Practical CPU overclocking

Processor overclocking methods

There are two methods of overclocking: increasing the frequency of the system bus (FSB) and increasing the multiplication factor (multiplier). At the moment, the second method cannot be applied to almost all serial AMD processors. Exceptions to the rule are: Athlon XP processors (Thoroughbred, Barton, Thorton )/Duron (Applebred), released before week 39 of 2003, Athlon MP, Sempron (socket754; downgrade only), Athlon 64 (downgrade only), Athlon 64 FX53/55. In serial processors manufactured by Intel, the processor overclocking is also completely blocked. by increasing the multiplier is the most “painless” and simplest, because only the processor clock frequency increases, and the frequencies of the memory bus and AGP/PCI buses remain nominal, so determine the maximum processor clock frequency at which it can operate correctly using this The method is especially simple. It’s a pity that now it’s quite difficult, if not impossible, to find AthlonXP processors on sale with an unlocked multiplier. Overclocking a processor by increasing the FSB has its own characteristics. For example, as the FSB frequency increases, the memory bus frequency and the AGP/PCI bus frequency also increase. Particular attention should be paid to the PCI/AGP bus frequencies, which in most chipsets are associated with the FSB frequency (does not apply to nForce2, nForce3 250). This dependency can only be circumvented if your motherboard's BIOS has the appropriate parameters - the so-called dividers responsible for the ratio of PCI/AGP to FSB. You can calculate the divider you need using the formula FSB/33, i.e., if the FSB frequency = 133 MHz, then you should divide 133 by 33, and you will get the divider you need - in this case it is 4. The nominal frequency for the PCI bus is 33 MHz, and the maximum is 38-40 MHz; setting it higher, to put it mildly, is not recommended: this can lead to the failure of PCI devices. By default, the memory bus frequency rises synchronously with the FSB frequency, so if the memory does not have sufficient overclocking potential, it can play a limiting role. If it is obvious that the RAM frequency has reached its limit, you can do the following:

• Increase memory timings (for example, change 2.5-3-3-5 to 2.5-4-4-7 - this can help you squeeze a few more MHz out of RAM).
• Increase the voltage on the memory modules.
• Overclock the processor and memory asynchronously.

Reading is the mother of learning

First, you will need to study the instructions for your motherboard: find the BIOS menu sections responsible for the FSB frequency, RAM, memory timings, multiplier, voltages, PCI/AGP frequency dividers. If the BIOS does not have any of the above parameters, then overclocking can be done using jumpers on the motherboard. You can find the purpose of each jumper in the same instructions, but usually information about the function of each is already printed on the board itself. It happens that the manufacturer himself deliberately hides “advanced” BIOS settings - to unlock them you need to press a certain key combination (this is often found on motherboards manufactured by Gigabyte). I repeat: all the necessary information can be found in the instructions or on the official website of the motherboard manufacturer.

Practice

We go into the BIOS (usually to enter you need to press the Del key at the moment of recalculating the amount of RAM (i.e., when the first data appears on the screen after rebooting/turning on the computer, press the Del key), but there are motherboard models with a different key for entering the BIOS - for example, F2), look for a menu in which you can change the frequency of the system bus, memory bus and control timings (usually these parameters are located in one place). I think that overclocking the processor by increasing the multiplier will not cause any difficulties, so let’s move on straight to raising the system bus frequency. We raise the FSB frequency (by about 5-10% of the nominal), then save the changes made, reboot and wait. If everything is fine, the system starts with a new FSB value and, as a result, with a higher processor clock speed (and memory, if you overclock them synchronously). Booting Windows without any incidents means that half the battle is already done. Next, run the CPU-Z program (at the time of writing, its latest version was 1.24) or Everest and make sure that the processor clock frequency has increased. Now we need to check the processor for stability - I think everyone has a 3DMark 2001/2003 distribution kit on their hard drive - although they are designed to determine the speed of the video card, you can also “drive” them for a superficial check of system stability. For a more serious test, you need to use Prime95, CPU Burn-in 1.01, S&M (more details about test programs below). If the system has passed testing and behaves stably, we reboot and start all over again: go into the BIOS again, increase the FSB frequency, save the changes and test the system again. If during testing you were “kicked out” of the program, the system froze or rebooted, you should “roll back” a step - to the processor frequency when the system behaved stably - and conduct more extensive testing to make sure that operation is completely stable. Do not forget to monitor the processor temperature and PCI/AGP bus frequencies (in the OS, PCI frequency and temperature can be viewed using the Everest program or proprietary programs of the motherboard manufacturer).

Voltage increase

It is not recommended to increase the voltage on the processor by more than 15-20%, but it is better that it varies within 5-15%. There is a point to this: stability increases and new horizons for overclocking open up. But be careful: as the voltage increases, the power consumption and heat dissipation of the processor increases and, as a result, the load on the power supply increases and the temperature rises. Most motherboards allow you to set the RAM voltage to 2.8-3.0 V, the safe limit is 2.9 V (to further increase the voltage you need to voltmod the motherboard). The main thing when increasing the voltage (not only on RAM) is to control the heat generation, and, if it has increased, organize cooling of the overclocked component. One of the best ways to determine the temperature of any computer component is to touch it with your hand. If you cannot touch a component without pain from a burn, it requires urgent cooling! If the component is hot, but you can hold your hand, then cooling it would not hurt. And only if you feel that the component is barely warm or even cold, then everything is fine and it does not need cooling.

