settings to touch are the settings in games as well, to play a game on low-end, it is always important to compromise quality for performance gained. Changing settings to low does not always get you the best performance or best low-graphics visual quality. Low settings often set everything to low, regardless of the impact on performance. Configuring settings yourself can often yield better results, as simply looking up the graphical settings in question online can yield you quality information about the setting, what exactly it does, and whether it has a performance impact that would be noticeable to you if you left it enabled.

What most individuals venturing into optimizing game-specific quality often miss is driver-controlled graphical quality. Pretty much all graphics drivers (Nvidia's, Intel's and AMD's) offer control panels in their drivers, which allow you to configure your ‘preferred performance' and override 3D applications' settings, which allows you to force downscaling, turn off certain sampling and antialiasing features and more, depending on the exact driver that you have. If configured properly, this can allow you to override certain limited or off-the-books settings that games force to be enabled, thus increasing performance.

You can most often enter driver control menus by right-clicking on your desktop and seeing an entry for your graphics vendor's settings (Windows only).

Mobile (laptop) low-end users may benefit from having their CPU power plans changed and set to the maximum while having ‘high-performance' mode on (Windows only). This can be achieved by entering the ‘Choose a power plan' settings menu through the search. In it, right off the start, you are supposed to change the power plan from ‘balanced' to ‘high performance'. When that is completed, enter plan settings through the ‘Change plan settings' button on the right side next to the high-performance entry, click on ‘Change advanced power settings', next up navigate to the ‘processor power management' dropdown menu, and change the minimum and maximum processor state to 100%, also ensuring that the system cooling policy is set to active.Most more advanced games don't just have limited in-game settings, they also have configuration files where the actual settings changes are recorded. It differs greatly between games, but most of the time the setting configuration file allows you to change your settings below or above that if what the in-game menu allows you to, thus possibly allowing for even more harsh optimizations to be applied.
If you have the need to apply harsh optimizations for games, for example in the cases of having a seriously low-spec machine, then it is worth a try to attempt to search around for custom configs and config guides for whatever specific games you play. Do keep in mind, a lot of these ‘low spec' configs just make sure that your game runs - visual quality is often completely discarded in favor of performance.

Windows users benefit heavily in terms of available system resource usage when they properly debloat their operating system. By default, Windows is bundled with a lot of bloatware (random applications), random services, telemetry, among other performance-heavy unnecessary software.
Properly debloating Windows can free up a substantial amount of disk space, mitigate random spikes in CPU usage (and overall CPU usage), as well as free up extra memory.
Shell scripts, methods, and user-made software exist to make this process easier, faster, and safer for an average end-user. Individual projects exist with instructions, but it is safer to follow a walkthrough for more inexperienced people: LowSpecGamer - How to REALLY make Windows 10 Super Light for Gaming

Keep in mind, heavily customizing Windows to minimize bloat may lead to system instability or certain features not functioning. If you want a bloat-free experience - it's recommended to switch operating systems.

(4) OPERATING SYSTEM

Many individuals have a false assumption that Windows is the crown of operating systems and that it is best suitable for gaming as well as daily use. I disagree, so do all of the experts. The cold truth that you may or may not want to hear is that GNU/Linux is simply faster, more optimized and does not cause much of an impact on your system, and delivers you more performance in games.

As mentioned previously, Windows having “bloat” (Referring to processes, software, or tasks running by software or the operating system, that are not necessary for operation, unwanted by the user, or overly impactful to the system) can massively impact performance in games or overall system responsiveness, and there is not much you can do about it on Windows. Even in high-end systems, Linux offers superior gaming performance when it comes to native ports of games, all-the-while being much more responsive and resource-friendly to lower-end systems, no matter the specifications. What it also offers is the freedom of choice, privacy, security, and knowledge about how your computer works.

In natively-running games, Linux can enable up to a 50% increase in performance, depending on the specifications at hand:

The main benchmarked example is CS:GO - A competitive first-person shooter developed by Valve, which offers good optimization and a native Linux port, which runs better on Linux than it does on Windows, with performance improvements ranging from +10% to +70%.

