What is overclocking? Overclocking is a method such that a wide variety of computer components is higher than the rated speed. For example, if you purchased a Pentium 4 3.2GHz processor, and you want it to run faster, you can overclock the processor to allow it to run at 3.6GHz. Why do you want to overclock? Yes, the most obvious motivation is to get more rewards from the processor. You can purchase a relatively inexpensive processor and overclock it overclock over the speed of the more than much processor. If you are willing to invest time and efforts, overclocking can save a lot of money; if you are a fanatic player like me, overclocking can bring you more fast than the faster processor that may be purchased from the store. Overclocking danger first I want to say, if you are very careful and know what to do, that is very difficult to cause any permanent damage to your computer through overclocking. If the system is too much, it will burn the computer or cannot start. But only it is difficult to burn the system only. However, there is still dangerous. The first is also the most common danger is fever. It will produce more heat when the computer component is higher than the rated parameter. If there is no sufficient heat, the system is likely to overheat. However, the general overheating cannot destroy the computer. The only thing that scrapped by overheating is the only thing that is repeatedly trying to make the computer run higher than the recommended temperature. As I said, you should try to suppress 60 c or less. However, there is no need to excessively worry about overheating problems. There will be a sign before the system crashes. Random restart is the most common signage. Overheating is also easy to prevent the use of heat sensors, it can display the temperature of the system. If you see the temperature too high, run the system at a lower speed or use better heat dissipation. I will discuss heat dissipation in this guide later. Another "danger" of overclocking is that it may reduce the life of the part. It will be reduced when a higher voltage is applied to the component. Small increase will not cause too much impact, but if you plan to make a lot of overclocking, you should pay attention to your life. However, this is usually not a problem, because any overclocking people are unlikely to use the same part to four or five years, and it is impossible to say that any part is not supported 4-5 years. Most processors are designed to use for 10 years, so in the brain of overclockfilters, the loss of some years to exchange performance is usually worthy of basic knowledge, in order to understand how to overclock, first, I must know how the system works. . The most common components used to overclock rate are the processor. When you purchase a processor or CPU, you will see its run speed. For example, the Pentium 4 3.2GHz CPU is running under 3200 MHz. This is a measure of how many clock cycles for the processor in one second. A clock cycle is a period of time, during which the processor can perform a given quantity instruction. So logically, the more clock cycles that can be completed in one second, the more you can process information, and the system will run faster. 1 MHz is one million clock cycles per second, so 3.2GHz processor can experience 3,200,000,000 or 3 billion clock cycles within every second. Quite can't afford, right? The purpose of overclocking is to increase the GHz level of the processor so that it can experience more clock cycles per second. The formula for calculating the processor is this: FSB (in MHz) × multiplier = speed (in MHz). Now explain what FSB and multiplier are: FSB (HTT * for the AMD processor is htt *), or the front-end bus is the channel of the entire system and the CPU communication. Therefore, the faster FSB can run, and it is clear that the entire system can run faster. CPU vendors have found a way to increase the FSB effective speed of the CPU. They just sent more instructions in each clock cycle.