Timings and frequency dividers

Timings are delays between individual operations performed by the controller when accessing memory. There are six of them in total: RAS-to-CAS Delay (RCD), CAS Latency (CL), RAS Precharge (RP), Precharge Delay or Active Precharge Delay (usually referred to as Tras), SDRAM Idle Timer or SDRAM Idle Cycle Limit, Burst Length . Describing the meaning of each is pointless and useless to anyone. It is better to immediately find out what is better: small timings or high frequency. There is an opinion that timings are more important for Intel processors, while frequencies are more important for AMD. But do not forget that for AMD processors, the memory frequency achieved in synchronous mode is most often important. Different processors have different memory frequencies as their “native” frequencies. For Intel processors, the following frequency combinations are considered “friends”: 100:133, 133:166, 200:200. For AMD on nForce chipsets, synchronous operation of the FSB and RAM is better, while asynchrony has little effect on the AMD + VIA combination. On systems with an AMD processor, the memory frequency is set in the following percentages with FSB: 50%, 60%, 66%, 75%, 80%, 83%, 100%, 120%, 125%, 133%, 150%, 166% , 200% are the same divisors, but presented a little differently. And on systems with an Intel processor, the dividers look more familiar: 1:1, 4:3, 5:4, etc.

Black screen

Yes, this also happens :) - for example, when overclocking: you simply set the clock speed of the processor or RAM (perhaps you specified too low memory timings) that the computer cannot start - or rather, it starts, but the screen remains black, and the system does not show any “signs of life”. What to do in this case?

• Many manufacturers build into their motherboards a system for automatically resetting parameters to nominal. And after such an “incident” with an inflated frequency or low timings, this system should do its “dirty” work, but this does not always happen, so you need to be ready to work manually.
• After turning on the computer, press and hold the Ins key, after which it should start successfully, and you should go into the BIOS and set the computer’s operating parameters.
• If the second method does not help you, you need to turn off the computer, open the case, find on the motherboard the jumper responsible for resetting the BIOS settings - the so-called CMOS (usually located near the BIOS chip) - and set it to Clear CMOS mode for 2-3 seconds, and then return to the nominal position.
• There are motherboard models without a BIOS reset jumper (the manufacturer relies on its automatic BIOS reset system) - then you need to remove the battery for a while, which depends on the manufacturer and model of the motherboard (I conducted this experiment on my Epox EP-8RDA3G: took out the battery, waited 5 minutes, and the BIOS settings were reset).

Information programs and utilities

CPU-Z is one of the best programs that provides basic information about the processor, motherboard and RAM installed in your computer. The program interface is simple and intuitive: there is nothing superfluous, and all the most important things are in plain sight. The program supports the latest innovations from the world of hardware and is updated periodically. The latest version at the time of writing is 1.24. Size - 260 Kb. You can download the program at cpuid.com.

Everest Home/Professional Edition (formerly AIDA32) is an information and diagnostic utility that has more advanced functions for viewing information about installed hardware, operating system, DirectX, etc. The differences between the home and professional versions are as follows: the Pro version does not have a RAM testing module (read/write), it also lacks a rather interesting Overclock subsection, which collects basic information about the processor, motherboard, RAM, processor temperature, motherboard board and hard drive, as well as overclocking your processor as a percentage :). The Home version does not have software accounting, advanced reports, interaction with databases, remote control, or enterprise-level functions. In general, these are all the differences. I myself use the Home version of the utility, because... I don't need the additional features of the Pro version. I almost forgot to mention that Everest allows you to view the PCI bus frequency - to do this, you need to expand the Motherboard section, click on the subsection with the same name and find the Chipset Bus Properties/Real Frequency item. The latest version at the time of writing is 1.51. The Home version is free and weighs 3 Mb, the Pro version is paid and takes 3.1 Mb. You can download the utility at lavalys.com.

Stability testing

The name of the CPU Burn-in program speaks for itself: the program is designed to “warm up” the processor and check its stable operation. In the main CPU Burn-in window, you need to specify the duration, and in the options, select one of two testing modes:

• testing with error checking enabled;
• testing with error checking turned off, but with maximum “warming up” of the processor (Disable error checking, maximum heat generation).

When you enable the first option, the program will check the correctness of the processor's calculations, and the second will allow you to “warm up” the processor almost to temperatures close to the maximum. CPU Burn-in weighs about 7 Kb.

The next worthy program for testing the processor and RAM is Prime95. Its main advantage is that when an error is detected, the program does not spontaneously “hang”, but displays data about the error and the time it was detected on the working field. By opening the Options -> Torture Test… menu, you can choose from three testing modes or specify your own parameters. To more effectively detect processor and memory errors, it is best to set the third testing mode (Blend: test some of everything, lots of RAM tested). Prime95 weighs 1.01 Mb, you can download it at mersenne.org.

Relatively recently, the S&M program saw the light of day. At first it was conceived to test the stability of the processor power converter, then it was implemented to test RAM and support for Pentium 4 processors with HyperThreading technology. At the moment, the latest version of S&M 1.0.0(159) supports more than 32 (!) processors and checks the stability of the processor and RAM; in addition, S&M has a flexible system of settings. Summarizing all of the above, we can say that S&M is one of the best programs of its kind, if not the best. The program interface has been translated into Russian, so it is quite difficult to get confused in the menu. S&M 1.0.0(159) weighs 188 Kb, you can download it at testmem.nm.ru.

