Processor clock overclocking. The best programs for overclocking the Intel processor. Overclocking an Intel processor with SetFSB

Everything written below is provided for general educational purposes only :) The author does not bear any responsibility for damage to anything (anyone) as a result of the actions referred to in this material.

I'm interested in the overclocking process itself. What needs to be done specifically?

First, carefully study the instructions for the existing hardware. Find the jumpers / jumpers / BIOS menu items that are responsible for the FSB frequency, memory bus, multiplier, divider for PCI and AGP. To taste - go to the manufacturer's website for a new firmware version for flashBIOS. Actually, everything - you can change the parameters within reasonable limits. Don't forget about cooling. The situation with AMD processors deserves a separate paragraph dedicated to it:

I'm not satisfied with the speed of my PC. I understand that overclocking will help me?

Not necessary. It depends on the specific programs you are working with. For example, for graphics packages (especially for 3DStudio or Maya), there will most likely not be enough memory (64 MB, it may run and will be, but it will not work, 128 MB is the minimum amount for such programs) than CPU clock speed , and for games it is more important which 3D accelerator is present in the system (although a weak processor will not be able to fully load a modern video card with work). But overclocking the system bus increases the speed of other components, so sometimes it helps a lot.

Should I "drive" my new ***-***MHz?

I do not advise you to do this out of sporting interest. If you are really not satisfied with the speed of work, then maybe you should buy more memory, which now largely determines the speed of work - for example, a fairly popular game :) Unreal Tournament is recommended to run on systems with 64MB of RAM least, I'm not talking about Windows 2k, which "loves" megabytes more than megahertz. For normal operation now you need at least 128 MB of RAM. But since there is no money, but you still want to make a "low cost upgrade", you should think about the consequences. It is unlikely that a new processor will cost you less than the same 64 or 128 MB of memory, and the increase rarely rises above 20-30 percent at best (which, however, is quite a lot :)).

What can happen to my system when overclocking?

The main enemy when overclocking a computer is temperature. The average processor (not overclocked) usually heats up to 40-50 degrees C, if you do not play Quake III _maximal_ t lies within 70-90, then this is still tolerable. One way or another, often tricks should be expected from other components. For example, the standard dividers for the PCI bus are 2, 3 and 4 (66, 100 and 133 MHz on the system bus, respectively), when set to 75 MHz (almost painlessly tolerated by any processor), the PCI frequency increases to 37.5 - in principle, there are no special objections no. But at 83 MHz at FSB, it increases to 41.5, which is not easily perceived by all boards (especially if there are a lot of them).

The frequency of AGP also increases - some video cards may not work.

Do not forget that the built-in IDE controller also "hangs" on the PCI-bus, so data loss on the hard drive is possible (more on that below).

It should be noted that "not all frequencies are equally useful" :) So, for example, overclocking a Celeron to FSB 100 MHz on a board with a BX chipset is a "personal business of the processor" (if the memory is PC100 or better). At the same time, if you overclock P3 to FSB 150 MHz on the same board, the increased load will fall on all nodes of the system, because everything they will work in non-standard mode. In the latter case, it is impossible to say anything about the stability of the work.

There are times when an overclocked processor burns out. Sometimes the motherboard is damaged as well. This is mainly due to the use of insufficiently high-quality components when assembling the system. In any case, in the case of overclocking "and (as, indeed, everywhere;)) you should be guided by common sense and not try to get a triple increase in performance. Especially when it's not about processor speed:

I overclocked the processor ... in general, it seems to me that it burned out. What to do?

First, you need to make sure that the problem is in the CPU. If smoke comes out from under the radiator and smells of burning, then, of course, there can be no particular doubts. But if the computer simply does not boot (only the BIOS splash screen or a black screen is displayed), then the reason may be different. For example, in a low-quality IDE controller or video card (let me remind you that when using non-standard system bus frequencies, AGP also starts to work in "overclocked" mode). You can try to unplug hard drives and CD-ROM cables, as well as sound cards, modems, etc., from the connectors on the motherboard. Or try using a POST Card (on some models of expensive motherboards, POST displays are built into the board itself). But keep in mind that some instances may simply not start at the FSB frequency that you have set. So the bus should be accelerated smoothly, and if the system does not want to work on one of the options, stop at the previous one.

One of my friends overclocked the computer and all the information on the hard drive "flew" from him. Why?

Some models of IDE drives that support UltraDMA are sensitive to the PCI bus frequency, and sometimes data can be lost when setting non-standard frequencies. At the same time, the hard drive itself, as a rule, remains operational, however, in some cases, servo marks may "go to the forefathers", after which it will be easier to throw away the hard drive than trying to fix it (fortunately, this is not very likely). You can usually deal with this by changing the operating mode of the hard drive (for example, by forcing it to work exclusively in PIO mode).

OK, I overclocked my *** - ***MHz to ***MHz. Turned on, it works. And now what?

An overclocked "stone" can work for some time, at first glance, normally or with rare freezes, and then burn out. The same can be said about other PC components. There is no guarantee that everything will work reliably. And yet - extreme modes shorten the "life" of the equipment. But despite the fact that the service life of most CPUs is 10 years: Although, again, it all depends on the overclocking conditions and the specific configuration. Try to work a little, run a couple of tests. If the results are satisfactory, you can relax by taking appropriate measures, which will be discussed below.

What ways of overclocking are available today?

The two methods of overclocking are to increase the multiplier and increase the clock frequency of the bus. The purpose of all this is the same - to force the processor to operate at a higher internal frequency than it was assigned by the manufacturer. For sixth-generation Intel processors, the first method is practically inapplicable (except for early models, but more on that below), everything is going to the point that the second one will be unavailable soon.Whether it will be or not - let's wait and see, but at the moment it remains only to increase the frequency (with or without an increase in the supply voltage).In the case of AMD, everything is different At the moment, Athlon and Duron processors do not have a hard limit on the multiplier, but increasing the bus frequency is almost impossible - the Alpha EV6 bus is used, in which data is transmitted on two signal edges, i.e. at an actual frequency of 100 MHz, the bus works, as it were, at 200. This whole system is very complex and exceeding the frequency parameters by more than 5 MHz often leads to a violation of its operation.

