So your interested in overclocking your new i7 or i5 cpu but your a little unsure of yourself. Hopefully this guide will help educate you a little.
Overclocking - The process of running a computer component at a higher clock rate (more clock cycles per second) than it was designed for or was specified by the manufacturer. So why would anyone want to overclock their CPU? Well there are a number of reasons to overclock with Increased Performance leading the Pack.
Increased Performance - Almost every task a PC can do from Gaming to Basic System Tasks will be faster. The increase in cpu performance also increases the bandwidth available to other components allowing them to reach their peak performance (bottlenecking).
A Cheap UpGrade - Overclocking your CPU is increasing the speed to that of a more expensive model or sometimes even faster than any model available.
Hobby - Overclocking can be Fun. There are numerous Benchmarks Threads here at Tech-Forums to participate in and Tweaking is required to move up the Leaderboard.
Educational - While some people just can't learn new tricks, as an Overclocker you will gain knowledge and hands-on experience with your PC before it's all over.
There are also a few Disadvantages to consider. With the biggest being you can damage your CPU and or other components if done improperly.
CPU Damage - Overvolting and Excessive Heat can permanently damage your CPU and other Components.
Increased Power Consumption - Depending on the amount of an overclock the increase in Power Consumption can vary.
Increased Spending - Overclocking can be Addictive. As you gain experience and confidence the Need For More Speed often includes Special Cooling Devices and Case Modifications.
Heat is probably the most important aspect of overclocking. As it pertains to overclocking, heat is a direct result of work or the ability to do work. While your sitting at your desk reading this your pc is producing heat and it has a cooling system designed to dissipate that heat based upon the the amount of work or heat that it can produce. As you overclock your cpu you are effectively increasing the amount of work it can do or the amount of heat it can produce. So we need to monitor the heat and if needed, regulate it. As a general rule we want to keep the maximum load temperature below 80dC with a target zone of 65-70dC.
Most motherboard manufacturers provide some type of hardware monitoring utility on the installtion cd. Some of these monitoring utilities are reliable and some not so much. Luckily, if you have doubts about the supplied utility there are a number of alternative monitoring utilities you can download.
CPU-Z - a freeware that gathers information on some of the main devices of your system. Highly Recommended CPUID HWMonitor - a hardware monitoring program that reads PC systems main health sensors : voltages, temperatures, fans speed. Highly Recommended.CPUID Real Temp -a temperature monitoring program designed for all Intel single Core, Dual Core and Quad Core processors.Real Temp - CPU temperature monitoring Core Temp - a compact, no fuss, small footprint program to monitor CPU temperature. http://www.alcpu.com/CoreTemp/
This subject could take pages to write so I'll make this as brief as possible while still being informative. Heat Regulation can be accomplished by several methods including air and water Cooling and a few more extreme methods like phase-change, dry ice or liquid nitrogen.
I'll start by saying that if your only wanting a small or mild overclock, your stock heatsink may be perfectly acceptable. If your planning on using the stock heatsink I recommend cleaning it with compressed air before you start and regularly cleaning it a few times a month. And, unless it's been done recently, you should apply a fresh coating of thermal paste between the cpu and the heatsink (more on thermal pastes below). If your wanting more cooling than your stock heatsink can provide, whether it's for a higher overclock or just for piece of mind, there are many very good options to choose from in various price ranges. The heatsink you choose should be based upon a few things like, available space, cpu model, price, availability, desired performance and newly released products. For this reason I recommend asking our forum members for their recommendations. I will link you to a few of my recommended manufacturers.
Becoming increasingly popular, water cooling is very effective if your willing to pay the price. A good setup can cost you $200+US and it may require some case modifications and it does require a good working knowledge of computer assembly. As for the cheaper, self contained water cooling systems, I find it impossible to recommend one at this time. In the future a new product may come out that changes my opinion, but at the present time they can be matched in performance at a lower price by a number of air cooling products. Once again if your interested in a water cooling system I recommend asking our forum members for their recommendations. Here are a few links to my personal favorite manufacturers.
DangerDen - Let us help you be cool!
