Considerations for cainiao becoming masters

Source: Tianji Forum

Statement:

1. overclocking is dangerous and may result in hardware decommission.

2. overclocking hardware reduces short service life

Super-frequency can improve the performance of the pc without spending a penny. It is also a means for computer enthusiasts to increase their understanding of hardware. Super-frequency is not just like a game. You can just click it to finish it. Super-frequency requires a systematic understanding of hardware knowledge, and it cannot be done by people without patience and basic knowledge.

If you want to exceed the frequency of a friend to have a certain understanding of BIOS, at least need to know what written in, or reference manual.

Basic knowledge:

CPU frequency = multiplier x external frequency (FSB)

For example, 2.0G (CPU frequency) = 200 MHz (external frequency) x10 (Frequency Doubling). Here we can set 1G to 1000 MHz.

Explain this formula:

FSB is the front-end bus. It can be understood that FSB is like a road leading to a city (CPU) in various villages (computer components). A multiplier can be understood as the number of FSB lines, therefore, we come to the conclusion that increasing the number of road lines or the width of the road can increase the traffic.

So: 250 MHz x 8 = 2G, or: 100 MHz x 20 = 2G
But: Obviously, the width of the road is better than that of more lines. The same is true in PCs. Generally, the latter formula is not used. Reducing FSB will greatly affect the performance.

However, FSB is inaccurate on some AMD CPUs, but the theory is the same. It should be noted that most CPU frequencies are locked and cannot be decreased or increased.

Some people say: I understand that FSB is finished, and that's why it is too unsuccessful. But in fact, many times the channel of the entire system is widened when FSB is increased, but the memory, AGP, and PCI systems cannot afford this load, resulting in overclocking failure.

How to solve this problem:

At this time, a good BIOS system has played a role, and Intel's original motherboard has almost no options for overclocking settings; the KT266, 333, 400, 600, 880, and K8T800 of the via series are not locked for PCI and AGP (except for K8T800pro), which brings some trouble to the overclock, almost all of the NF series are locked.

The most important part is the memory:

When the formula 250 MHz x 8 = 2 GB is used, your DDR400 memory is changed to DDR500. If the memory is unbearable at this time, the overclock will fail.

How can this problem be solved?

In the BIOS overclocking settings, there will be a crossover setting, that is, FSB: RAM, and the ratio is set, such as, and so on. Note that when using this ratio, there will be some performance loss. When a small amount of super-frequency is used, good memory can be tolerated, so there is no need to use a crossover. Therefore, the good memory is much more expensive than the miscellaneous card memory.

Voltage:

When everything is ready, the limit will be reached, and the voltage will be increased (this is also the root cause of short life reduction, the impact of small pressurization is very small ). Super frequency of any component is limited by voltage. Voltage addition is effective but dangerous. Never greedy. When you smell it, everything is too late.

Heat Dissipation problems:

This problem is often the first cause of failure. Therefore, make sure that your cooling system has sufficient capacity before overclock. Or change a stronger heat dissipation system. Currently, the price of water-cooled equipment is not very high.

Stable operation is the key:

Without stability, everything is white. I heard that the super-frequency madman has surpassed a CPU to 5 GB, but I don't believe it is stable. The general stability refers to the continuous operation for 24 hours.

TIPS: Why is it unstable to make the PCI/AGP bus run out of specifications?

The major cause of instability caused by over-capacity PCI bus is that it forces components with very strict allowable deviations to run at different frequencies. The PCI specification is usually set to 33 MHz. Sometimes it requires 33.3MHz, and I believe it is close to the real specification. The primary victim of high PCI speed is the hard disk controller. Some controller cards have higher allowable deviations than other cards, so they can run at increased speeds without obvious damage.

However, the onboard controllers (especially SATA controllers) on most boards are extremely sensitive to high PCI speeds. If the PCI bus runs at 35 MHz, there will be damage and data loss. Most of them can cope with 34 MHz. In fact, the super specification is smaller than 1 MHz (depending on how the motherboard is rounded to 34 MHz ...... For example, most boards may report a 34 MHz PCI speed at any FSB between 134 and MHz. The actual range is from 33.5MHz to 34.25 MHz, and the change may be larger based on changes in the clock frequency of the motherboard.

In a higher FSB and a higher divider, the range may be larger ). Sound cards and other integrated peripheral devices are also damaged when the PCI bus exceeds the specifications. The ATI graphics card has a much lower allowable deviation (directly related to PCI speed) on a high AGP speed than nVidia cards ). Remember, most Realtek LAN cards (based on PCI and occupying expansion slots) are set to run securely at any frequency from 30 to 40 MHz.