What is a "fixed multiplication factor"?

The internal frequency at which the processor operates is determined as follows: the system bus frequency is multiplied by a factor. For example, the multiplier for Celeron 400 is 6 (6*66~400). If earlier it was possible to overclock the CPU frequency by increasing the multiplier, now we do not have this possibility. Of the old processors, the multiplier is closed for some batches of Pentium 120 and 133. For all new Pentium IIs, the coefficient has become limited from above (i.e., for Pentium II 266, coefficients up to 4 inclusive are possible, but not higher). 100% multiplication blocked for SL2W8 300 Mhz PII OEM and SL2W7 266 Mhz PII OEM. There is no way to unlock it, even with ABIT BH-6 and B21. Starting with Celeron, all Intel processors come out with a hard-coded ratio (ignoring the value set on the motherboard). This also prevents overclocking on the bus to some extent. it is impossible, for example, to set the 5*100=500 MHz mode on the same Celeron 400 (which would give a good increase in speed, almost painlessly for the processor). This does not yet apply to AMD processors, in which it is fixed, but can be changed by an overclocker (see below).

True, there is one thing here - if this is a new processor from trial batches, the coefficient is usually not fixed there yet. Yes, and such processors overclock much better than their later serial counterparts.

Is there a way to get around this limitation?

For Intel Pentium II and later processors, generally not. It is believed that Abit B*6 motherboards allow this, but the method adopted by them does not work with processors released in 1999 and later.

Some thoughts by Dmitry Tyurin:
There are some considerations on this issue. Specifically, I would like to change the multiplier on the Celeron-266 to 3-3.5 (so that either 112*3.5 or 133*3 works). After a long reading of the datasheet from Intel and the remarks of various people at the conference, the following is obtained: when the power is turned on, when determining the type of processor, it tells the BIOS its multiplier and the BIOS sets it on the corresponding legs of the processor (signals LINT, LINT, A20M#, IGNNE#; legs - B16, A17, A5, A8) L or H values ​​(Intel does not explain what L and H are, but most likely banal 0 and 1). All this passes by jumpers with a multiplication factor (SoftMenu), apparently, the corresponding legs are cut off from the multiplexer. Now why do I think so: one person in the conference wrote that on the Gigabit mother he changed the processor multiplier, but before the first cold reboot. Intel says that the factor can be changed in low power mode, and gives possible values ​​(4, 4.5, 5 but who believes Intel :-)). Perhaps Abit BH6 works on the same principle. The idea is simple - to prevent the BIOS at boot from correctly determining the multiplier by sticking or planting legs B16, A17, A5, A8 on GND. It would be interesting to know if someone has already done such an experiment.

Contact B21...

Many motherboards (in particular, those manufactured by Intel itself) do not allow you to set the FSB frequency manually, choosing it automatically. Contact B21 (in slot processors) indicates the frequency required by the processor. The way to get around this is to isolate this contact (for example, with tape). It is also possible to use a socket processor on an adapter that has the possibility of such blocking initially.

It should be noted that most modern boards ignore FSB auto-detection, allowing you to set the desired value from the BIOS or jumpers.

What is the difference between OEM and Retail processor delivery options? I heard retail is better at chasing?

In the OEM version, the kit contains only the CPU in a plastic package, and it is, accordingly, cheaper. Retail (or boxed, boxed) comes in a colorful box, which contains installation instructions, a cooler (and a pretty good one), and, of course, a processor :). It cannot be said that the chips themselves are somehow different. The cooler plays an important role in overclocking. Boxed processors usually use AAVID coolers, which provide better cooling than the no-name ones you'll most likely be offered when buying an OEM option. On the other hand, in the case of OEMs, you can try to find the most optimal cooler, as well as experiment with different brands of thermal paste and achieve better cooling (in the end).

What processors are best known for overclocking?

In general, such qualities of the CPU differ from sample to sample, but there are some models that have a higher average overclockability indicator. Examples are Pentium 166MMX (at one time it worked at frequencies up to 250 MHz), Celeron 300A and 333 PPGA (they work stably even with a frequency increase of one and a half times, at an FSB frequency of 100 MHz, or even higher). It is worth considering that the ability to work at a higher clock frequency does not always provide much greater performance. For example, the Celeron 660 runs up to 1 GHz, while running slower than the PIII-700 and PIII-500E overclocked to 750 MHz.

AMD has its hits too. So, for example, after the cessation of production of K6, a certain number of K6-2 350 were marked at 200 and 233 MHz (in order to fulfill orders for processors of this frequency). In many cases, they were able to overclock to 400-450 MHz (ie, actually twice).

Which cooler is best for an overclocked CPU?

If the processor is boxed - the one that came with it in the kit. If the manufacturer of the cooler installed on the CPU cannot be identified, you will have to spend some money (perhaps up to $30) on a high-quality fan. Examples are ElanVital, AAVID, TennMax, AVC products.

I heard there is such a program - CPUIdle. What is it for?

The point of using it is that it monitors processor idle periods (idle) and turns it off using the HLT instruction, which is found in almost all new CPUs. At this time, the heat dissipation of the crystal decreases, which prolongs its life, even if it operates in normal mode (not overclocked). If your computer has the MotherBoard Monitor program and the ability to control the processor temperature, CPUIdle works with it, automatically switching the processor to suspended mode when the thermal parameters go beyond the set limits.

