| 摘要 |
<p>When the processor of a vector computer detects an optimization instructing message which instructs the optimization of the method for performing access to a list at compiling time, the processor automatically changes the method for accessing. The bank contention associated with vectorization can be avoided and the access speed to the list can be accelerated by, for example, copying the data to be accessed in a working array, randomly changing the accessing order, or performing scalar loading.</p> |