Benchmark sources:

Counter-Strike: Global Offensive Benchmark - Linux vs Windows

Counter-Strike Global Offensive | Pop!_OS 20.04 vs Windows 10 | GTX 1070 + R7 2700X

CS:GO Windows vs Arch Linux

Multiplatform games like those running on Java utilize system resources much more efficiently, and in turn produce much greater framerates, as OpenJDK is well known to be superior than Oracle Java JDK/JRE.

The easiest example here is Minecraft Java, a multiplatform game based around Java, utilizing OpenGL on all platforms for rendering. To say that Windows is bullied here is an understatement.

Minecraft Benchmark - Linux vs Windows

Minecraft 1.16.3 - Linux vs. Windows (FPS)

If you are concerned about system resources, privacy, security, and overall experience, trying out GNU/Linux for the first time is good practice, and then you'll decide for yourself - whether it is better for you.

Linux gaming is only a recent major stride, it is still improving, and some games are not compatible with it - those being the games that employ DRM (digital rights management tools) or anti-cheats which are designed for Windows only.

Linus Tech Tips - Linux gaming is BETTER than windows?

Linus Tech Tips - Microsoft Should be VERY Afraid - Noob's Guide to Linux Gaming

Steam natively supports GNU/Linux and incorporates SteamOS, which utilizes a compatibility layer named Proton to run Windows games on GNU/Linux distributions by emulating the Windows API and translating DirectX to Vulkan. You can check Proton support of Steam games here: ProtonDB

So far, 75% of the top 1000 games are supported on Proton with Gold+ playability - meaning that at the very least, it runs perfectly out of the box with minimal tinkering needed.

Contrary to propaganda and popular belief, GNU/Linux does not require years of computer experience or for you to study terminal commands or use one. GNU/Linux's interfaces like KDE Plasma and Gnome deliver you a desktop environment that is more optimized, clean, customizable, and easier to use than that what Windows 10 provides.

Great starting point distribution, with out-of-the-box support for Steam, Proton, and removes the need for the end-user to even touch the command-line/terminal or perform anything of complexity - featuring guides, massive community support: Pop!_OS by System76 - an easy to install, easy to use, elegant and gaming-ready distribution.

(5) COOLING, SYSTEM MAINTENANCE

Computer cooling can be seen as an art form, ranging from basic OEM cooling solutions to stock coolers, aftermarket coolers, to full-fledged water cooling quad-radiator beasts. Anyone with a computer should already understand the importance of temperature when it comes to hardware.

The importance of system cooling stems from the fact that semiconductors and most complex hardware such as the CPU and GPU can be potentially damaged by high temperatures, or have their lifespans drastically reduced due to prolonged system thermal stress. Modern processors and graphics cards employ two systems that both have to do with the response to temperatures:

1. Boosting (Also known as: “Turbo boost”, “Turbo”, “Precision Boost”, “Thermal Velocity Boost”).

Boosting is a common system in place in all modern CPUs and GPUs, which allows them to have on-demand increases in clock speeds (Amount of cycles) above the norm or “base clock”, meaning that whenever there is a need (high CPU usage) and the CPU's temperatures are under control and there is sufficient wattage to the CPU, it raises the clock speed above the base clock - thus yielding higher performance. As stated, all modern processors have this feature, and in specification sheets, the main values which are of concern are the “Base clock” and “Boost clock”. The CPU will not boost automatically above the boost clock, even if temperatures or voltages allow, anything outside of the boost clock is usually referred to as “outside of processor specification”.

2. Thermal throttling (Also known as just “throttling”).

Thermal throttling (separate from current or voltage throttling) is a feature whose purpose it is to adapt component (CPU, GPU) performance and at the same time the component's TDP (Thermal design power, or in a really basic form - the amount of heat outputted by a component) to that of what the component's cooling system can sustain. Simply put - if your component is being cooled inadequately and temperatures get up to 80-95 degrees (variable by component, vendor, model) - the component will automatically slow itself down to reduce the total heat produced - thus preventing the component from going over the maximum rated temperatures, which can damage the component irreversibly.