In general, using this utility will reduce the temperature of the processor by about 10 C, although if you overclock the processor in order to play Quake, the CPU will not be idle and this utility will have almost no effect, except for temperature control and emergency shutdown.

It should be noted that the HLT function is already built into Windows NT/2000 and many UNIX-like systems, and the ability to "raise an alarm" in case of overheating is built into the BIOS of some motherboards.

The CPUIdle website provides a list of supported hardware, but I will say right away that all more or less modern processors work with this program.

How to ensure that the processor does not overheat?

To do this, there are a lot of programs that allow you to monitor t CPUs, boards, fan speeds, etc., but the main condition is that your motherboard supports this feature - almost all new ones have it. Here are the addresses where you can get programs for monitoring the CPU:

  • MotherBoard Monitor is one of the best, freeware.
  • BCM Diagnostics is a set of programs for evaluating PC performance, but the main feature is the presence of Hardware Monitor.
  • Winbond Hardware Doctor is no different, it allows you to monitor all parameters at the same time and warn if they go beyond the established limits.

All this and much more ;) can be found at www.tucows.com and other similar servers.

How can I lower the temperature of the "iron" during overclocking?

There are many ways to do this - from removing the case cover to installing a liquid nitrogen cooling system :). But I will list the most available:

  • First of all, you need to check the processor fan. Perhaps the heatsink has accumulated dust, and the cooler is noisy like a tractor and makes a strange tapping sound - then you just need to take action, regardless of whether you overclock your system or not. If all of the above is true, then remove the heatsink along with the cooler (in most cases it is attached to the CPU socket, if it is a Socket, if it is a Slot, to the processor cartridge). It is advisable to remove the fan (for the slot - it is highly not recommended!) And clean it from dust and debris. The same should be done with the radiator. Remove the remnants of the old thermal paste from the crystal and heatsink, you need to apply a new one in a thin layer so that it does not spread. Then assemble everything back to its original state. Naturally, you need to act carefully, without making excessive efforts.
  • The same operation will not hurt to carry out with the power supply fan, as well as with the video card cooler (if any).
  • It should be your normal practice to remove dust from the case at least once every two months. Especially a lot of it accumulates in the power supply, this has a bad effect on heat dissipation, so sometimes you need to look there too.
  • You can get a software-cooler for your CPU almost for free - this will help lower the t of the processor by several degrees.

These are, so to speak, general measures.

  • Installing a powerful radiator and cooler will radically help, but you will have to spend money. When choosing a cooling device, you need to look at the number of fins and the size of the radiator (the best option is needle), the diameter of the fan. Naturally, a good cooler should not make too loud noise and vibrate.
  • You also need to take into account such a thing as free space in the PC case - some especially monstrous devices can rest against the power supply or something else.
  • For AMD Duron and Thunderbird processors in the "new" Socket462 cases, the cooling device must be chosen ESPECIALLY carefully, because there are known cases of mechanical damage to the crystal due to too much clamping force on the radiator.

Well, a very expensive solution to the problem is the installation of water cooling. This is already exotic - it's probably easier to buy a more powerful processor for the money :)

Does the case type - AT or ATX - affect overclocking performance?

In general, yes. ATX cases have a more thoughtful arrangement of the power supply, which allows you to lower the temperature inside the case. In addition, many motherboards provide the ability to automatically turn off in the event of an abnormal temperature parameters of the CPU. Although if you have an AT standard system unit, this does not mean that you need to throw it away and buy ATX - these advantages, IMHO, are not always worth the amount by which the latter are more expensive than the former.

And if I don’t even think about something like that (the computer is dear to me, like a memory :)), is it worth taking care of all these things - cooling, different programs?

In any case, it won't hurt. The processor heats up well and in normal mode, if the cooler fails, it can burn out. If you really care about the "health" of your computer - pay attention to it.

What equipment DOES NOT run AT ALL? The so-called black list.

There is no hardware that cannot be overclocked at all. It's just that some models run worse, some better. The first concerns the IBM/Cyrix 6x86/6x86MX (M1/M2) processors. These are characterized by instability in the overclocked state and strive to burn out at the first opportunity. Older AMD K6s also don't overclock well.

Overclocking is difficult for Intel motherboards, in which almost all settings are automated and you cannot set them manually (you can only switch the FSB frequency - 66/100 / (133) MHz, some do not even have this feature).

Why increase the CPU voltage?

For more efficient acceleration. This allows you to achieve normal operation of the processor with an increase in the frequency of the system bus, but at the same time increases the chances of "burning" it due to increased heat dissipation. Of course, this is not recommended, but sometimes there is simply no other way to achieve stable operation.

The scheme for increasing the supply voltage is different for Intel and AMD processors. Let's take a look at Celeron and Pentium II/III first. The motherboard determines the voltage to apply to the CPU based on a signal from the processor itself. There are, however, some motherboards that allow you to set this value manually with some steps. But if your sample is not one of those, you need to seal the corresponding contacts on the processor with something (or insulate the "legs" if the processor is for a socket). For Athlon and Duron things are a little different. Changing the voltage values ​​​​is carried out by soldering the resistors on the processor board (for Slot) or closing the contacts on the case (for Socket). For slot processors, there is also a special device connected to the internal connector of the processor cartridge, which allows you to set different voltage and multiplier values, but I have not come across it.

Which processors are chasing more efficiently - under Slot or Socket?

Processors in PPGA (Plastic Pin Grid Array, designed for Socket) and FC-PGA designs have lower heat dissipation than SECC (Single Edge Contact Cartridge, for Slot). The socket ventilation system is more efficient, on the other hand, a more powerful heatsink or a dual cooler can be installed on the slot processor.

However, the question is rather theoretical: the release of processors for Slot 1 is gradually being curtailed.

What is the difference between overclocking AMD processors (Athlon, Duron)?