Swiftech - Swiftech
Phase-Change, Dry Ice and LN2 cooling methods are all extreme methods used to cool the cpu to sub-zero temperatures. Of the (3) only the phase-change option is even remotely considered an option for everyday use. And then you should be a highly experienced overclocker with money to spend on the more than occasional replacement part. You can check out my last phase-change torture test here. Operation SuperPi - Phase-Change The only phase change I've seen for sale lately is this unit made by OCZ. OCZ Technology | Products | Cooling Products | OCZ Cryo-Z
Thermal Compounds and Lapping
Regardless of the cooling method, thermal paste should be used between the cpu and the base of the cooling device used. The exterior metallic surfaces contain small imperfections that prevent the two surfaces from mating together for optimal thermal transfer. Thermal paste will fill in the small imperfections and provide a thermal interface to maximize thermal transfer. There are two main methods of applying thermal paste. Method 1 involves manually spreading the thermal paste with a spatula like device (flexible card) or with your finger (use a plastic baggie as a finger glove to avoid direct contact with the paste). Method 2 involves applying the thermal paste (about the size of a small pea or a grain or 2 of rice) to the cpu and letting the weight of the heatsink spread the paste. Once again because of new products and different usages I recommend asking our forum members for recommendations on the various types of thermal compounds, but you usually won't go wrong with Tunig TX-2.
Sometimes the imperfections in the surfaces are so severe that thermal paste alone will not provide a suitable thermal interface for proper thermal transfer. In this situation you'll want to consider lapping one or both surfaces. Lapping is the process of smoothing and leveling the surface of the cooling device or cpu's heat spreader with an abrasive surface such as sandpaper. Bewared lapping the surface will void the warranty on the product involved. Here are a few YouTube videos to watch. The process is the same regardless of the type of cooling device.
As an overclocker you will be pushing your cpu beyond the manufacturers operating specifications.Stability is our ultimate goal. If you can't play your games or run your programs then what good was any of it. The first step in aquiring a stable overclock is to have your system stable at its default settings. A fresh install of Windows with all the Service Packs and Recommended Updates is a good way to start. If that isn't an option then it doesn't hurt to go thru the Spyware Asylum Guide Spyware Asylum. Next you want to make sure that your drivers are up to date for all the components in your system. Your bios may also have an update available, they are often released with compatibility and performance inprovements, but a failed bios flash can leave your system inoperative. At this point in time you should be ready to perform a preliminary stability test of your system. There are many utilities available that you can use for stability testing (links below),choose one and run it for an hour. LinX will heat your system up very fast and usually crash within the first ten minutes if something is wrong, so for your preliminary run of LinX set it to run for 20 minutes.
There are just to many different bios settings spanning way to many brands and models of motherboards for me to explain and recommend settings to them all. So I'll be covering the basic settings that should be available for most of the users reading this, thou the names may be slightly diferent.
PC Health Status - Hardware Monitor - System Monitor
This heading should bring up a screen showing you the values each sensor on the motherboard is detecting. As you make changes to your various voltage settings these values will reflect those changes upon a system restart. Also located under this heading are your Smart Fan settings. Unless you've replaced all the fans on your motherboard with water cooling blocks I recommend adjusting each fan setting to Manual / 100%.
This heading displays all the onboard devices built into your motherboard. As a rule of thumb, if you don't use it, turn it off.
Frequency / Voltage Control
This Heading contains the settings for Frequency and Voltage settings as implied, but it also contains the the Memory and CPU Features for my motherboard as well. Some of these settings may be located under a different heading in your bios.
CPU Clock Ratio / Multiplier
This value multiplied by the CPU Host Frequency, also known as the Base Clock(BCLK) or Quick Path Interconnect(QPI), determines the CPU Frequency. Using the i7 920 as an example, the default multiplier is 20 and the default BCLK or QPI is 133MHz, therefore 20x133MHz=2660MHz or 2.66GHz. This value can be set at it's default or a lower value, only the Extreme Edition cpu's can use a higher value.
CPU Host Frequency
This is the cpu bus speed or Base Clock(BLCK), previously known as the Front Side Bus (FSB). Increasing this affects the core frequency, CPU-northbridge frequency and QPI frequency. The default value is 133, using the i7 920 with a Multiplier of 20 as an example, changing this value results in the following CPU Frequency.
BLCK x CPU Clock Ratio = CPU Host Frequency
With older chipsets the MCH Strap or the Memory Controller Hub Strap would allow you to select a value which then loaded a pre-defined set of values for the memory subsystem. A lower setting would increase performance but it sacrificed stability or compatibility and a higher value would have the opposite effect. So while I'm not positive, I think the MCH Strap setting is basically operating in the same manner with the X58 and P55 chipsets.