IMPORTANT: While components become more efficient each generation at providing more cycles (performance) for less power consumption, it is important to understand that a component requires more voltage for more cycles to be sustained. If you increase the number of cycles greatly and the voltage is not high enough to sustain that amount, the CPU will become unstable and can randomly shutdown, that is why during overclocking it is also important to fine-tune the voltage to ensure system stability. Voltage correlates to the heat produced by the component, also known as the component's TDP. While more cycles require more voltage, if the processor is running at a lower power state, the voltage can drop (along with the temperatures). This is how thermal throttling reduces temperatures, simply put - slowing down a component can allow you to decrease the voltage on it, which in turn leads to lower heat production.

In undercooled components, what can often happen is that the component simply does not have the headroom to boost itself to the maximum speed, or the thermal solution is so inadequate that the component has to throttle during load, which can heavily tank performance.

GOOD TO KNOW:

1. Mobile PCs (laptops).

Speaking from history and logic, laptops tend to have more temperature-related problems than desktop computers. This is a laptop's Achilles heel - even high-performance laptops take pride in attempting to be as slim as possible, but, the downside to this is that with a more slim design and a bigger battery, there is usually less space for a laptop's cooling components, which in standard cases are heat pipes leading from the CPU/GPU ‘die' (referring to the actual block of silicon on which the processor's circuit is fabricated on, and the part where the heat is generated) to usually the sides or back of laptops, where they connect to fins (metal spread out to maximize contact with flowing air and conduction of heat from the heat pipes), in almost all cases this is accommodated by a fan of sorts taking cooler air from the bottom of the laptop through breather holes and blowing it through the fins - the fast-moving air conducts heat from the fins, thus carrying the heat generated by the component out of the system (this is referred to as air cooling).

Due to the limited size of the heat pipes, fan, and heatsink/fins, laptops have more limited cooling efficiency than desktop computers. Due to this, laptops have their own processor series (mobile processors) - which feature reduced performance or voltage to compensate for weaker cooling and power delivery. Even in optimized laptops, some designs undermine the cooling or choose hardware with simply higher TDP than the cooling system is rated for, this can sometimes lead to thermal throttling after prolonged use or intense use. 

2. Desktop PCs.

Desktop PCs have the advantage of size when it comes to cooling, the size allows desktops to house big air coolers, tons of airflow potential, water cooling components, etc.

Desktop PC cooling factors are:

* The case and its modularity: allows one to have one face of the case for taking in cool air, moving it through the case and the internal cooling components, then another face for pulling the now-heated air out of the case.

* The individual component cooling systems: both the CPU and graphics card have their own heatsink and fan on them, the CPU often comes with an in-box cooler (free) for non-overclocking and basic usage purposes, which can be swapped out for an aftermarket cooler. While the graphics card often has a beefier and long term cooler attached to it, that is specifically tailored to the graphics card's shape and components and is not meant to be replaced.

GOOD TO KNOW: 

It's good practice to understand the basic terminology between different types of graphics processors.

* Dedicated graphics cards (dGPU) are defined as having their own dedicated memory, as opposed to utilizing system memory. It is a common practice nowadays to refer to any PCI-E (PCI-Express is a type of ‘bus' or ‘connector' used to interface high-speed components such as graphics processors to the CPU) graphics card as a dedicated graphics card, usually because all modern PCI-E GPUs are dedicated in nature, as they have onboard video memory, as opposed to utilizing the system memory.

* Integrated graphics processors (iGPU) are defined as not having onboard video memory, and utilizing solely system memory as video memory. They are commonly found integrated into the CPU dies (Intel HD graphics, AMD APU series), or present on the motherboard and linked via the north chipset bridge (Northbridge refers to a part of the chipset that interfaces and connects the CPU, memory, and high-speed interfaces such as PCI-E).