The above-mentioned tester programs are designed to check the processor and RAM for stability and identify errors in their operation; they are all free. Each of them loads the processor and memory almost completely, but I would like to remind you that programs used in everyday work and not intended for testing can rarely load the processor and RAM so much, so we can say that testing occurs with a certain margin.

The author does not bear any responsibility for the breakdown of any hardware of your computer, as well as for failures and glitches in the operation of any software installed on your computer.

It so happened that in almost twenty years of IT practice I have never had to deal with overclocking - somehow everyone had other interests. However, when choosing a configuration for another new (although now far from new) computer, for some reason I settled on an Intel processor with an open multiplier - i5-2500K. Why I did this, I don’t remember now, perhaps I intended to figure out in my old age what this overclocking is. And then one evening, when there was nothing to do, I realized that the moment had come, and I delved into studying the issue, and the next evening I applied what I had learned in practice. That's what I'm going to report on.

Overclocking theory

Overclocking issues have been of interest to humanity all the time from the moment computer technology came to the masses. The main driver of overclocking is the spirit of competition, passion, and the desire to achieve better results than others. Well, its main object is innocent processors, which are subjected to inhuman loads in order to obtain these same results. There are two main ways to overclock a processor. The first is to increase the frequency of the BCLK clock generator, which, through multipliers, determines the operating frequency of the processor, memory, buses and bridges. This option is, in principle, universal, but it has many nuances and limitations associated with a specific processor and motherboard, so so that your experiments do not lead to the death of the computer, you need to carefully understand everything. The second method is to change the processor multiplier, the same one by which BCLK is multiplied to obtain the operating frequency. This path is much safer (only the operating mode of the processor is changed, and not the entire system) and simpler (essentially one parameter is responsible for overclocking), but there is one thing: the multiplier must be unlocked (allowed for change) by the processor manufacturer.
Initially, Intel processors had an open multiplier, but in the 90s of the last century, after a series of scandals related to the relabeling of processors by unscrupulous suppliers, when slow processors were overclocked and sold at the price of faster ones, the company blocked the multiplier. Since then, the unlocked multiplier has only been found in top “enthusiast” models, which, naturally, were not cheap. The situation changed fundamentally with the advent of second-generation Intel Core (Sandy Bridge) processors - their line included models with an unlocked multiplier for the mass consumer, which received the K index. Initially, the cost of the K and non-K variants of one processor differed quite significantly, but now it has practically disappeared to no (for example, the difference between Core i5 3570 and Core i5 3570K today is 150 rubles).

So, Intel itself has opened the way for “at-home”, fast and highly skilled overclocking. It would be a sin not to take advantage of such an opportunity, and I began my experiments. As I already said, my long-suffering home computer was used as a test bench once again; by the way, it was completely unprepared for overclocking; rather, on the contrary, it was chosen for reasons of efficiency and noiselessness.

Experiment

According to the specifications, the i5-2500K operates at multipliers from 16 to 56. With standard parameters and using SpeedStep, we have 16x at idle and 34x under load. Now let's start the process. “Home” overclocking has become so homely that it can now be done directly from Windows, without going into the BIOS. But we will still be oldfags to begin with - only BIOS, only hardcore! However, we won’t get much hardcore – we only need one parameter there; in the BIOS of my ASUS P8Z68-V LX motherboard it is called CPU Ratio and is located in the CPU Power Management menu. To overclock the processor above standard values, you will also need to enable the Turbo Mode option (it has nothing to do with Intel Turbo Boost, which, on the contrary, is recommended to be turned off).
The first overclock was tiny, up to 36x, in order to mark my entry into the ranks of overclockers. However, there was no fanfare, and nothing at all happened except the frequency in the CPU monitor. The temperature also remained unchanged. The next level is 40x, a significant figure; until recently such a result (when overclocked on the bus) was considered a grandmaster. The height was taken without the slightest effort and without changing the voltage on the processor. But the temperature, unfortunately, crept up and reached 68 degrees at 100% load. There is nothing to be done; the cooling system installed on the computer has shown itself to be completely unsuitable for overclocking.

Step three. 44x, that is, 1 GHz increase. Having made my face like a brick, I started the computer. “Well, no, that’s enough,” he replied and flew into the blue screen. It is necessary to increase the processor supply voltage. I immediately raised it to 1.4 V so that it would be enough. Now I decided to operate through the GUI on Windows. In the AI ​​Suite software supplied with the ASUS motherboard, the Turbo V EVO component is responsible for overclocking. To operate, this program uses the TPU (TurboV Processing Unit) controller on the motherboard. The TPU module is so intelligent that it can itself, without human intervention, overclock the system to the highest possible parameters. Thus, overclocking technology, from the point of view of the “dummies”, has reached its highest point, when to get the result it is enough to press one button “make sure everything goes well”.
I was not able to really test the 4.4 GHz mode, because just a few seconds after starting a full load, the temperature rose to the maximum permissible, and I was forced to interrupt the experiment. However, I have no doubt that with normal cooling the operation of the processor would be stable - numerous experiments of other users convince me of this. If we talk specifically about the i5-2500K, absolutely everyone’s processors work up to 4.5 GHz, the result of 5 GHz is quite common, and the most stubborn ones reached 5.2 GHz. Let me emphasize that we are talking about stable operation under heavy (test or real) load. Thus, we are dealing with more than 50% increase in frequency with minimal material and mental costs.

