AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |
Back to Blog
![]() If the data in the cache line is used only once before it is evicted, performance will be poor. Try to place frequently accessed data sequentially in memory, remembering that the cost of fetching a new data value from main memory requires a cache line fill. ![]() It is best to configure them as noncached and nonbuffered memory, which forces the processor to read the peripheral device on every memory access, rather than use what would be stale data from cache. Memory-mapped peripherals frequently fail if they are configured to use cache or the write buffer. For more information on regions and the configuration of cache and write buffers operation within them, refer to Chapter 13 if you are using an ARM processor core with a memory protection unit and Chapter 14 if you are using an ARM processor core with a memory management unit. Most regions in a memory system are configured to have both the caches and write buffer enabled, taking maximum advantage of the cache architecture to reduce average memory access time. ![]() Here are a few simple rules to help write code that takes advantage of cache architecture. CHRIS WRIGHT, in ARM System Developer's Guide, 2004 12.7 CACHES AND SOFTWARE PERFORMANCE
0 Comments
Read More
Leave a Reply. |