entries
Thursday, January 29, 2009- Overclocking – Theory
Overclocking depends on several factors. Cooling, voltage, architecture, and hardware limits all determine how much a device can overclock.
What is a ‘clock cycle’?
A ‘clock cycle’ is the time it takes for an instruction to be processed by a device. These devices are measured in Hz (e.g. MHz or GHz) which describe how many clock cycles are performed per second. For example, a 3GHz processor performs 3,000,000,000 clock cycles per second.
Intel CPUs
Intel CPUs are overclocked via the FSB (Front Side Bus) and CPU multiplier. The FSB controls the communication between the CPU, RAM and PCI busses. There are two types of FSB; one determined by the CPU called the ‘Rated FSB’ and the other is the ‘Bus Speed’. ‘Bus Speed’ is what is used to calculate clock speed. When you hear/read people referring to FSB, it is most likely referring to the Bus Speed. ‘Rated FSB’ (or ‘effective FSB’) is a little more confusing. Rated FSB is the name given to the result of the Bus Speed when a CPU uses double or quad pumping. Double pumping is double the Bus Speed and Quad Pumping is four times the Bus Speed.
IMPORTANT : Overclocking PCI busses is not recommended as it can easily cause system instability. Most motherboards lock the PCI bus so it cannot be altered.
The ‘multiplier‘ is a number which is multiplied by the Bus Speed to determine the CPU clock speed. The multiplier is often locked, meaning it cannot be changed. Sometimes the multiplier is partly locked, which only allows a smaller number, or completely unlocked (quite rare these days).
Let’s examine a practical example:
Let’s say the Bus Speed of a computer is 300MHz and the CPU multiplier is 11x. To determine the clock speed that the CPU is running at, we multiply the FSB by the multiplier.
11 (Multiplier) x 300 (Bus Speed) = 3300MHz (3.3Ghz)
However, Intel CPUs from Pentium 4 upwards are ‘quad pumped’. This means that the 300MHz Bus Speed is effectively running at 1200MHz. This, however, does not affect overclocking. When adjusting values we are only adjusting the Bus Speed.
You’re probably thinking right now, “Why do I need to know about the Rated FSB if I never use it?” Truth be told, you don’t. It does help, however, when looking at Intel’s specifications for their CPUs. Intel often advertises their CPU FSB speeds in Rated FSB instead of Bus Speed. So knowing how to find the Bus Speed from this data can help.
AMD CPUs
Although the technology used in AMD CPUs differ from Intel processors, the technique for overclocking is still very similar. The main difference is that there is another factor to consider called the LDT Bus. This separate bus allows communication between the CPU and the motherboard chipset.
The Technique Behind HyperTransport
HyperTransport can be considered as AMD’s version of the FSB. All AMD CPUs with HyperTransport are treated as ‘double pumped’. With Intel CPUs, the FSB is on the motherboard, but with AMD CPUs, the FSB is on the actual CPU. The Bus Speed is called the ‘HT Bus’ when working with AMD processors. There are two busses in an AMD CPU, the LDT Bus (Lightning Data Transport) which communicates with the motherboard chipset, and the HT Bus which communicates with the RAM and determines the clock speed of the CPU. The LDT Bus is sometimes known as the ‘HT Link’. The LDT Bus Speed is derived from the HT Bus and ‘HT Multiplier’. Confusing? Here are some examples:
Let’s say the HT Bus is 200MHz, the multiplier is 11x, and the HT multiplier is 5x for an AMD CPU.
To find the LDT Bus frequency we find the product of the HT multiplier and the HT Bus:
200 (HT Bus) x 5 (HT Multiplier) = 1000MHz (LDT Bus)
To find the clock speed of the CPU we find the product of the HT Bus and the CPU multiplier:
200 (HT Bus) x 11 (CPU Multiplier) = 2200MHz (2.2GHz)
As you can see, the method is similar to Intel CPUs except with the addition of the LDT Bus.