* External graphics processor (eGPU) refers to a graphics processor not housed inside of a system. This concept usually only applies to laptop computers, as most of the time - laptops do not provide a way to switch out or upgrade the graphics processor that originally shipped with the product (With a notable exception being “MXM laptops” - which provide a connector that lets you fully upgrade a laptop's graphics card). Laptops of all ranges often have high-speed interfaces that can be potentially used to connect a dedicated graphics card, most laptops utilize a PCI-E 1x slot as a way to connect a modem or wireless modules to them. Higher-end laptops can have external ‘Thunderbolt' connectors - which offer high speeds of 10-40 Gb/s depending on the exact version of the connector.

With Thunderbolt 3, it is entirely possible and easy to have a full desktop dedicated graphics card connected to your laptop PC. Companies such as Razer offer ‘external GPU docks' - which allow you to adapt Thunderbolt 3 USB-C to PCI-E 16x.

Both active (cooling systems including fans mounted to heatsinks to dissipate heat more efficiently) and passive cooling systems (cooling systems that have no moving parts) collect dust and dirt over time. Dust on heatsinks and fans can directly impact the cooling performance. Dust is oftentimes pulled into the PC from the air or surrounding surfaces by the fans and can settle on the blades of fans and the heatsink fins. While in smaller amounts, this does not cause any problems with temperatures, years of no maintenance is guaranteed to make your system all gunked up.

Intel stock cooler ("Intel stock thermal solution") after years of active use. - showing dust buildup along with the fan and top of the heatsink, which prevents airflow from going through most of the heatsink.

The issue with dust buildup preventing proper heat conduction or airflow is a problem regardless of whether you are on a laptop or desktop.

The obvious solution is to clean your system every once in a while. The interval of a thorough cleaning session varies between sources, but it is recommended to do it at least once every two years, oftentimes gamers and other individuals do it once a year.

IMPORTANT: Cleaning a PC is straightforward, but you have to know the “where?” and “how?”. 

In laptop computers, the most straightforward way to clean does not always require you to open them up. The main area to focus on is the heatsink and its fins, which are located in front of the blower-style fan. Air heat conduction occurs in that area, and due to constant airflow, dust tends to get stuck in between the fins, forming a layer of dust or entire ‘dust bunnies'. The best way to get rid of them, and the fastest, is to use compressed air - which can be bought for just a few dollars. Blowing compressed air into the fins from the outside means that most dust stuck will be decoupled and set loose in the direction of the fan. I recommend doing this when the computer is running (and preferably on a higher load, so the fan(s) are spinning at higher speeds), so the fan immediately expels the dust when it is decoupled. Watch your face, you do not want to breathe in most of what's there. Turn the laptop on, blow into the fins from the outside, wait a few seconds, repeat a few times, and you are done. Of course, to those without compressed air or those who want a more thorough cleaning process, it is then recommended to open the back of the laptop up (If unsure how to or worried about breaking something, look up the model of your laptop online - there is almost always a video on how to disassemble every laptop). Once you have the back exposed, you should have a clearer view of the fan intake and exhaust, allowing you to clean those with various tools, it doesn't really matter what you use - as long as you can scoop the dust out, you can even blow into the heatsink or fan manually. Do not attempt to remove the cooler assembly (heat pipes, fans) UNLESS you have thermal paste available to you. Thermal paste is a sticky and toothpaste-like substance that is placed on the component's die or IHS (integrated heat spreader, commonly found only on desktop computers) - the function of thermal paste is to seal the imperfections and gaps between the contact points of the component and the heat pipes. Removing/separating the two ruins the pre-applied thermal paste, meaning that if no new one is applied and the cooling system is placed back in - there are going to be serious temperature issues due to lack of proper contact between your components (CPU, GPU) and the heat pipes. If you have thermal paste available to you, you can then carefully remove the heat pipes and fans from the laptop. Removing the cooling assembly carefully is a must - it is possible to chip off part of the CPU or GPU die if you are too rough. Having the thermal assembly removed, you can proceed to complete maintenance:

1) Use compressed air or other tools to ensure that the area in between the fins is clear from dust, if you can separate the fins from the fan, you can clean it by washing it with higher-pressure water and letting it dry properly.