Results and conclusions

As expected, the results of the computational tests climbed linearly as the frequency increased. For example, I chose the CPU Queen integer “chess” test. As you can see, with maximum overclocking, our processor “pushed” not only the extreme first-generation i7, but also the server Xeon (although it was initially inferior to both).

Someone might be wondering what happened to the Windows Experience Index? Almost nothing, it increased by only one tenth, from 7.5 to 7.6. However, do not forget that for Windows 7 the maximum index value is 7.9, so a big jump could not have happened.

Now let's try to answer the question, who needs this overclocking - except for overclockers themselves? However, it was answered before us: first of all, to fans of computer games. Experiments have shown that processor power at standard frequencies is not enough to power top-end video cards, especially if there are several of them, and as the frequency increases to a certain limit, gaming performance also increases. Saturation occurs, by the way, at our “home” 4-4.5 GHz; it is at this frequency that the processor ceases to be the “bottleneck” of the entire system. In addition, people dealing with heavy media content, and, of course, respected fans of distributed computing will definitely be happy with the extra gigahertz. I note that all categories of citizens will have to vigilantly monitor the temperature of the processors and their cooling system - otherwise a slight “zilch” and smoke is guaranteed.

Today we will look at such an interesting thing as overclocking. Let's understand the essence, methods of achieving it and why it is needed.

Overclocking is an English word that literally translates as “overclocking,” but when translated into Russian it means overclocking. It happens for both processors and video cards.

In this article we will look at overclocking methods for the processor.

The essence of overclocking

Overclocking the processor is carried out by increasing the core frequency and bus frequency.

There are two very important nuances:

• The core frequency must be proportional to the bus frequency;
• The bus frequency should not be higher than the core frequency.

Overclocking is usually needed by enthusiasts, gamers, or benchers - people who are obsessed with Benchmark results and want them to always be higher than others.

Overclocking a processor is a very delicate matter. Here you can't go left or right. Overclocking can definitely be compared to jewelry making.

If you do less, the computer won’t turn on; if you do more, it will burn out.

But we will talk about all the disadvantages a little later, but for now let’s talk about the advantages:

• An increase in FPS (frames per second) in games, albeit small (by 5-10 frames);
• Increased PC performance in general;
• Increased application launch speed;
• An overclocked processor is a good temporary replacement for a video card.

Why temporary? Because you can’t sit on the built-in (integrated) video card forever; eventually the entire motherboard will burn out, or at best one processor.

And it’s not bad, because the processor will work due to the large amount of RAM, and if you don’t have it, then there is no point in overclocking the processor.

Overclocking a processor (and more often a video card) sometimes does not make sense even if it has already been overclocked at the factory.

If you bought a processor (video card) not from an official manufacturer, such as ATI or NVidia, then they overclock it and sell it a little more expensive, citing the fact that it is supposedly more productive.

Such cases especially occur in foreign online stores Amazon, EBay and others.

Ways to achieve the goal

You can overclock the processor either through the BIOS or directly from the system. Each method has its pros and cons, and we will now look at them.

Overclocking via BIOS.

First, you need to go into the BIOS. This is done when the computer starts with the corresponding button.

In our case, it is launched by the Del button. You can recognize your button if you carefully read the labels at the initial stage of starting the computer.

Entering the BIOS is usually the first on the screen and is often written Run BIOS Setup. But, again, it all depends on the BIOS version and your motherboard.

Once you enter the BIOS, you should find a menu such as Advanced. There will be two items: CPU Tuning and PCI Clocking (they may be called differently for you).

By going into each of them, you can change the core and bus frequencies accordingly. But be careful: a processor with a factory core speed of 200 MHz will not be able to overclock to 300 or 290.

Do not suddenly increase the frequency. You should do this gradually, adding 2-5 MHz at a time.

Overclocking via BIOS has one advantage: if you make a mistake, you can reset everything to factory settings with several reboots or through a special lever on the board.

You need to pull the lever in the right direction, wait 5 seconds and the BIOS will return to its previous state.

Another plus is that the computer will not freeze during overclocking, unlike overclocking through programs. But the obvious disadvantage of such overclocking is that you will have to reboot all the time to apply the changes.

But still, this method is safer than overclocking through programs.

Overclocking through the system (programs).

If overclocking through the BIOS seems difficult to you, then you can try overclocking through programs.

You can download all the described programs both on official websites and on forums and file hosting sites absolutely free.

We would like to warn you that not a single program in the world is immune from errors, so all responsibility will rest solely with you (the programs will also remind you of this).

So, let's start with the program for AMD processors.

It's called AMD Overdrive.

The program is very easy to use, it warns the user of danger and has a number of interesting functions, including providing general and detailed information about each part of the computer, changing the frequencies of RAM, processor, benchmark, stability test, cooler management, and Auto Clock.

The last one is the most interesting. We have only seen this function in this program.

Its essence is that it itself adds the required number of hertz and tests the processor. If everything goes well, then it adds the next 1-5 megahertz and so on.

In addition, if the system suddenly freezes, Auto Clock remembers the last normal value and shows it in the menu next time. This is perhaps the only feature of the program.

The next program is MSI AfterBurner.

The peculiarity of the software lies in the interface: it is very colorful and original. In addition, you can use different skins. There are about five of them.