The process itself is very different from that of a PII/III or a Celeron. The main feature is that the internal multiplier is not rigidly fixed. Its value is determined by the position of the resistors (for Slot A) or copper conductors on the case (for Socket A). With some skill, these parameters can be changed. True, for a slotted Athlon, you need to open the cartridge, and the procedure for soldering resistors and connecting the necessary contacts with conductive tracks is quite complicated. But it is possible and realistically feasible at home. This should only be done if you do not care about the warranty, because the chance of damaging the presentation of the processor is quite high. For a slot processor, you will have to tinker with the resistors on the processor board, which are located at the top of it. This must be done with a low-power soldering iron, VERY carefully. With an instance for a socket, everything is simpler - just open the copper jumpers located on the case near the core and close in a certain combination to obtain the required multiplier. Some motherboards don't even need that.

Overclocking AMD models designed for Socket/Super7 is similar to overclocking Celerons and PII/III, except that they do not have a multiplier limit and can be set using jumpers on the motherboard.

What's the difference between different processor cores - like Mendocino and Coppermine?

It is, and quite serious - different cores are, generally speaking, different processors. They have different characteristics and behave differently when overclocked. Here is a brief description of modern Intel's CPU cores:

Klamath 0.35 µm, PII 233-300 MHz Used with 512 KB external cache (running at half core frequency) Used in the first Pentium II. First processor for Slot1 (Single Edge Contact Cartridge). The multiplier is only limited, but not rigidly fixed, which allows you to set frequencies up to 112 MHz on the bus. Operable at frequencies up to 350 MHz (not always).
Deschutes 0.25 µm, PII 266-450 MHz Used with 512 KB external cache (runs at half core frequency) Standard - 66 and 100 MHz FSB, but works well at 112 MHz (and sometimes more). Basically, it depends on the type of external cache chips. Cartridge - SECC and SECC2 (providing better ventilation).
Covington 0.25 µm, Celeron 266-300 MHz In fact, the same Deschutes, but without the second level cache. Due to this, it accelerates well (up to one and a half times).
Mendocino 128k L2-cache (internal, core clock), 0.25 µm, Celeron 300A-533 MHz Placing L2-cache on the same chip with the core had a positive effect on overclocking ability. In some cases it was even possible to get a twofold increase (Celeron 333->666)
Katmai 0.25 µm, PIII 450-600 MHz Use with 512 KB external cache (runs at half core frequency) From the point of view of our topic, almost indistinguishable from Deshutes. The only thing: an improved technical process allowed to bring the frequency up to 600 MHz, while for Deschutes more than 500 is a rarity. Models with index "B" are designed for FSB 133 MHz.
Coppermine 256 KB L2-cache (internal, core clock), 0.18 µm, PentiumIII 500 MHz or higher Compared to Katmai, the technological process has been changed and the cache memory now operates at the same frequency as the processor (as in Celeron). External frequency - 100 and 133 MHz, overclocking up to 150 is possible.
Coppermine128 128 KB L2-cache (internal, core clock), 0.18 µm, Celeron 533A and above Coppermine with half the cache and 66 MHz FSB. There are no advantages over the "adult" PIII, the same can be said about overclocking.

Table for AMD processors:

K6-2 (K6-3D) 0.25 µm, K6-2 266-333 MHz Supports 66, 95 and 100 MHz FSB. Overclocking is highly dependent on the specific sample (but on average the potential is low). The coefficient is not fixed in principle: 300 is obtained either as 66x4.5 or 3x100, and 333 as 66x5 or 95x3.5
K6-2 CTX 0.25 µm, K6-2 200-550 MHz An improved version of the K6-2 core. Somewhat faster and better overclockable. Processors with a frequency of 200 and 233 MHz (actually relabeled 350) are especially famous for this, often overclocking to 400-450 MHz.
K6-2+ 128 KB L2 (at core frequency), 0.18 µm, K6-2+ 450-550 MHz Contains an integrated second-level cache, produced according to a new technical process. The latter allows you to achieve speeds above 600 MHz without much difficulty.
sharptooth 256 KB L2 (at core frequency), 0.25 µm, K6-III 400-500 MHz In fact, this is K6-2 CTX, but with an integrated second-level cache. The large die area and high power consumption prevent high clock speeds from being achieved. Out of production.
K7 0.25 µm, Athlon 500-1000 MHz Use with 512 KB external cache (runs at 1/2, 2/5 or 1/3 of the core clock) The first AMD processor under Slot. The system bus is EV6 (200 MHz DDR), unstable to overclocking. The multiplier can be changed, but the procedure is not easy.
Thunderbird 256 KB L2 (at core frequency), 0.18 µm, Athlon 700 MHz or higher Significantly improved K7, available both in slot and socket versions. Overclocking results are pretty good. It correlates with the old K7 in much the same way as Coppermine with Katmai.
Spitfire 64 KB L2 (at core frequency), 0.18 µm, Duron 600 MHz or higher Thunderbird with reduced cache. Available only in Socket (462-pin) version. Accelerates great.

How suitability for overclocking depends on the manufacturing technology - 0.25, 0.18?

The more perfect the technology, the smaller the size of the crystal itself, the power consumption, and, hence, the temperature. This parameter is presented in micrometers, the smaller the number, the better the overclocking qualities of this core (and, therefore, the processor itself) will be.

One has only to take into account that if the manufacturer has already brought the core frequency almost to the upper limit, it will be difficult to overclock the processor. For example, Pentium III 450 often overclocks to 600 MHz, while Pentium III 600 is almost impossible to overclock - this frequency is actually the limit for the Katmai core (and for the memory used as a cache).

What is stepping?

Stepping means the internal version of the processor. When fixing minor flaws or errors in the microcode, a CPU modification is released that has a new version number. Usually, the more stepping, the more stable it behaves and the better the processor overclocks.

What do the letter indexes of Pentium processors mean?

