Page 48 - Computer_Science_F5
P. 48

When  the  processor  requests  data,  the  two adjacent levels at a time. A block
              cache controller first checks the cache.  is the fixed-size chunk of data that the
              If the data is found, a cache hit occurs,  cache transfers between itself and other
              and the data is quickly delivered to the  memory levels.                             Chapter One: Computer architecture
              processor. If not, a cache  miss occurs,
          FOR ONLINE READING ONLY
              and the cache controller then checks the  Cache line: A unit of data storage within
              main memory. If the data is found in main  the  cache. It acts  like  a container  that
              memory, a main memory hit occurs, and  holds a block and is specifically designed
              the data is transferred to the cache and  for  efficient  data  transfer  between  the
              then to the processor. If the data is not  cache and other memory levels.
              in the  main  memory, a main  memory
              miss occurs. The data is retrieved from  Cache  hit: An  event  in  which  specific
              storage, transferred to the main memory,  data or instructions are searched and
              then to the cache, and finally delivered  found available  in  the  accessed cache
              to the processor.                        memory. In this scenario,  the  cache
                                                       transfers (copies) the requested data to
                                        Cache          the processor.
                              CPU       control
                                         logic
                                                       Cache miss: Event in which specific data
                                                       or instructions is searched and not found

                 Block                                 in the  accessed  cache  memory. In this
                                                       scenario, the data is fetched from lower
                                                       levels and transferred to the processor.

                   Block is                            Hit rate: the percentage of times the data
                  transferred
                                                       is found in the cache without needing to
                                                       access lower levels.  A higher hit rate
                                                       indicates efficient data retrieval.

                                                                            Totalhits
                                                           Hitrate =
                                                                     Totalhits+Totalmisses
                Figure 1.39: Memory interfacing between
                           cache memory and main       Miss rate:  the  percentage  of  times  the
                           memory                      data is not found in the cache and needs
                                                       to be fetched from lower levels. A lower
              Memory hierarchy and interfacing  are    miss rate is desirable.
              enabled by the following key concepts;
              Unit of information transfer: Memory        Missrate =       Totalmisses
              is  organised  into  fixed-size  blocks.  A            Totalhits+Totalmisses
              memory hierarchy can consist of multiple   Miss penalty: the extra time it takes to
              levels, but data is copied between only   access data from a lower level compared


                                                     39
               Student’s Book  Form Five



     Computer Science Form 5.indd   39                                                      23/07/2024   12:32
   43   44   45   46   47   48   49   50   51   52   53