A. Reading order
CPU to read a data from the Cache to find, if found, immediately read and sent to the CPU processing; If you do not find, with a relatively slow speed read from the memory and CPU processing, while this data where the data block transferred to the Cache, you can make vgp-ac19v25 to read the entire data from the Cache, without having to call to memory.
Read Cache hit rate is very high (most of the CPU up to 90%), means that the CPU to read the next data is read mechanism allows the CPU to 90% in the Cache, only about 10% need to be read from memory. This saves CPU time to read the memory directly, and also allows the CPU to read data without waiting for the basic. Overall, the CPU reads the data sequence is first Cache memory.
Two. Cache classification
The front of the Cache as a whole to consider the classification analysis. Intel from the Pentium Cache separate, usually divided into a cache L1 and two cache L2.
In the concept of the past, L1 Cache the Cache integrated on the CPU, known as the chip. , Also in the L1 Data Cache (I-Cache) and the Instruction Cache (D-Cache). They are used to store data and the implementation of these data, instructions, and two Cache can be accessed by the CPU to reduce the conflict caused by contention Cache to improve processor performance.
P4 processors use an advanced level instruction Cache – dynamic tracking cache. Sony Vaio computer batteries is directly connected to the execution units and dynamic tracking engine, dynamic tracking engine can quickly find the implementation of instruction and the instruction sequence is stored in the track buffer, thus reducing the decoding cycle of the main execution loop to improve the processing computational efficiency of the device.
Previous L2 Cache not integrated in the CPU on the motherboard or CPU integrated with a circuit board, also known as off-chip Cache. However, P III began the process to improve L2 Cache is integrated in the CPU core, clocked at the same rate, ending the history of L2 cache and CPU gap frequency to equal the performance of L2 cache and L1 Cache higher transmission speed. L2Cache only store data, regardless of data Cache and instruction Cache. CPU core does not change the circumstances, to increase the capacity of L2 Cache can make performance upgrade, the same core CPU high and low end points is also often play tricks on the L2 Cache, showing the importance of the L2 cache. CPU L1 Cache and L2 Cache only difference is that the reading order.
3. Read hit rate
CPU in the Cache find useful data hit, this time called when there is no CPU required for data in the Cache (miss), the CPU to access memory. In theory, a 2 Cache CPU, read L1 Cache hit rate of 80%. In other words, the CPU L1 Cache find useful data accounted for 80% of the total data, the remaining 20 percent is read from the L2 Cache. Can not accurately predict the data to be performed, the read L2 hit rate is also 80% (from L2 read useful data for the total data, 16%). Some data had to call from memory, but this is a very small proportion. In some high-end areas of the CPU (like the Intel Itanium), we often hear L3 Cache, it is the design for the data read L2 Cache misses – the Cache in the CPU with L3 Cache, only about 5% of the data needs to be called from memory, which further improves the efficiency of the CPU.