They are deciphered quite simply: the index "E" (embedded) means the cache memory built into the processor core (i.e. the Coppermine core), and "B" (bus) - the 133-MHz system bus. EB, respectively, both. This is done to distinguish models with the same clock speed, but with different cache or system bus parameters, as well as to indicate processors based on the Katmai core that support 133 MHz FSB.

It would sometimes be difficult to figure it out without letter indices - in particular, there are as many as four different Pentium III 600.

How do all these abbreviations stand for - SECC, FSB, FC-PGA?

SECC - Single Edge Contact Cartridge "Knife" type of processor socket, or Slot. SECC2 The same as in the previous case, but with improved case cooling. SEPP - Single Edge Processor Package Almost the same as SECC, but without the plastic case. Used in Celeron. PPGA - Plastic Pin Grid Array. Processor pin connector (Socket). FSB - Front Side Bus Processor bus (external). Sometimes this concept is confused with the memory bus, but the frequency of the external CPU bus may not be equal to the frequency of the memory exchange bus. FC-PGA - Flip Chip Pin Grid Array Socket type of Intel processors, practically the same as PPGA (however not fully pin-compatible with it). SDRAM - Syncronous Dynamic Random Access Memory Type of memory used as RAM in most modern PCs. DDR-SDRAM - Double Data Rate SDRAM Double data rate. New type of memory. The speed of operation increases due to the transmission of information on both fronts of the signal, which at the same frequency allows you to double the peak throughput. SRAM - Static RAM Used as a processor cache. Much more expensive and faster than DRAM (particularly due to the fact that it does not require time to regenerate the content)

What about Socket->Slot adapters?

We can only say one thing: this provides more opportunities for installing new generation processors in slot motherboards. When buying a new system, it is better to take a motherboard with a socket type connector (it is cheaper, and slot cards are gradually being phased out). In addition, there is one more point: not all adapters support high FSB frequencies (for example, 133 MHz). But a more powerful heatsink can be attached to the processor installed in the adapter. Also, some advanced models have the ability to configure CPU voltage and other parameters (for example, blocking B21).

You also need to take into account that cheap adapters (like motherboards) do not have a temperature control function (more precisely, they are not able to transmit the temperature sensor readings built into the processor to the motherboard) - the main parameter during overclocking. This problem is solved by using an external sensor, but the accuracy is reduced.

What programs can be used to determine the speed of a computer?

One of the best programs of this kind is Quake III :) Here "no one will be left behind" - the memory bus, the video chip, and the processor are intensively used (you can try software rendering - it loads the CPU more).

Special programs for this in most cases can be downloaded for free from the Web (3DMark, WinBench, WinStone). You can also try the speed of work on real applications used by many in their work, for example, PhotoShop. Produced by applying various effects (Gaussian Blur, Render Texture, Radial Blur) to large files and recording the time spent on working out the effects. This allows you to really evaluate the gain in speed.

Just don't use the utilities that are part of multifunctional kits for this, for example, the SysInfo benchmark from the Norton Utilities kit, which sometimes give completely unrealistic results.

When you buy a laptop, based on its cost, you expect good performance. And still it's not enough. But you can increase the processing speed of the central processing unit (CPU) from that declared by the manufacturer. Therefore, the question arises: how to overclock the processor on a laptop in order to get more efficient work for the same money. There are several options available and safe, which we'll cover in this article.

Let's start with the question "why"

It seems that the laptop is only 3 years old and it has never failed when performing any tasks (play a new shooter, watch the latest video rentals, transcode videos).

But now it can not cope with half the needs. What will you have to do - change your laptop? But you can try to "revive" your electronic friend by overclocking the processor. Performance will increase slightly. If you do everything right, the results will please. In addition to increasing the clock frequency, memory starts to work faster, and as a result, the speed of applications increases slightly.

But overclocking a laptop processor is half the battle. You have to pay for everything in this life:

  • Performance will increase, but power consumption will also increase at the same time. This means that battery life will inevitably decrease.
  • In addition, now the laptop will heat up much more. You will have to think about the cooling system or, at least, do not block the special slots from the bottom and side.
  • CPU life is likely to decrease.

Windows also improves performance

Overclocking the processor in a laptop is a bit difficult, but possible. Mobile device manufacturers, of course, thought about protection and took care in advance to ensure the best performance at maximum frequency, when you need to speed up work. When the processor is idle, the frequency automatically drops. But you can improve performance without harming the laptop using system tools by changing the power mode.

To do this, the Windows operating system has a software tool - " Power supply". You can find it by going to Control Panel. The figure below shows the window that will appear in Windows 7 or 8.1.

You need to go to the section Power supply» and select « High performance».

This is how you can "overclock" the laptop processor without the risk of breaking something. The performance gain will be immediately noticeable.

Overclocking via BIOS

In some models, it is possible to overclock the laptop processor with standard tools from the BIOS. To get into this system, you need to turn on the laptop and press a certain key on the keyboard. A hint of which button to press will be visible for several seconds on the monitor screen. For example, on the screen of the hp monitor, the inscription shown in the figure below will appear.

After this condition is met, the initial menu will appear, which will indicate the key that you need to press to get into the BIOS.

Consider the sequence of actions that must be performed to overclock the processor:


Be warned that most often manufacturers block the CPU to prevent the user from changing the clock speed on their own.

Overclocking with dedicated apps

For relatively old laptops, it is possible to overclock the processor using a small program paired with the program Prime95.

Further implementation of any overclocking method is associated with the risk of damaging the laptop. All actions must be performed with extreme caution, in small steps.

You need to understand that the maximum possible to achieve a slight increase in the frequency of the processor - within 10-15%. A further increase is possible if the cooling system is prepared and the power supply of the chip is changed. Since during acceleration, along with an increase in frequency, heat generation also increases. By the way, modern processors have a two-level overheating protection system. If the temperature threshold is exceeded, the processor automatically forcibly lowers the frequency and voltage, which leads to a decrease in heat generation. If the temperature does not fall below 95–110º, the laptop either turns off or freezes.

