| 摘要 |
Data to be retrieved from a non-volatile storage medium is identified by considering previously stored cache data in a cache buffer is disclosed. When a request for data is received from a host, the requested data is ear-marked for retrieval, with a desired amount of pre-read data and a desired amount of pre-fetch or read-ahead data. It is then determined whether any of the data is already in the cache buffer, and how much of a gap, if any, exists between the desired data and that already stored in the cache buffer. If some of the data is already in cache, then a controller has to satisfy the read request with a single continuous retrieval of data from the storage medium, such that the data as stored in the cache buffer is not interrupted by gaps within the cache buffer. In this manner, buffer usage rate is increased while maintaining the same cache hit rate in a particular access pattern to evaluate hard disk drive performance. In another aspect of the invention, the amount of pre-read and pre-fetch data can also be adjusted to decrease the number of times the read head misses the data read start point and is forced to wait for the disk to rotate a revolution to begin reading.
|