The Technique Behind the LDT Bus
The trick with the LDT Bus is to keep it as close to stock as possible. Most CPUs come with a stock HT Multiplier of 5x when the HT Bus is 200MHz (200 x 5 = 1000MHz). Essentially, you should try and keep the LDT around 1GHz as most motherboards do not allow too much overclocking of this bus. When overclocking, the HT Multiplier may have to be lowered to 4x so that the HT Bus can be raised without making the LDT Bus unstable. 1100MHz is the maximum frequency the LDT Bus should reach before dropping the HT multiplier (this can vary depending of the model of motherboard). Because the LDT Bus communicates with other busses such as PCI, SATA, USB, it is essential that they are locked to prevent instability. Most modern motherboards lock these busses automatically.
Let’s say you want to get you HT Bus to 250MHz so your CPU with 11x multiplier can reach 2.75GHz. If the HT Multiplier is set to 5x, the result will be the following:
CPU clock speed: 250 (HT Bus) x 11 (CPU Multiplier) = 2750MHz (2.75GHz) LDT Bus: 250 (HT Bus) x 5 (HT Multiplier) = 1250MHz (1.25GHz)
Notice how the LDT Bus is set too high. To fix this, just lower the HT Multiplier:
LDT Bus: 250 (HT Bus) x 4 (HT Multiplier) = 1000MHz (1GHz)
This result is much better. If you are overclocking and over the limit for the LDT Bus, yet lowering the HT Multiplier results in a frequency lower than stock, you have two options. Either do a stress test (explained later in the tutorial) with the higher HT Multiplier and check if it’s stable, or lower the HT Multiplier anyway. Underclocking the LDT Bus is not as bad as it sounds, as long as it stays close to 1GHz. Further HT Bus overclocking will bump the LDT up anyway.
For example, let’s say the same CPU is using a HT Bus of 225MHz instead:
LDT Bus: 225 (HT Bus) x 5 (HT Multiplier) = 1125MHz (1.125GHz)
This is just above the LDT Bus recommended limit, so you can either stress test it to check stability, or lower the HT multiplier:
LDT Bus: 225 (HT Bus) x 4 (HT Multiplier) = 900MHz
900MHz is still efficient enough for the CPU to run without a bottleneck, but if stable, the 5x HT Multiplier is recommended. How to determine stability will be discussed later.
`updated on- 8:19 AM
Tuesday, January 27, 2009- AMD sees loss of over $3 billion in fiscal 2008
AMD has been struggling for years now with profitability and competing with rival chipmakers. The world's second largest chipmaker has seen hard times and things still aren't looking up for the company.
AMD has posted its Q4 2008 and fiscal 2008 earnings reports, and things don’t look good. For Q4 2008 AMD reports revenue from continued operations was $1.162 billion. While that sounds like a lot of revenue, the number is down 35 percent from Q3 2008 and down 33 percent from Q4 2007.
The chipmaker reported a net loss of $1.424 billion for the quarter or $2.34 per share. The loss for continuing operations for the quarter was $1.414 billion. Q4 2008 revenue was down 28 percent sequentially with Q3 process technology license revenue of $191 million excluded.
AMD reports that for its fiscal 2008 year ending on December 27, 2008 it produced revenue of $5.808 billion and posted a massive net loss of $3.098 billion. The end of fiscal 2008 marks the second straight year where AMD posted a net loss in excess of $3 billion. The net loss in fiscal 2007 totaled $3.379 billion.
The chipmaker notes that the results for its continuing operations include a negative impact of $996 million resulting from the write-down of ATIs value, formation of The Foundry Company, restructuring charges and other items.
AMD President and CEO Dirk Meyer said in a statement, "Although industry visibility is poor, our priorities remain clear and achievable. We remain focused on further reducing our breakeven point through targeted restructuring actions while ensuring we execute our highly-competitive product and technology roadmaps. We made significant progress toward the creation of ‘The Foundry Company’ in the quarter, and anticipate closing the transaction in February. We expect our ongoing restructuring actions and asset smart strategy, combined with the strength of our innovative product offerings, will leave us well positioned for a global market recovery."
AMD's main rival, Intel, announced yesterday that it would close some of its fabrication plants and consolidate manufacturing to improve its financial outlook.
Why should I overclock?
The Need for More Performance
There comes a time when your PC no longer has the grunt to run software comfortably due to lack of performance. Older or low end PCs suffer this problem when a demanding piece of software is installed. It is important to locate what the problem is before resorting to overclocking. The problem may well be lack of RAM, which overclocking cannot fix.