CPU-Z program

Before overclocking, you need data about the chip installed in the laptop. This utility will help CPU-Z. This information is required for the program.

SetFSB utility

Deliberately designed for fast and easy CPU overclocking. With its support, you can safely change the system bus frequency without rebooting the operating system, bypassing the BIOS.

The program has an interface that is quite understandable for work, and the entire overclocking process takes place by step-by-step movement of just ONE slider.

If this laptop is supported by the program, the chip frequency data will be visible in the lower right corner.

The sequence of actions is extremely simple: increase the bus clock frequency in small steps and test it with the program Prime95.

Prime95

A small utility that can measure the performance of a computer. The measurement process is based on performing calculations of Mersenne primes. This action uses all the capabilities of the laptop.

Optionally, you can check both the RAM and the processor itself. While the program is running, you must be prepared for the fact that the computer will noticeably slow down.

The frequency increase is made in small steps until a freeze occurs. After the performance has been saved, Prime95 testing must be terminated and exited from the CPU setup program.

Conclusion

If everything went well, you can stop there. But this is not the whole complex of works. Performance depends not only on the frequency of the processor, but also on the frequency of the memory. It can also be increased by selecting the necessary timings. Tips from friends and Internet searches will help you set up your laptop. Overclocking without prior preparation can be harmful. For game lovers, the next step is overclocking the graphics card. The main thing is that all actions are thoroughly thought out, and then the efforts will not be in vain.

Related videos

CPU for overclocking | How does a processor become an overclocking legend?

Since the birth of the IBM-compatible PC, some processors have been positioned as exceptionally suitable products for aggressive overclocking. Some models are famous for their outstanding overclocking potential, while others are famous for their low price. We even remember a few unique examples where features that were initially disabled on the chip could be unlocked.

We decided to take a little digression into history and compile a list of some of the most interesting ones in terms of overclocking the CPU.

CPU for overclocking | Intel i486

Although overclocking existed before the appearance of this chip, this process became much more interesting with the advent of Intel 80486 due to its flexible clock speed settings and the first implemented internal clock frequency multiplier in the i486 DX2 model. Introduced in 1992, the DX2 was available in three versions: 40 MHz (20 MHz x2), 50 MHz (25 MHz x2), and 66 MHz (33 MHz x2). Computer enthusiasts could buy a cheaper version of the i486DX2-40 and raise the clock speed to 25-33 MHz using a jumper on the motherboard, resulting in the performance of the flagship i486DX2-66 model.

It may not seem like much today, but overclocking like this provided a 60% increase in frequency when computer manufacturers were paying $600 for 486DX2-66 in batches of 1000, and the cost of a CPU upgrade kit could exceed $1000. Buying the i486DX2-40 and DX2-50 saved hundreds of dollars, making overclocking a very attractive option for PC enthusiasts.

CPU for overclocking | Intel Pentium 166 MMX

Intel released the Pentium MMX in 1997, equipping it with an extended instruction set and twice the L1 cache (a whopping 32 KB at the time) compared to the first generation of Pentium processors. In addition to being noticeably faster than their predecessors, these chips also offered extensive overclocking capabilities. The top-of-the-line Pentium MMX 233 was about $600 at the time of release, but the 166 MHz version was $200 cheaper and could usually reach 233 MHz with little effort. Many of these CPUs were able to hit the 250 MHz mark with an FSB boost of up to 83 MHz, making the Pentium MMX 166 a top processor at a reasonable price.

CPU for overclocking | Intel Celeron 300A

Despite its advanced age, the Celeron 300A is still respected in overclocking circles, and it is this chip that is responsible for the fact that many people joined the ranks of overclockers in 1998 (among them there are employees of the site). The processor was made on the Mendocino core, designed for low-budget PCs. Intel decided to cut costs by locating the L2 cache directly on the CPU die, instead of using an external cache card like it made for its top-end Pentium II processors. Although the Celeron only had 128KB of L2 cache instead of the Pentium II's 512KB, the on-chip cache meant that it ran at the processor's own frequency, and in many cases this gave the Celeron 300A an advantage over more expensive CPUs. In addition, the $180 Celeron 300A had incredible overclocking potential: increasing the FSB from the factory frequency of 66 MHz to 100 MHz made it possible to reach 450 MHz, on par with the $500 Pentium II 450. For the first time in history, overclockers could get the performance of a flagship CPU for less than $200 with a minor tweak. Not surprisingly, the Celeron 300A is fondly remembered by members of the overclocking community, to which it was directly involved.

CPU for overclocking | Pentium III 500E

If Celeron brought a very large number of advanced PC users into the ranks of overclockers, then the Pentium III 500E successfully continued his work. This chip, introduced in 2000, was made using a 180 nm lithographic process, equipped with a 256 KB L2 cache, and led to a change in the Slot 1 interface to a more modern Socket 370. Unlike the cut-down Celeron processors, the Pentium III 500E (cost $240 at the time of release) was identical in terms of architecture to the Pentium III 750 MHz ($800). Naturally, it provided aggressive overclocking to 750 MHz by simply increasing the FSB to 150 MHz, coming close in performance to the rare and expensive ($1000) Pentium III 1 GHz.

