Baidu know
Baike cardcache is part of the CPU, it exists in the CPU accesses data in the CPU speed is very fast, one second to access,[url=http://www.style-jackets.com/]Moncler Jackets[/url], process one billion instructions and data (terms: CPU frequency 1G), and memory is much slower, faster memory can reach tens of megabytes on the good, the speed difference between the two can be seen how big the cache is to solve the speed of the CPU speed and memory speed difference issue
directory
About precautions because of physical memory in the cache memory is only a small number of copies of the data was the most frequently accessed data is not static cache and secondary cache on a slower CPU frequency profile faster CPU frequency if the CPU compared to a furniture factory in town, and the storage system than for the suburban wood factory, then the reality is that timber plants farther away from the furniture factory, even with a larger truck to transport wood, furniture factories have to lay-off to wait for the wood sent. In this case, one solution is to set up a small warehouse in the city, in which to place some of the most commonly used wood furniture factory. This is actually a furniture factory warehouse Of course, the larger the warehouse, storage timber, the more the better, because even some things not commonly found in the warehouse. If we do not need the timber warehouse, outside of the timber from the factory will continue to find, but have to wait for a furniture factory. I think now we have to understand what I mean, as opposed to the warehouse on the L1 cache, the CPU time can be quickly read and write, so the storage is the most common CPU code and data (we will explain how to choose the . L1 cache is faster than system memory is much faster because it uses SRAM, the memory element with four to six single-crystal transistors. It also makes very high cost of SRAM, it can not be used with the entire storage system. In the majority of the CPU, L1 cache and core together on a single chip. If the furniture factory in our case, like factories and warehouses on the same street. This design allows the CPU to the fastest place to get from the most recent data, but also makes the This data is not needed if the CPU L1 cache, that is, The faster the processor, the greater the gap between the two. If you use that kind of high-frequency Pentium4 processor, to obtain data from memory is equivalent to Physical memory that is memory memory the most frequently accessed by the CPU data and instructions are copied into the CPU cache, this CPU can not often go to as pick on the line, and the cache memory is much faster than the speed of this note is to be noted: because of the cache memory is only a small replica of a part of the data, because the cache is only a small part of the memory copy of data, so the CPU to find data to the cache, there will not find the case (because the data is not copied into the cache from the memory to go), then the CPU will still find the data into memory, so the system would slow down the speed, However, the data copied into the CPU cache will go to the next not to the memory fetch. The most frequently accessed data is not static 2,[url=http://www.style-jackets.com/]Moncler Jackets[/url], because over time, the most frequently accessed data is not static, that is, the data just is not frequent at this time has been frequently need access to or just the most frequent data, we not frequent, so that cached data should always be replaced in accordance with certain algorithms, so as to ensure that cached data is accessed most frequently and on a secondary cache to cache distinguish between these two concepts, we first look at the relative RAM ram and ROM, RAM is the power-down after which it disappeared what kind of information, ROM at power-down after that kind of information does not disappear RAM was divided into two One is static RAM, SRAM; one is dynamic RAM, DRAM. The latter the former is much faster than memory speed, we generally use dynamic memory RAM. Some say a rookie, in order to increase system speed, the cache is not on line yet to expand,[url=http://www.style-jackets.com/]Moncler Jackets[/url], the greater the expansion, more data cache, the faster the system does not do caching are usually static RAM, the speed is very fast, but the static RAM integrated, low (store the same data, static RAM, dynamic RAM size is 6 times), the price is high (with the capacity of dynamic RAM, static RAM is four times),[url=http://www.style-jackets.com/]Moncler Jackets[/url], shows that the expansion as a static RAM Cache is a very foolish act,[url=http://www.style-jackets.com/]Moncler Jackets[/url], but in order to improve system performance and speed, we must expand the cache, so there is a compromise approach, do not expand the original static RAM cache, but the increase as the number of high-speed dynamic RAM cache, these high-speed dynamic RAM speeds faster than conventional dynamic RAM, but the static RAM cache than the original slow, we had called a static ram buffer cache, while the later increase in dynamic RAM called cache memory. Level cache and secondary cache memory contents are accessible copies of high frequency data (maps), they exist to reduce the high-speed CPU for slow memory access. CPU find data or instructions are usually the order is: first-level cache to find, and then find the secondary cache to find, if you can not find a memory only to find extended reading: 1 Open Category: CPU, computer hardware I have to improve the Related articles：
没有评论:
发表评论