2) Use compressed air or a cloth (preferably soaked in isopropyl alcohol) to clean the blades of the blower-style fan, or even the walls if possible. While it is possible to wash the fan, as well, it is a risky maneuver if you do not know the exact techniques for safely drying electrical components, so it is not recommended to do when you are inexperienced.

3) Clean off the CPU's (and GPU's, if present) dies from old thermal paste, as well as the thermal paste left on the bottom-facing part of the heat pipes. It is recommended to do this with a cloth or q-tip soaked in isopropyl alcohol (you should be smart enough to avoid water near sensitive electronics when possible…). Once the isopropyl alcohol has dried and you are done with cleaning other parts of the thermal solution, you then apply a small drop of thermal paste. For a first-timer, it is always recommended to watch a tutorial on how to do it, as too much and too little can cause inconveniences: FSixTech - BEST METHOD to Replace your Laptop's Thermal Paste!

After it is done, you can remount your thermal solution, minding proper screw tightness, plugging fans in, etc.

In desktop computers, cleaning one out requires you to open up the side panel. Once the side panel is removed, you are greeted with every component in your system, for starters, you can see your CPU heatsink (and probably fan), one or more case fans, a dedicated graphics card if you have one installer. Cleaning all of them out does not require you to remove the individual coolers, and compressed air is the best tool that can be used here. The two most sensitive components that are to be cleaned are the CPU heatsink and the GPU heatsink. The process is easy with a bottle of compressed air, as you can just blow between the fins of both of them and it decouples the dust and dirt. Additionally, you should also clean out your whole case from dust, as it can settle on the motherboard, power supply, or other components - but it is up to you whether you want an entirely clean system. Since the placement of the graphics card is often with the cooler facing downwards, it is often recommended to temporarily remove it to clean it. Do not attempt to take the CPU and GPU heatsinks off without having thermal paste ready, the why is explained in the laptop section.

(6) OVERCLOCKING, UNDERVOLTING

The section is in a pre-release state, the information provided here may have not been fact-checked fully yet, or is incomplete / not ready for proper usage.

This section is still missing on individual overclocking guides, stability and hardware lifespan warnings.

Overclocking and undervolting are a set of changes made to the way that your hardware operates. A good overclock can drastically improve your components' performance, while a good undervolt can reduce the amount of heat that your components generate.

Clock speed, also formally known as the ‘clock rate' describes the frequency (measured in Hz) at which the component's synchronization pulses fire. Synchronization pulses in components like the CPU, GPU, and RAM serve the same purpose as in any other circuitry - it coordinates the speed at which operations in the circuit fire, or in more primitive terms - how fast the component is at processing or completing a specific task. Increasing the clock speed of a component like the CPU, GPU or RAM then theoretically increases the performance or ‘speed' of the component.
The faster a component operates, however, the more power is required to keep it stable. In most modern components this is achieved through modifications in voltage. When it comes to overclocking, usually with a substantial overclock - an overvolt is needed to keep it stable, while in cases of reducing power draw (or temperatures), voltage is usually lowered.

In a lot of OEM machines (such as laptops or prebuilt computers with non-standard motherboards), changing the components' clock speeds or voltages is often not an option. Most laptop chipsets do not give you an option to change the CPU multiplier or anything else.

Companies like Intel purposely disable the option to change the clock speed or voltage on most of their processors, in order to market and sell more ‘K-series' (“unlocked”) and ‘X-series' (“extreme edition”) processors - which are more expensive. Meanwhile, AMD supports tuning the CPU clock speed and voltage even on their entry-level processors. It is also of importance to note that not all motherboards support overclocking - in the case of AMD AM4 motherboards, B and X series support overclocking, while the A series does not. On the Intel side of things, only Z and X series chipsets support these aforementioned features. That being said, if you own a laptop, overclocking is almost always not an option. This does not however mean that a lot of desktops are left out - if you have an entry-level desktop from even 6 years ago - it does not mean that overclocking is not supported, so it is important to look up your processor and your motherboard to be able to make the decision.