With the setting at Auto the MCH Strap is automatically adjusted to match the current memory speed setting. Adjust this setting higher if your stable overclock is lower than expected within your current temperature range.
As an example, your at 20x190MHz=3.8GHZ and your using DDR3 1600 MHz memory with a 2:8 ratio which give you a memory speed of 1520Mhz. Increasing the MCH Strap to 1600 may allow you to reach 20x200=4.0GHz with a resulting memory speed of 1600MHz. Or you can lower the MCH Strap to a value below your present memory speed for a possible increase in memory performance.
CPU Uncore Frequency
This is the speed of the L3 cache and the Integrated Memory Controller, BLCK x Multiplier. The Multiplier should always be set at a minimum of 2X's the memory ratio or divider that you currently using. With the memory divider at 2:8 you would set the Uncore Frequency to 16X, 2:10 at 20X, 2:12 at 24X and so on. CPU-Z refers to the CPU Uncore Frequency as the NB Frequency.
All of the major components of a PC operate at frequency rates. If you look over the information above again you'll notice we have a CPU Host Frequency, CPU Uncore Frequency, Memory Frequency, QPI Frequency and the PCIE Frequency. In order for your PC to operate correctly all of these frequencies, which are all different, have to sync up in order to communicate with each other. At stock speeds the system usually has no problems doing this. But when you start overclocking and speeding up all of the various bus lines it get's harder and harder for the system to keep in-sync. The memory controller has the hardest job here, it's receiving data from the cpu (data and control info) that it has to sync up with the memory bus and send it to the memory, then it receives data from the memory that it has to sync up with whatever bus line it's going to next. The memory controller does all of this sync'ing by introducing a data delay when necessary. But at higher frequencies these delays can can get timed wrong causing an imbalance in the data flow. This is were the Clock Skew setting comes into play, by adjusting the Clock Skew you are adjusting the length, in time, of the delay.
This setting is generally left alone with air cooling as heat is the limiting factor. If your at 4GHz and 75dC at full load then adjusting your Clock Skew may get you to 4.2GHz but the extra heat from the overclock is still limiting you to 4GHz. The last time I adjusted the Clock Skew I was using a Phase-Change and my cpu was operating in the 4.6-5GHz range.
Note: The PCIE bus line is not affected by the Clock Skew, by adjusting the PCIE Frequency you are more or less sync'ing it to the system.
This is the speed of the pci-e bus, the default is 100MHz. As the BCLK is changed this value is automatically adjusted to remain as close to 100MHz as possible. But at certain BLCK settings this value can flucuate between 99-101MHz. I've found that a setting below 100 can cause instability while a value of 101-110Mhz can actually increase stability.
Intel SpeedStep / CxE Function / C-STATE / CPU TM Function
These are Power Saving Features designed to reduce the speed of the cpu to save power under various conditions. these should be disabled for overclocking.
Execute Disable Bit
When this is enabled the cpu will not execute any code residing in areas of the memory marked by the OS as non-executable. This is used to prevent certain types of malicious software from taking over computers by inserting their code into another program's data storage area and running their own code from within this section (Buffer Overflow Attack). I disable this feature, but generally speaking it shouldn't effect your overclocking.
This is a feature built into the cpu to support virtualization software like Microsoft's Virtual PC. If your not using it, disable it.
At stock speeds Turbo Mode will alter the CPU Clock Ratio / Multiplier depending upon the cpu load and temperature. When you start overclocking Turbo Mode is reduced to a simple increase of 1 to your existing CPU Clock Ratio / Multiplier.
Intel HT Technology
HT or Hyper-Threading is Intel's term for simultaneous multithreading. For each processor core that is physically present, the operating system addresses two virtual processors, and shares the workload between them when possible. Enabling this will increase your performance in multithreading applications. Disabling this feature will decrease the cpu core temperatures and may increase your overclock.
QPI Frequency Selection
This setting is actually a Multiplier, 4.800 GT/s = 36, 5.866 GT/s = 44, 6.400 GT/s = 48. The \b QPI Frequency Selection X BLCK = The QPI Link \b0 (or QPI Clock Rate). With a mild overclock or at stock speeds this Multiplier can be raised to increase the Data Bandwidth of the QPI bus. When overclocking set this to Auto or 4.800 GT/s (36).