Therefore, the CPU vendor has a way to send two instructions for each clock cycle (AMD CPU), or even the Intel CPU, instead of each clock cycle to send an instruction. Then consider the CPU and see the FSB speed, it must be recognized that it is not really running at that speed. Intel CPU is "four-core", that is, they send 4 instructions each clock cycle. This means that if you see 800MHz FSB, the potential FSB speed is actually only 200MHz, but it sends 4 instructions per clock cycle, so it reaches an effective speed of 800MHz. The same logic is also suitable for AMD CPUs, but they are just "two cores" means that they only send 2 instructions per clock cycle. So the FSB of 400MHz on the AMD CPU is composed of 2 instructions from the potential 200MHz FSB. This is important because the CPU real FSB speed will be processed during overclocking, not a valid CPU speed. The multiplier portion of the speed equation is also a number, and the total speed of the processor is given by the FSB speed. For example, if there is a CPU having 200MHz FSB (real FSB speed before multiplying two or four) and 10 times, then the equation becomes: (FSB) 200MHz × (multiplier) 10 = 2000 MHz CPU speed, or It is 2.0GHz. On some CPUs, such as Intel since the processor since 1998, multiplier is locking cannot be changed. In some, such as the AMD Athlon 64 processor, the multiplier is "capping lock", that is, the multiplication is changed to a lower number, but it cannot be improved to higher than the initial. On other CPUs, the multiplication is completely released, which means it can change it to any desired number. This type of CPU is an overclocking best because it can easily overclock in the CPU by increasing the multiplier, but now very rare. Increase or decrease the multiplication over the CPU is much easier than the FSB. This is because the multiplier and FSB are different, it only affects the CPU speed. When changing the FSB, it is actually changing the speed of each individual computer part and the CPU communication. This is all other components in the overclocking system. This may bring a wide variety of problems when the components that are not intended overclocking are exceeded. However, once you understand how overclocking occurs, you will know how to prevent these problems. * On AMD Athlon 64 CPU, the term FSB is really useful. There is no FSB in nature. FSB is integrated into the chip. This makes the FSB and CPU communication than Intel's standard FSB method is much faster. It may also cause some confusion because FSB on Athlon 64 may sometimes be said to be HTT. If some people are talking about increasing the HTT on the Athlon 64 CPU, and is discussing the speed of the regular FSB speed, the HTT is considered as FSB. To a large extent, they operate in the same manner and can be considered as the same thing, and HTT as an FSB is considered to eliminate some possible confusion. How to overclock so now I now understand how the processor reaches its rated speed. Very good, but how to improve this speed? Overclocking is the most common method to pass BIOS. You can enter the BIOS when you press the specific key when the system is started. The most common keys used to enter the BIOS are the delete key, but some may use the f1, f2, other f buttons, Enter, and other keys. Before the system starts to load Windows (any OS), there should be a screen to display what the button to use at the bottom. Assume that the BIOS supports overclocking *, once it goes to the BIOS, you should be able to use all settings you need to use the overclocking system.
The most likely to be adjusted is: multiplier, FSB, RAM delay, RAM speed, and RAM ratio. At the most basic level, your only thing to try is to get the highest FSB × multiplier formula you can reach. Completing this easiest way is to increase multiplier, but that is not implemented on most processors, because the multiplication is locked. Second method is to increase FSB. This is quite limited, all RAM issues that must be processed when improving the FSB will be described below. Once the speed limit of the CPU is found, there is not only one choice. If you really want to push the system to the limit, you can reduce the multiplication in order to raise the FSB. To understand this, imagine a processor with a 2.0GHz, which uses 200MHz FSB and 10 times. So 200 MHz × 10 = 2.0GHz. Obviously this equation works, but there are other methods to get 2.0GHz. You can increase the multiplier to 20 and reduce the FSB to 100 MHz, or you can reduce the FSB to 250 MHz and reduce the frequency to 8. Both combinations will provide the same 2.0GHz. So, is it two combinations that should provide the same system performance? no. Because FSB is a channel used to communicate with the processor, it should make it high as possible. So if the FSB is reduced to 20 MHz, it will still have a clock speed of 2.0GHz, but the rest of the system and processor communication will be much more slower than before, resulting in loss of system performance. Ideally, in order to increase the FSB as high as possible, it should be reduced. In principle, this sounds very simple, but it will become complex when including the rest of the system, because the other parts of the system are also determined by the FSB, the first thing is RAM. This is also what I have to discuss in the next section. * Most retail computer vendors use motherboards and BIOs that do not support overclocking. You will not be able to access the required settings from the BIOS. Tools allow for overclocking from the Windows system, but I don't recommend them, because I never tested RAM and its impact on overclocking, as I said before, FSB is the path to the system and CPU communication. So the increase in FSB also effectively overclocking the rest of the system. Parts that are most affected by the increase in FSB is RAM. When purchasing RAM, it is set at a speed. I will use the form to display these speeds: PC-2100 - DDR266 PC-2700 - DDR333 PC-3200 - DDR400 PC-3500 - DDR434 PC-3700 - DDR464 PC-4000 - DDR500 PC-4200 - DDR525 PC-4400 - DDR550 PC -4800 - DDR600 To understand this, you must first understand how RAM works. RAM (Random Access Memory, Random Access Memory) is used as a temporary storage for files that require quick access. For example, when you load the plane in the game, the CPU loads the plane into the RAM so that it quickly access information when any need, not from the relatively slow hard drive. It is important to know that RAM is run at a speed, which is much lower than the CPU speed. Today, most RAM runs at a speed between 133 MHz to 300MHz. This may be confused because those speeds are not listed on my chart. This is because RAM vendors imitate the practice of CPU vendors, trying to let RAM send twice as information in each RAM clock cycle *. This is the origin of DDR in the RAM speed level. It represents Double Data Rate (twice the data speed). Therefore, DDR 400 means that RAM operates at a validity of 400 MHz, and 400 in DDR 400 represents the clock speed. Because each clock cycle is sent twice, it means that its real operating frequency is 200MHz. This is like AMD's "two core" FSB.
