| 摘要 |
In the case of dynamic memories, the stored information must be refreshed at the latest after certain time intervals, as otherwise a loss of information can occur. Since this refreshing comprises reading and re-writing a line of memory, additional time is required for this. The aim of the invention is to minimise these lost times. This is achieved by departing from previous concepts and detecting separately for each memory line whether it is imminent that the refresh time will be exceeded (Fig. 2). The respective memory line is refreshed again only just before this time elapses. Consequently, the minimum of refresh cycles for the memory is achieved. It is advantageous in this case that whenever the memory is addressed frequently, additional refresh cycles can be virtually dispensed with. <IMAGE>
|