There is also such a program as MSI Kombustor.

This is a program (or rather, an addition to the program) that allows you to test the number of frames per second. If you download AfterBurner from the official website, they will also offer Kombustor. Of course, you can refuse it.

In addition, the utility provides the function of recording video and creating screenshots, but more on that another time.

Initially, after installing the program, you will not be able to change anything. You will only be able to observe the graph and sliders, which you can only drag to the left, which even after confirmation will not lead to anything.

If you want to do overclocking with this program, then you should play around with the settings.

To do this, go to the settings, in the first tab, find the “Compatibility Properties” section and check the box in the “Allow video adapter management and monitoring” line.

But if you have an AMD processor, then you will have to check the box in the “AMD Compatibility Modes” section and check the “Expand the limits for the official overclocking mode” checkbox there. After that, save the settings.

The program will ask you to restart the application, but there is a catch.

The entire computer will restart, so it’s easier to cancel the confirmation and restart manually. After this you will be able to overclock the processor.

Intel XTU (Extreme Tuning Utility).

The latest Intel XTU (Extreme Tuning Utility) program only works with Intel processors.

If you try to install the program on a computer with a processor from another company, the installation will not complete, notifying you that the user tried to install the program on incompatible hardware.

The program is no different in its capabilities from the previous ones. You can download it from the official Intel website.

Risks of overclocking

Of course, no matter what you do with your computer, no matter how much you “pervert” it, there will always be risks. And at overclocking they are the worst.

If you do everything wrong, you may simply be left without a computer.

The most important and only risk from which all others increase is overheating.

You should always keep an eye on your CPU temperature. This can be done both through overclocking programs and through special other monitoring programs, for example, Speccy from Piriform or Core Temp.

If you decide to do overclocking, then you must make sure that you have a normal one in your system unit. If the temperature rises above 70 degrees, the consequences will be dire, the CPU will burn out.

Well, along with the processor, the motherboard, RAM and video card can burn out. So, as they say, measure twice, cut once. Be careful and do not demand the impossible from your computer.

I hope the article helped you fully understand this difficult task of overclocking a processor. I wish you all success, see you again!!!

Throughout the development of the entire human race, stones have been our integral companions. Axes, arrowheads... pyramids in the end! Silicon alone is worth something - after all, it was thanks to it that we got fire. Albeit not so long ago, but already in the name of the development of the computer industry in the “Bronze” Age, people decided to torment their “stones” again. How it all started, we are afraid to even think. Either since the ancient Z80, or later, on the 286/386 series of processors, at some point a certain group of people discovered a new exciting activity, or rather, became the founder of a new direction - overclocking. The word, strictly speaking, is not ours; it is translated from English as “promotion”. Our definition has taken a slightly different form - acceleration, that is, increased productivity. We will tell you what it is and how it happens in this article.

Where did it start

In those glorious years, when prices for computer components literally went through the roof, processors were not so easy to overclock. If now overclocking a computer is practically no difficulty - the presence of a keyboard and the appropriate software allows you to do this in literally a few minutes - then increasing the clock frequency occurred with the use of a soldering iron, rearranging jumpers and short-circuiting the legs of the processors. That is, at that time, overclocking was available only to a select few - brave, dedicated and experienced technicians.

But it wasn't just processors that could be overclocked. Graphics cards and RAM came next, and more recently, enthusiasts have achieved improved optical mouse performance.

Why is this necessary?

And, in fact, why are we going to do something? Let's add up all the pros and cons in order to understand whether we really need it? The advantages include the following points:

• Increased productivity has never bothered anyone before. Its increasing quantity cannot be accurately predicted; it all depends on the components used. For example, the gain from overclocking a processor with a powerful graphics card almost always increases speed in 3D applications. Although, even without the goal of improving gaming performance, computer productivity as a whole will extend to archiving, transcoding, video/audio editing, arithmetic calculations and other useful operations. But the gain from “tuning” the memory will most likely not be as big as from overclocking the processor or video card.
• Many of the concepts you will learn during overclocking will provide invaluable experience.

And here is the other side of the coin:

• There is a risk of destroying the equipment. Although it depends on your hands, the quality of the components used and, finally, the ability to stop in time.
• Reducing the life of overclocked components. Here, alas, nothing can be done: with increased voltage and a very high frequency, coupled with poor cooling, the service life of the hardware can be reduced by half. This may seem unacceptable to many, but there is one detail: on average, the life of a modern processor is ten years. Whether it is a lot or a little, everyone decides for himself. We just remind you that as of today, progress has reached such a speed of development that a processor released two or three years ago is considered unacceptably outdated. What can we say about five...

Basic Concepts

Having designed a processor, the manufacturer creates a whole series (line) with different characteristics, often based on one single processor. Why, tell me, do two identical processors have different frequencies? Do you really think that the company that produces them manages to program each processor to a certain frequency? Of course, there is another way. The frequency of the younger processors in the line can easily reach even the older ones, moreover, sometimes exceeding it. But hidden problems lurk on all sides, one of which is the question of successful selection of the “stone”... however, this is another story, which we will tell about next time. Because to further study the material, it is necessary to familiarize yourself with all the terms that will appear in the text one way or another.

BIOS(Basic Input-Output System) - Elementary input/output system. Essentially, it is an intermediary between the hardware and software environments of the computer. More specifically, it is a small configuration program containing settings for all the hardware contents of your computer. You can make your own changes to the settings: for example, change the processor frequency. The BIOS itself is located on a separate chip with flash memory directly on the motherboard.