So back to RAM. There is a speed listing the DDR PC-4000. The PC-4000 is equivalent to DDR 500, which means that the PC-4000 RAM has a 500MHz effective speed and a potential 250 MHz clock speed. So what is overclocked? As I said before, when I improve the FSB, I will effectively override all other things in the system. This also includes RAM. Rated RAM in PC-3200 (DDR 400) is running at a speed of up to 200MHz. This is enough for people who are not overclocking, because FSB will not exceed 200MHz anyway. However, when you want to raise the FSB to more than 200MHz, the problem will appear. Because RAM is rated at a speed of up to 200MHz, it may cause the system to crash at a speed of up to 200MHz. How do I solve this? There are three solutions: using FSB: RAM ratio, overclocking RAM, or RAM is set at a higher speed. Because you may only know the last one in the three choices, I will explain them in the future: fsb: RAM ratio: If you want to increase FSB to a higher speed support than RAM, you can choose to let RAM run in fsb Lower speed. This is done using the FSB: RAM ratio. Basically, the FSB: RAM ratio allows you to select a ratio between FSB and RAM speeds. Suppose you are using PC-3200 (DDR 400) RAM, I mentioned it before it runs at 200MHz. But you want to increase FSB to 250MHz to overclock CPU. Obviously, RAM will not support elevated FSB speeds and it is likely to cause system crash. In order to solve this, the 5: 4 FSB: RAM ratio can be set. Basically this ratio means that if the FSB is running in 5MHz, then RAM will run only under 4MHz. More simply, the ratio of 5: 4 is changed to 100: 80 ratio. Then the FSB is running under 100MHz, RAM will run only under 80MHz. Basically this means that the RAM will run only 80% of the FSB speed. So as for 250 MHz target FSB, running in 5: 4 FSB: RAM ratio, RAM will run under 200MHz, which is 80% of 250MHz. This is perfect because RAM is rating 200MHz. However, this solution is not ideal. Run FSB and RAM with a ratio results in time difference between FSB and RAM communication. This causes deceleration, and if the RAM is running at the same speed, it will not appear. If you want to get the maximum speed of the system, use the FSB: RAM ratio is not the best solution. Overclocking RAM overclocking RAM is really simple. The principle of overclocking RAM is the same as overclocking CPU: allowing RAM to run at a higher speed than it is set. Fortunately, there are many similar things between the two overclocking, otherwise the RAM overclock is much more complicated than the imagination. To overclocking RAM, you only need to enter the BIOS and try to make RAM run at a rate ratio. For example, it is possible to try to make the RAM of the PC-3200 (DDR 400) run at a speed of 210 MHz, which exceeds the rated speed of 10 MHz. This may be fine, but in some cases, the system will cause the system to crash. If this happens, don't panic. By increasing the RAM voltage, the problem can be quite easily solved. The RAM voltage is also referred to as VDIMM, which is adjustable in most BIOS. Use the smallest available increment to improve it and test each setting to observe if it is running. Once a running setting is found, you can always keep it, or try further RAM.