Memory Feature / DRAM Timing Control
Memory Frequency / DRAM Frequency
This setting will be displayed as a Ratio (2:6, 2:8, 2:10, 2:12, etc...) or as a Multiplier (6, 8, 10, 12, etc...). If you are presented with a ratio then you simply take the second number and that is your Multiplier. At the default BLCK setting of 133 we see that each Multiplier corresponds to the following Speeds.
BLCK x Memory Multiplier = DRAM Frequency
133 x 6 = 798 (800 MHz)
133 x 8 = 1064 (1066 MHz)
133 x 10 = 1330 (1333 MHz)
133 x 12 = 1596 (1600 MHz)
133 x 14 = 1862 (1866 MHz)
133 x 16 = 2128 (2133 MHz)
So why do some manufacturers give us ratio's, just what is that all about? Using 2:10 as an example, for every (2 MHz of the BLCK) : the memory bus speed (increases by 10 MHz). So at the default BLCK speed of 133 MHz we have 66.5 sets of 2 MHz multiplied by 10 equaling 665 MHz.
(133 MHz/2) x 10 = DRAM Bus Speed
66.5 MHz x 10 = 665 MHz Bus Speed (This will match the memory speed shown by CPU-Z)
665 MHz x 2 = 1330 MHz (our memory is Double Data Rate, DDR, so we multiply by 2).
When you bought your memory it was rated to run at a certain speed (1333, 1600, 2000) in MHz with a pre-defined set of latencies (for example, 7-7-7-24, 8-8-8-24, 9-9-9-30). These latencies represent the following bios settings:
CL - CAS#Latency - refers to the column of physical memory in an array of capacitors ( a grid comprising of rows and columns) used in DRAM modules. The latency refers to the active amount of clock cycles that must be expended to take a request. Combined, CL sends data from the memory controller, has it read to the memory location, and output to the modules output pins.
tRCD - RAS to CAS Delay - The memory controller selects a bank, followed by a row location (using the Row Address Strobe, or RAS) and a column location using CAS. It represents the time in cycles for issuing a command and active read \\ write commands.
tRP - Row Precharge Time - Row Precharge Time represents the minimum allowable time taken between any active command and the read \ writes of the following bank on the memory module.
tRAS - Min RAS Active Time - Represents the amount of time taken between a row being accessed and deactivated. A tRAS row must be allowed to complete before being deactivated, setting this option too low can result in data corruption as the row is closed down too soon.
Depending upon your specific bios and the settings provided you may have multiple options available to adjust. Of those options we will only be looking at two (2) more of these.
tRFC - Row Refresh Cycle Time - This parameter determines the amount of cycles it takes to refresh a row in a memory bank, again if this is set too low it will cause corruption and set too high will result in a loss in available bandwidth, but increase stability.
CR - Command Rate - When the MC (Memory Controller) first tries to access memory, it has to latch onto a memory bank, known as CS (Chip Select). Then it proceeds to find the column (CAS), the Row (RAS), and then return the data to the CPU. Now, 1T means it takes 1 clock cycle to find a memory bank, vs. 2T where it takes 2 clock cycles to find the memory bank.
There was a time not that long ago when adjusting your system for maximum performance meant aquiring the lowest possible latencies, even at the expense of a few MHz in cpu speed. This is no longer true. With the memory controller built into the cpu we no longer have the bottleneck associated with the older technology. Instead we have a wider direct path between the cpu and memory controller which results in fast access to a tremendous amount of bandwidth. So, while we still want the fastest possible latencies, we no longer have to sacifice any cpu MHz's. I would rather run at higher latencies at 4.0 GHz than slightly lower latencies at 3.9 GHz.
This is the voltage to the cpu cores. With proper cooling most cpu's can easily reach 3.2 - 3.8 Ghz with little to no increase at all. At speeds above 3.6 GHz the voltage requirements can rise quickly depending upon the stepping. Older i7 920 C0 stepping cpu's often require significantly more voltage than their D0 counterparts.
Increase the CPU VCore when: BSOD error code ***STOP: 0x00000101
BSOD error code ***STOP: 0x00000124 (try raising CPU VTT 1st)
LinX produces errors and the results are very simular
LinX errors happen within the 1st minute
LinX BSOD within the 1st minute
VDroop Control / Load Line Calibration
VDroop is the flucuation in voltage that the cpu needs as it changes load states. When VDroop Control is disabled or Line Load Calibration is enabled, the CPU VCore voltage setting is a minimum value that may flucuate under load. When VDroop Control is enabled or Line Load Calibration is disabled the CPU VCore voltage is a maximum value the system trys to keep constant at idle or load. I recommend disabling VDroop Control for Air Cooling as the system will run cooler under idle conditions. I only recommend enabling VDroop Control for systems pushing maximum speeds with Water or a more Extreme Cooling Option.