FSB(Front Side Bus) - The system or processor bus is the main channel for communication between the processor and other devices in the system. The system bus is also the basis for determining the frequency of other computer data buses, such as AGP, PCI, PCI-E, Serial-ATA, as well as RAM. It is this that serves as the main tool in increasing the CPU (processor) frequency. Multiplying the processor bus frequency by the processor multiplier (CPU Multiplier) provides the processor frequency.

Beginning with Pentium 4, corporation Intel began to use technology QPB(Quad Pumped Bus) - aka QDR(Quad Data Rate) - the essence of which is to transfer four 64-bit data blocks per processor cycle, i.e. with a real frequency of, for example, 200Mhz we get 800Mhz effective.

At the same time, the once competing AMD Athlon transmission occurs on both edges of the signal, as a result, the effective transmission speed is twice as high as the real frequency; 166Mhz in the Athlon XP gives 333 effective megahertz.

The situation is approximately the same in the line of processors from AMD- K8, (Opteron, Athlon 64, Sempron (S754/939/AM2)): the FSB bus has been continued, now it is only a reference frequency (clock generator - HTT), multiplying by a special multiplier we get the effective frequency of data exchange between the processor and external devices. The technology was named Hyper Transport - HT and is a special high-speed serial channel with a clock frequency of 1 GHz at "double" bit rate (DDR), consisting of two unidirectional buses with a width of 16 bits. The maximum data transfer rate is 4 Gbit/s. Also, the processor frequency, AGP, PCI, PCI-E, Serial-ATA are generated from the clock generator. The memory frequency is obtained from the processor frequency, thanks to a reduction factor.

Jumper It is a kind of “contact closure” assembled in a miniature housing. Depending on which contacts on the board are closed (or which are not closed), the system determines its own parameters.

CPU

CPU multiplier(Frequency Ratio/Multiplier) allows us to achieve the final processor frequency we need, while leaving the system bus frequency unchanged. Currently, in all Intel and AMD processors (except Athlon 64 FX, Intel Pentium XE and Core 2 Xtreme) the multiplier is locked, at least upward.

CPU cache(cache) - a small amount of very fast memory built directly into the processor. The cache has a significant impact on the speed of information processing, since it stores data that is currently being executed, and even those that may be needed in the near future (this is managed by a data prefetch unit in the processor). The cache comes in two levels and is designated as follows:

L1- first level cache, the fastest and least capacious of all levels, directly “communicates” with the processor core and most often has a divided structure: one half for data ( L1D), the second - instructions ( L1I). The typical volume for AMD S462 (A) and S754/939/940 processors is 128Kb, Intel S478\LGA775 - 16Kb.

L2- the second level cache, which contains data ousted from the first level cache, is less fast, but more capacious. Typical values: 256, 512, 1024 and 2048Kb.

L3- in desktop processors it was used for the first time in the Intel Pentium 4 Extreme Edition (Gallatin) processor and had a capacity of 2048Kb. It has also found a place in server CPUs for quite some time, and should soon appear in the new generation of AMD K10 processors.

Core- a silicon chip, a crystal consisting of several tens of millions of transistors. He, in fact, is a processor - he is engaged in executing instructions and processing data coming to him.

Processor stepping- new version, generation of processor with changed characteristics. Judging by the statistics, the greater the stepping, the better the processor overclocks, although not always.

Instruction Sets- MMX, 3DNow!, SSE, SSE2, SSE3, etc. Since 1997, with the introduction by Intel of the first MMX (MultiMedia eXtensions) instruction in the history of processor manufacturing, overclockers have received another way to increase performance. These instructions are nothing more than the concept of SIMD (Single Instruction Many Data - “one command - a lot of data”) and allow no less than the processing of several data elements with one instruction. By themselves, of course, they will not increase the speed of information processing, but with the support of these instructions by programs, a certain increase is noted.

Technical process(manufacturing technology) - along with various optimizations carried out with each new stepping, reducing the technical process is the most effective way to overcome the processor overclocking limit. It is designated by a strange combination of letters “µm”, “nm”. Example: 0.13\0.09\0.065µm or 130\90\65nm.

Socket(Socket) - A type of processor socket for installing the processor into the motherboard. For example, S462\478\479\604\754\775\939\940\AM2, etc.

Sometimes manufacturers use alphabetic names along with the numeric name, for example S775 - also known as Socket T, S462 - Socket A. Such apparent confusion can be a little disorienting for a novice user. Be careful.

Memory

SDRAM(Synchronous Dynamic Random Access Memory) - a system for synchronizing dynamic memory with random access. This type includes all RAM used in modern desktop computers.

DDR SDRAM(Double Data Rate SDRAM) - An improved type of SDR SDRAM with double the amount of data transferred per clock.

DDR2 SDRAM- further development of DDR, allowing to achieve twice the frequency of the external data bus compared to the frequency of DDR microcircuits with an equal internal operating frequency. All I/O control logic operates at half the baud rate, meaning the effective frequency is twice the actual frequency. It is produced using a thinner 90 nm process technology and, along with a reduced nominal voltage to 1.8V (from 2.5V for DDR), consumes less energy.

