Rubber buffer for Honda in the design of the CPU, the DC load capability of the general output line can drive a TTL load. In the connection, an address line or data line of the CPU may be connected to multiple memory chips, but the memory chips are The MOS circuit is mainly a capacitive load, and the DC load is much smaller than the TTL load. Therefore, in a small system, the CPU can be directly connected to the memory, and a buffer is required in a large system.
Honda rubber buffer any program or data to be used by the CPU, must first be placed in the main memory (memory), that is, the CPU only exchanges data with the main memory, so the speed of the main memory largely determines the operating speed of the system. . During a short period of time during a program, the address generated by the program tends to be concentrated in a small range of address space in the memory. The instruction addresses are originally distributed continuously, and the cyclic block and sub-blocks are repeatedly executed multiple times, so access to the contents of these addresses naturally has a tendency to be concentrated in time.
The concentration of the data distribution of the Honda rubber buffer is not as obvious as the program, but the storage and access of the array and the selection of the work unit can make the memory address relatively concentrated. This phenomenon of frequent access to localized memory addresses, and the lack of access to addresses outside of this range is known as the localized nature of program access. From this nature, it can be seen that the information set accessed in this local range changes slowly with time. If the information that is frequently accessed in a certain address range is collected in batches, the information can be read from the main memory in batches. The high-speed access to the small-capacity memory is stored for the program to be used at any time during the period to reduce or no longer access the slower main memory, which can speed up the program.