IMPORTANT: Overclocking (and subsequently overvolting) can result in system damage. The clock speeds of a component are limited for a reason, going above them may not be safe for the component, its power delivery, or cooling system. Increasing the voltage on a component produces more heat - if the heat is not controlled and the temperatures are too great, your component may be damaged from high temperatures or have its lifespan reduced. Applying unreasonable clock speeds or voltages to the CPU from the BIOS can prevent your computer from POSTing (essentially making your computer unbootable) - if this happens, it is necessary to reset the BIOS settings by either jumping the CLR_CMOS headers or reseating the CMOS battery on your motherboard - if you do run into an issue where you cannot boot due to incorrect or faulty BIOS settings - make sure to remember to always clear CMOS to be able to boot back into your computer again. Generally - there is not much risk in overclocking as long as you follow standard procedures, but due to the nature of overclocking (or modifying voltage) - you cannot ever be guaranteed to have successful results.

It is natural for overclocking to increase the performance substantially. Modern processors and components are mainly designed for overclocking when it comes to the high end. While it is not recommended by most manufacturers - tinkerers among other individuals still prefer to do it.
More modern examples (as for CPUs), such as the Ryzen 5 2600 or the Ryzen 3 3100 allow you to overclock by 100-500 MHz just using the stock heatsink, but of course - those are not exactly low-end processors, they are well within the bounds of mid-range. This is why it is important to research your processor and chipset before jumping into anything. If it is overclockable, it is more than likely that someone has done it and released a video on it, so it is highly encouraged to see what people have done with the processor. The same story goes for graphics cards, although, most dedicated graphics cards (even low-end ones) allow you to overclock or at the very least allow you to set it somewhere within the maximum boost clock - same as for CPUs, always look up guides or tutorials on overclocking your components, they usually include the information on how to perform the overclock and the kind of performance improvement you can hope to achieve.

Undervolting, however - only brushes by the concept of overclocking. Unlike overclocking, undervolting is the process of reducing the voltage applied to the component, but usually not decreasing the clock speed. In the manufacturing process, components have their voltages set higher than what the thin line of system stability is, just to compensate for certain component imperfections or usage fluctuations. Reducing this gap between system stability and the preset voltage allows for a component to perform cooler - with the goal of either reducing temperatures of power draw. In certain cases, if the component is under high temperatures, the temperature reduction from undervolting may help the component perform better.

To perform either overclocking or undervolting on a processor, there is no unified documented process of doing it. Factors such as brand, model of processor, chipset all apply. It is important to research your components to be able to understand what your components can achieve, what performance improvement to expect, and the potential risks.

You can make open suggestions on the Google Documents version of this. Feel free to add more content, correct my mistakes, improve explanations, or anything else to improve the document. Just make sure to leave your name somewhere as for the credits.

Texas#1111 - 771007665257709599

Documented changes

Release 1.0.0 - 2021-01-30

The initial release of the document, with hardware and software bottlenecks already well described, but missing a lot of entries and information on game/OS/driver tinkering, as well as no information provided about overclocking or undervolting.

Release 1.0.5 - 2021-01-31
Small fixes in text readability, added more to the system tinkering section, laid out the groundwork for overclocking and undervolting.

Release 1.1.0 - 2021-02-01
Small fixes in text readability, expanded headers on system tinkering, bottlenecks. Expanded on overclocking and undervolting to warn about risks and other important information.

--- Reply reserved for upcoming content or other form of important information (2)

--- Reply reserved for upcoming content or other form of important information (3)

--- Reply reserved for upcoming content or other form of important information (4)

--- Reply reserved for upcoming content or other form of important information (5)