Real and effective memory frequency- with the advent of DDR and DDR2 memory, the concept of real frequency came into our lives - this is the frequency at which these modules operate. The effective frequency is the one at which the memory operates according to the specifications of the DDR, DDR2 and other standards. That is, with double the amount of data transmitted per clock cycle. For example: with a real DDR frequency of 200Mhz, the effective one is 400Mhz. Therefore, in designations it is most often listed as DDR400. This trick can be considered nothing more than a marketing ploy. Thus, we are given to understand that since twice as much data is transferred per clock cycle, it means the speed is twice as high... which is far from true. But for us this is not so important; there is no need to delve into the jungle of marketing.

Real frequency, MHz	Effective frequency, MHz	Bandwidth, Mbps
100	200	1600
133	266	2100
166	333	2700
200	400	3200
216	433	3500
233	466	3700
250	500	4000
266	533	4200
275	550	4400
300	600	4800
333	667	5300
350	700	5600
400	800	6400
500	1000	8000
533	1066	8600
667	1333	10600

Designation of memory by theoretical bandwidth - when buying memory, along with the usual designations like DDR 400 or DDR2 800, in our case you can see names such as PC-3200 and PC2-6400. All this is nothing more than a designation of the same memory (DDR 400 and DDR2 800, respectively), but only in theoretical bandwidth, indicated in Mb\s. Another marketing ploy.

Memory designation by access time- the time during which information is read from a memory cell. Denoted in "ns" (nanoseconds). In order to convert these values ​​into frequency, you should divide 1000 by the number of nanoseconds. Thus, you can get the real operating frequency of the RAM.

Timings- delays that occur during operations with the contents of memory cells, given below. This is by no means their entire number, but only the most basic:

• CAS# Latency (tCL) - the period between the read command and the start of data transfer.
• tRAS (ACTIVE to PRECHARGE command) - the minimum time between the activation command and the closing command of one memory bank.
• tRCD (ACTIVE to READ or WRITE delay) - the minimum time between the activation command and the read/write command.
• tRP (PRECHARGE command period) - the minimum time between the closing command and reactivation of one memory bank.
• Command rate (Command Rate: 1T/2T) - command interface delays resulting from a large number of physical memory banks. Manual configuration is currently only possible on non-Intel chipsets.
• SPD (Serial Presence Detect) is a chip located on the RAM module. Contains information about the frequency, timings, as well as the manufacturer and date of manufacture of this module.

Theory

How exactly we will exceed the nominal processor frequency, you already guessed, right? Everything is as simple as a donut: we have a system bus (aka FSB or clock generator - for AMD K8) and a processor multiplier (aka multiplier). We simply change the numerical values ​​of one of them and at the output we obtain the required frequency.

For example: we have a certain processor with a standard frequency of 2200MHz. We begin to think, why is the manufacturer so greedy when in the same line with the same core there are models with 2600MHz and higher? We need to fix this! There are two ways: change the processor bus frequency or change the processor multiplier. But first, if you do not even have basic knowledge of computer technology and are not able to determine the standard FSB frequency or its multiplier by the name of the processor alone, I advise you to use a more reliable method. There are programs specifically for this that allow you to obtain comprehensive information on your processor. CPU-Z is the leader in its segment, but there are others. You can equally well use SiSoftware.Sandra, RightMark CPU Clock Utility. Using the resulting programs, we can easily calculate the FSB frequency and processor multiplier (and at the same time a bunch of previously unknown, but damn useful information).

Let's take, for example, the Intel Pentium 2.66GHz (20x133MHz) processor on the Northwood core.

After simple operations in the form of raising the FSB frequency, we get 3420MHz.

That's how it is! We can already see how convolutions have begun to swarm in your minds, multiplying unthinkable numbers by monstrous coefficients... not so fast, friends! Yes, you understood everything perfectly: to overclock we will need to either increase the multiplier or the system bus frequency (and best of all, immediately, and, most importantly, more - approx. hidden internal greed). But not everything is so simple in our life, there are plenty of obstacles in our lives, so let’s get acquainted with them before we begin.

You already know that most processors on the market have a locked multiplier... well, at least in the direction we would like - towards an increase. Only lucky owners of AMD Athlon 64 FX and some Pentium XE models have this opportunity. (Options with rare Athlon XPs released before 2003 are not considered). These models can drive their already “high-frequency” “stones” practically without any problems (fussing with memory and insufficient FSB frequency reserve on the motherboard). The unlocked multiplier in this series of processors is nothing more than a gift to users who have given quite a lot of money. Everyone else who is not able to spend $1000 on a processor should go (no, not through the forest) just a different route...

Increasing the FSB or clock frequency. Yes, this is our savior, which in almost 90% of cases is the main tool for overclocking. Depending on how long ago you purchased your processor or motherboard, your standard FSB speed will vary.

Since the first Athlon from AMD and Intel Pentium on the S478, a 100MHz system bus has been the standard. Then the Atlons switched first to 133, then 166 and finally ended their life on a 200Mhz bus. Intel also did not sleep and gradually increased the frequencies: 133, then immediately 200, now 266, and even 333MHz (1333Mhz in QDR terms).

That is, having a modern motherboard with good potential for increasing the frequency of the clock generator (in fact, this quartz, which controls the FSB frequency, can also be referred to as PLL), everything becomes extremely simple - this is an increase in the frequency itself. To what extent and how to actually change it, we will talk a little later.

We hope you haven't forgotten what FSB is? No, we don’t mean the megahertz at which it works, but the immediate meaning. FSB is a system bus that connects the processor to other devices in the system. But at the same time, it is the basis for shaping the frequency of other buses, such as AGP, PCI, S-ATA, as well as RAM. So what does this mean? This means that when we increase it, we will automatically increase the frequencies of AGP, PCI, S-ATA and RAM. And if increasing the latter within reasonable limits only benefits us (currently only motherboards based on the NVIDIA nForce4 SLI Intel Edition chipset can overclock the processor regardless of memory), then we can overclock S-ATA, PCI and AGP with PCI-E completely no need. The fact is that they are quite sensitive to such experiments and respond to us with very unpleasant consequences. The ratings of these buses are: PCI - 33.3Mhz, AGP - 66.6Mhz, SATA and PCI-E - 100Mhz. And it is highly not recommended to significantly exceed them. Unstable operation of the same S-ATA can lead to loss of data from your S-ATA drive!

That is, this is a very significant limitation... there was. But the point is this: realizing the benefits of such a miscalculation, some chipset manufacturers decided to fix this problem themselves. It all started with the use of special dividers that automatically switch PCI and AGP buses to the nominal at 100, 133, 166... ​​MHz. (and such interesting situations arose in which the processor was stable at 166Mhz, initially working at 133, but at 165 - not at all!), now you understand why. But not everyone learned this lesson. You don’t need to look far for examples: the VIA K8T800 chipset, released at the beginning of the Athlon 64 era. Having very good functionality and price, it simply cannot fix PCI\AGP\S-ATA frequencies when increasing HTT. That is, you will not get more than 220-230Mhz increase in the clock generator. That's so sad, gentlemen. Be careful not to fall for such a chipset (although it is a little old).

Thus, we put an end to this section of the article and move on to the next one. We looked a little at the theoretical part, plus a few nuances that may come your way. It's time to get down to business. At the same time, figuring out along the way what other sticks need to be taken out of the wheels.

To be continued…

Website about overclocking computers - What is overclocking and what does it mean?

Introduction

WHAT IS OVERCLOCKING

What is overclocking, or, more simply put, PC overclocking? Overclocking is a user action that is aimed at increasing computer performance through three main methods of overclocking: 1-increasing the voltage supplied to a certain device, 2-increasing the frequency of the bus or memory, 3-installing new drivers.

WHY IS OVERCLOCKING NEEDED?

Let's assume that you bought your computer 2-3 years ago. Time does not stand still, and more and more new user needs appear that are not satisfied by the hardware of the computer. But you don’t have a ton of money, and the toad is choking, and you are not going to buy new hardware. Now the question is: what will you do? Should you work on your computer in the primitive text editor “Word and Deed” (as my computer science teacher does) or will you still think about overclocking your computer? Most people will answer: “I’m not a computer science teacher, and I can use the materials on this site to improve the work of my iron friend (computer, of course).” Also a very important detail - overclocking is FREE! It is provided by component manufacturers. But more on this in the next subparagraph.

WHY OVERCLOCKING IS POSSIBLE

So, why is overclocking provided by the hardware manufacturer itself? But the point is this: at a certain factory, for example, at NVidia, a batch of GeForce 6600 video cards was brought to life. Some of them did not pass the test (their temperature exceeded the permissible limit), and in order not to throw these video cards in the trash, NVidia decided to release them in light under a lower "name" - GeForce 6200, for example. And we can apply one of the ones mentioned in the section WHAT IS OVERCLOCKING ways to “grow” 6200 to 6600!

OVERCLOCKING THEORY

Let's look at the first way to increase productivity - increasing the voltage supplied to some device. Everyone knows that when the voltage supplied to a light bulb increases, it begins to burn brighter. The same goes for PC components. They, of course, do not start to glow brighter, but there is an increase in performance, and a noticeable one. Only, unfortunately, the processor and video card are more expensive than a light bulb, and if you increase the voltage above critical, they may be damaged. So be careful! The practical part will be discussed in the section ACCELERATION .

Second way - increasing the system bus frequency. The easiest and most effective way to overclock. The action takes place above the motherboard. There are also three ways to overclock this type:

increasing the frequency using a DIP switch on the motherboard;

increasing frequency using BIOS;

increasing the frequency by software.

The first option is the most idiotic. Not all uteruses have these switches, so screw them. The second option is for more or less advanced lamers and programmers in this environment. The easiest way is the third. Agree, it’s easier to sit in Windows and look at all these beautiful bells and whistles than to sit in BIOS, the interface of which is slightly better than in DOS. The second method will be discussed in the section ACCELERATION . The third method, that is, installation files of programs for overclocking, see the section FILES .

Third way - installing new drivers. Drivers provide communication between devices, and, therefore, the newer the drivers, the faster the device will work. Drivers can be found on the website of the manufacturer of your components.

COOLING

There are several types of cooling: air, water and cooling using Peltier elements. Let's look at these methods in more detail. . Air cooling is used on almost all computers. A radiator and fan are used for cooling. Such an assembled device is called cooler. The best cooler manufacturer today is Zalman. Coolers are more expensive than regular ones, but they are more reliable and have higher performance. Water cooling is less safe than air cooling, but it is more efficient. These coolers are also easy to install, which is why they are widely used among overclockers. Cooling using Peltier elements is the most exotic way to overclock your computer. It is for those who will not spare money to get the most out of your computer. True, this method has some disadvantages, the main one of which is condensation.