CPU for overclocking | AMD Athlon and Duron 600 (Thunderbird/Spitfire)

The first generation Athlon was a cartridge that hides the processor board with the CPU and cache chips installed. The cartridge was installed in the Slot A slot connector with 242 contacts. Since the design of the cartridge was completely closed to the user, a separate device called Gold Finger was used to unlock the multiplier, with which it was also possible to change the CPU voltage. These processors themselves had excellent overclocking potential, but in 2000 they were replaced by the next generation on the Socket A Thunderbird/Spitfire core, and multiplier overclocking became easier thanks to the famous L1 bridges. All that had to be done was to connect four small bridges on the CPU case with a graphite pencil (or better yet, using a special conductive pen) to unlock the multiplier. Duron 600 for $80 could be overclocked to 1 GHz, which brought it very close in performance to Athlon 950 ($360). The cost of CPUs that are of interest in terms of overclocking has dropped below $100.

In addition, the more expensive Athlon processors could be overclocked to over 1 GHz at a time when Intel's top Pentium III models were relatively overpriced, if they could be found at all: Intel processors over 1 GHz were extremely rare for months. after the announcement. With the advent of Thunderbird's successor, the Palomino-based Athlon processor, the pencil bridge trick became obsolete, but that was after Athlon and Duron were able to attract a huge number of overclockers to their camp.

CPU for overclocking | AMD Athlon XP-M 2500+

With AMD locking down the multiplier on desktop CPUs, overclockers have realized just how much multiplier potential the mobile versions have. At $25 more than desktop CPUs, Barton's mobile processors offered a lower stock Vcore (1.45V) and an adjustable multiplier. As a result, an Athlon XP-M 2500+ processor running at 1.83 GHz could often be overclocked to 2.5 GHz without much effort. Some overclockers were able to reach 2.7 GHz when overclocking this processor.

CPU for overclocking | Intel Pentium 4 1.6A

The first Pentium 4 processor was based on the little-known Willamette core, a design that failed to impress at launch and was even a step backward in some performance and power consumption tests. But in 2001, the Willamette was replaced by the Northwood architecture, which has twice the size of the L2 cache (512 KB) and is based on a thinner 130-nm process technology.

For the first time, computer enthusiasts began to reconsider their opinion about the Pentium 4 precisely at the time of the heyday of Northwood - due to the increased scalability of this architecture. The Pentium 4 1.6A sold for about $300 and easily overclocked to 2.4 GHz with a factory cooler. This was slightly faster than the flagship $560 1.8GHz Pentium 4.

CPU for overclocking | AMD Opteron 144

While AMD's Athlon 64 processors provided excellent performance, they usually didn't have the same solid overclocking potential as the Pentium 4. However, in 2005, AMD introduced a 1.8 GHz version of the Opteron 144 for under $150. Opteron processors have always been server and workstation oriented chips that require expensive register memory. However, the Opteron 144 was a version for regular socket 939 single-processor boards that used unbuffered memory. Just as importantly, he had incredible overclocking potential. Many copies could be overclocked to 3 GHz, while the highest performing Athlon FX-57 models had a frequency of 2.8 GHz and cost $1000.

CPU for overclocking | Intel Pentium D 820 and 805

In 2005, Intel's Pentium family was often inferior in performance to AMD's Athlon 64 line. So, the most budgetary Pentium D 820 processor was estimated at $240, which is about a hundred dollars cheaper than the Athlon 64 X2 4200+.

Although the performance of the budget Pentium left much to be desired at factory frequencies, it was a full-fledged dual-core processor that in capable hands reached 3.8 GHz, and some instances even conquered the 4 GHz bar.

In 2006, the $130 Pentium D 805 processor was born - the same processor that we overclocked to 4.1 GHz in the article "Overclocking Pentium D 805: $130 Dual-Core 4.1GHz Processor". The Pentium D was able to switch the attention of enthusiasts to Intel, and this is in an era of AMD dominance.

CPU for overclocking | Pentium Dual Core/Core 2 Duo E2000/E6000/E8000

Back in 2006, the release of Core 2 Duo processors based on the Conroe architecture allowed Intel to reclaim the industry leader's crown while ushering in the golden age of overclocking. If we were to dedicate a page to every model in the lineup that had outstanding scalability, this article would be at least twice as long.

Let's start with the budget Pentium Dual Core, which was essentially a version of the Core 2 Duo with the L2 cache cut down to 1 MB. Pentium Dual Core E2140 (1.6 GHz) and E2160 (1.8 GHz) cost $80 and $90 respectively at launch, and easily conquered the 3 GHz line. The Core 2 Duo E6300 (1.866 GHz) cost less than $200 at launch but could be overclocked to around 4 GHz, on par with the flagship Core 2 Duo E6700 (factory clocked at 2.667 GHz) for $580.

Later in the Core 2 lifecycle, the Wolfdale core, which was made to 45nm, allowed processors like the 3GHz Core 2 Duo E8400 to break the 4GHz barrier with minimal resistance. This by no means applies to all Core 2 models, but in our memory there was not a single representative of the line that would not have good overclocking capabilities.

CPU for overclocking | Intel Core 2 Quad Q6600

Core 2 Quad Q6600 was introduced in 2007. But even now, there are enthusiasts who still use the capabilities of this quad-core processor, making it somewhat of an anomaly in the rapidly changing world of technological progress.

Based on the revolutionary Core 2 architecture and 65nm process technology and factory clocked at 2.4GHz, this CPU reaches the mid-3GHz frequency range with little difficulty. This was surprising at the time, given the complex architecture of a quad-core CPU.

Although the Q6600 was priced at $850 at launch, by 2010 it had dropped to $200, making it a popular choice among computer enthusiasts on a budget. In 2011, the Q6600 was replaced by the Core 2 Quad Q9550, another CPU with an excellent reputation among overclockers.