CPU VTT / QPI/DRAM Voltage
This is the QPI or cpu core to memory voltage. This voltage must always stay within .5v of the VDimm voltage. If you set the VDimm voltage to 1.65v then the CPU VTT must be set to a minimum of 1.15v, at 1.70v the CPU VTT must be at a minimum of 1.20v and so on. Try to keep this voltage below 1.45v unless you have added a direct cooling solution for the Vreg/Mosfets (Vreg = Voltage Regulator).
Increase the CPU VTT when:
BSOD error code ***STOP: 0x00000124 "general hardware failure"
LinX errors happen only after 10 min or more
LinX hangs but does not BSOD
LinX reboots system without BSOD
CPU PLL VCore
This is the clock generator voltage, it keeps the CPU clock in-sync with BLCK. I've seen reports that raising this voltage may help when increasing the BLCK, cpu multiplier or if you notice the QPI Link bouncing frequently. I've actually found that reducing this voltage increases stability. The default is 1.8v and I'm running at 1.75v.
IOH PLL VCore
Input/Output Hub Phase Locked Loop voltage. I haven't found any benefit to raising this voltage. I suspect raising this voltage may help at higher BLCK settings when you notice memory or video flucuations (Northbridge related issues).
QPI PLL VCore
This is the Quick Path Interconnect Phase Locked Loop voltage, it keeps on-chip memory controller in-sync with BLCK. Try raising this voltage a little when a CPU VCore or CPU VTT adjustment does not help.
Increase the QPI PLL when:
Memory Clock speed is increased by the multiplier/ratio LinX errors without a BSOD after a few minutes
This is the main voltage to your memory.Raising this voltage may help when overclocking your memory speeds or at high BLCK values. This voltage must be kept within .5v of the CPU VTT or QPI/DRAM Voltage (see above).
This is the main Northbridge voltage which controls the connection between the CPU, Memory, GPU and PCIE 2.0 bus. Raising this voltage may be necessary at higher BLCK and CPU Uncore Frequencies. Raising this voltage can also help with SLI stability issues.
IOH/ICH I/O Voltage
This is the IOH to Intel Southbridge termination voltage. Raising this voltage can help with SLI stability issues.
This is the Southbridge voltage which connects the standard pcie bus, as in your drives, add-in cards and on-board devices to the Northbridge (see IOH VCore above). This voltage may need to be raised to keep the Southbridge in sync with an overclocked Northbridge.
This setting may be available for the CPU, VTT or both voltages. Higher frequency settings provide a cleaner or more of a straight line voltage. Let's say you've set your VCore at 1.3v, in reality this voltage may flucuate between 1.29 and 1.31 volts (as an example) but if you choose a higher PMW Frequency this flucuation may be reduced to 1.295 to 1.305 volts. Cleaner power is always better, but raising this value will cause the Vreg/Mosfet to run hotter (additional cooling maybe required).
Applying Your Knowledge
The general rule of thumb to follow when overclocking is to start slow and proceed even slower. Start by increasing your BLCK by 5MHz and note the resulting CPU Host Frequeny, CPU Uncore Frequency and Memory Frequency. Adjust any settings needed to keep all of these values within specifications. Once you have made your preliminary bios adjustments save your settings and exit the bios. Note any error screens/BSOD you may get and make the appropriate adjustments as needed. Once you've successfully booted into Windows start up your Temperature Monitoring program and LinX, then run LinX for a minimun of 20 minutes or if it fails note the error screen/BSOD info and make the appropriate bios adjustments. Once you have successfully passed LinX, if your Load temperatures are still within the safe zone (below 70dC) then you can once again increase your BLCK and repeat everything above. As your overclock increases and the error screen/BSOD increase in frequency you'll want to increase your BLCK in smaller and smaller amounts. Once you have reach your maximun overclock you'll want to do a final Stability Test. LinX is brutality hard on your system, I won't advise anyone to run it for more than 1 hour. Instead, for long duration testing, I'll recommend using Prime95 or OCCT and then run a 12 or 24 hour test.