To Relieve a Bottleneck
Bottlenecking is the term used to describe a component in your PC which is restricting other components due to lack of performance or efficiency. Overclocking can relieve a bottleneck if the problem lies with the CPU, RAM or GPU clock speeds. Bottlenecking is impossible to overcome completely, but it can be minimised with careful planning. Remember that a PC is as fast as its slowest component.
To Exercise a Known Safe Clock Speed
Some low to mid-end devices are underclocked versions of higher-end models. These lower-end models can then be overclocked by the end user to get similar (or possibly better) performance than the higher-end models. This is common with CPUs which use a binning process to determine what clocks CPUs get. This is great for the consumer because it allows us to spend less and then overclock to get significantly higher speeds. Some of these lower-end models have a safe known clock speed which they can perform stably on stock voltages. Voltages and stability will be discussed further on in the tutorial.
For Fun, Hobby or Competitive Reasons
Some people enjoy getting the most out of their system without spending a cent. Others spend money on complex cooling setups such as water cooling and vapour phase change units. Cooling is important when overclocking and will be discussed later in the tutorial. Overclocking competitions are becoming popular and allow skilled overclockers to showcase their talent.
To Learn More About your System
Overclocking allows you to understand how components communicate and function. It also allows you to learn what some of those jargon words mean such as DDR (double data rate).
`updated on- 6:40 AM
Monday, January 26, 2009- ok today got a pic to show u guys a simple one its the computer rig of melvin teo's new comp total spendeture 2k including screen pics will be uploaded later one for now will talk abit about overclocking
Introduction to overclocking
Overclocking is the technique used to increase the speed of a device beyond stock. ‘Stock’ is the name given to the default specifications of a device. Once you have altered the device it is no longer considered stock unless it is returned to its initial specifications. Overclocking describes the alteration of clock speed above stock (hence the name overclocking) of any given device, usually RAM and CPUs, but also other components such as GPUs and PCI busses. When a devices clock speed is decreased from stock (or from an overclock), it is called underclocking.
NOTE: Not all hardware can be overclocked. Research the components you wish to overclock using a search engine such as Google, and check if they are capable of being overclocked. Searching the model number of your motherboard along with "overclock" should help you determine whether the system is capable.
DANGER of overclocking
Overclocking is safe. There, I said it. I will not attempt to scare you away from overclocking. I believe that if you learn properly you can overclock without trouble, as long as you know your limits. This tutorial explains how to determine these limits. I am not liable if you damage anything by following this tutorial. Overclock at your own risk. Please note that overclocking can decrease the lifespan of your components.
Overclocking and Warrantee
Before you leap into the world of overclocking, check that you aren’t voiding your warrantee. Most devices will forfeit your warrantee when overclocking is performed, so the main rule of thumb is to not overclock anything you cannot afford to replace. Even though overclocking is fairly safe to perform, it is not very smart to risk breaking something if you cannot replace it.
What Determines CPU Performance?
The performance of a CPU is determined by its frequency (or clock speed), architecture and cache. The architecture of CPUs is constantly evolving, and with this we get faster processors that run cooler and more efficiently. Multiple core CPUs are a prime example of architecture improvement. Two or more cores allow data to be processed more efficiently therefore increasing performance, sometimes even with smaller clock speeds. The performance difference between single and multiple core CPUs is dependant on the software and OS you are running. With overclocking we can’t change architecture because it is a physical component. The only way to alter architecture is to buy a new CPU altogether. Clock speed is the rate in which data is processed. It is common for people to get confused with this idea and assume that a bigger clock speed CPU runs faster than a lower clocked CPU. This is not necessarily true, especially when there is a major architecture difference. An interesting comparison of CPUs at the same clock speed but with different architectures can be found here.If you were to compare two identical CPU models, each with different clock speeds, the higher clocked CPU would be the faster processor.
so to simply round it up if u are super lazy to read what i can summarise its simple just the effects of overclocking, the results of overclocking, warranty, and the danger of overclocking.
`updated on- 9:39 AM
Saturday, January 24, 2009- haha ok start this web for mostly any or everyne who visits my blog http://foreveraliveunabletodie.blogspot.com and is wodnering to knwo any or everything to know about COMPUTERS.
at here shaun's personal blog for computers i will update mostly any or everything i wan or like to talk about regarding any and everything about computer products.
STAY TUNE. ^^
`updated on- 9:20 AM