CPU for overclocking | Intel Core i7-920

Intel's Nehalem architecture was introduced in 2008 along with the Core i7 brand. The quad-core Core 2 Quad processors have proven themselves well, but the rethinking of Hyper-Threading has allowed the Core i7 to take a step forward in the types of workloads associated with parallel computing. In addition, the LGA 1366 platform is equipped with a three-channel memory subsystem, and the memory controller is implemented directly in the processor itself.

flagship model Core i7-965 Extreme(3.2 GHz) was sold for $1000 and had an open multiplier. But the $285 Core i7-920 (2.67GHz) offered an identical architecture for less than a third of that price. Although it had a locked multiplier, it was possible to raise the frequency to 4 GHz through overclocking with BCLK. In fact, the Core i7-920 is still quite productive and provides stable operation after overclocking, which indicates the long life of the Nehalem architecture and the X58 Express platform.

CPU for overclocking | AMD Phenom II X2 550 and X3 720 Black Edition

AMD's flagship Phenom II model has never had overclocking potential (the overclocking efficiency did not reach 4 GHz). But the processors in the Black Edition line at least made configuration easier thanks to an open multiplier. The Phenom II X2 550 and X3 720 had their own unique features, namely, in some cases, they allowed additional cores to be unlocked if the motherboard used supported this function.

While some of these processors did have defective cores that were impossible to bring back to life (which made such "overclocking" a lottery), a great many were capable of running as quad-core processors at times over 3 GHz. Back in 2010, when the top quad-core Phenom IIs cost $180, you could take a chance and often end up owning a higher-end processor for $100. In the worst case, for relatively little money, you became the owner of a dual-core or triple-core CPU, which was still easily overclocked thanks to an open multiplier.

CPU for overclocking | Intel Core i5-2500K

Intel introduced its chips on the architecture Sandy Bridge in 2011, and they were based on a 32-nm process technology. Compared to the top Core i7 models, the Core i5 processors lacked the 3MB shared L3 cache and Hyper-Threading feature. None of these measures resulted in a significant performance difference, except in highly concurrency load scenarios.

On the other hand, the Core i5-2500K includes an unlocked multiplier, which makes it possible to overclock the CPU from the factory frequency of 3.3 GHz all the way up to 4.5 GHz using air cooling. We think the $225 price tag is reasonable considering the high performance potential of this chip. Even today, the relatively meager advantages of the architectures make the 2500K a worthy choice for computer enthusiasts.

Unlike AMD products, Intel processors have less overclocking capabilities. For Intel, stability is a priority, which in turn reduces the likelihood of a permanent CPU failure if the speed is increased excessively. The company does not offer its customers special programs for increasing the frequency, but this can be done using special applications from other developers. Next, we will look at several such programs and describe how to overclock an Intel processor.

Processor overclocking options from Intel

There are only two ways to improve CPU performance

  1. Using Third Party Applications
  2. Overclocking using motherboard settings.

In the first case, you will need to choose a suitable program, since not all applications will be able to overclock a particular model. Using BIOS settings, in turn, is the safest way to increase performance and does not carry a large number of risks compared to the first option. Inexperienced users are not recommended to change the speed settings, as there is a risk of processor failure.

Checking for suitability for overclocking the processor

It is not always possible to increase the frequency of the CPU, and even if there is one, then you need to determine the acceptable limit for increasing it. The most important thing to pay attention to is the temperature of the processor, which should not exceed 60 - 70 degrees Celsius. Due to overheating, to monitor it, you can use the special program AIDA64, for this you will need to do the following:

Intel Processor Overclocking Option #1: CPUFSB

This is a universal application, using which you can quite easily increase the speed of the CPU. The program supports many models and processors manufactured by various companies. In addition, there is support for the Russian language.

To overclock, you will need to do the following:


Intel Processor Overclocking Option #2

This application has a more intuitive interface and is capable of overclocking Intel and AMD processors of various modifications. To use it to increase the speed, you will need:


Intel processor overclocking option #3 :BIOS

You can overclock the processor using the BIOS settings mat. boards, if provided by the hardware manufacturer. You will need:


The procedure for increasing the speed of the processor through BIOS settings may look different, depending on the model of the mat. boards, but the principle remains the same - you need to turn off automatic settings, and then set your own parameters. The most important thing when overclocking is to monitor the temperature and avoid excessive overheating.

The ability to overclock Intel Core-series processors may be slightly lower than that of competitors from AMD. However, Intel focuses on the stability of its products, not performance. Therefore, in case of unsuccessful overclocking, the probability of completely damaging the processor is lower than that of AMD.

Unfortunately, Intel does not release or support programs that can speed up the CPU (unlike AMD). Therefore, you have to use third-party solutions.

There are only two options for improving the performance of CPU cores:

  • Using third party software, which offers the possibility of interacting with the CPU. Even a user who is on “you” with a computer can figure it out here (depending on the program).
  • Using the BIOS is an old and proven method. With some models of the Core line, programs and utilities may not work correctly. In this case, BIOS is the best option. However, inexperienced users are not recommended to make any changes in this environment on their own, because. they affect the performance of the computer, and it is difficult to roll back the changes.

Find out the suitability for overclocking

Not in all cases, the processor can be accelerated, and if it is possible, then you need to know the limit, otherwise there is a risk of damaging it. The most important characteristic is the temperature, which should not be higher than 60 degrees for laptops and 70 for desktop computers. We use software for these purposes:


Method 1: CPUFSB

- a universal program with which you can easily increase the clock speed of the CPU cores. Compatible with many motherboards, processors from different manufacturers and different models. It also has a simple and multifunctional interface, which is fully translated into Russian. Instructions for use:


Method 2: ClockGen

- a program with an even simpler interface, which is suitable for accelerating the work of Intel and AMD processors of different series and models. Instruction:


Method 3: BIOS

If you have a bad idea of ​​what the BIOS working environment looks like, then this method is not recommended for you. Otherwise, follow these instructions:


Overclocking Intel Core series processors is a little more difficult than overclocking AMD chipsets. The main thing when overclocking is to take into account the recommended degree of increase in frequency and monitor the temperature of the cores.

Have questions?

Report a typo

Text to be sent to our editors: