2.8.4. How the mapfile memory model calculates wait states

The memory map file specifies access times in nanoseconds for nonsequential/sequential reads/writes to various regions of memory. By inserting wait states, the map memory model ensures that every access from the ARM processor takes at least that long.

The number of wait states inserted is the least number required to take the total access time over the number of nanoseconds specified in the memory map file. Consider this when designing your system.

For example, with a clock speed of 33MHz (a period of 30ns), an access specified to take 70ns in a memory map file results in two wait states being inserted, to lengthen the access to 90ns.

If the access time is 60ns (only 14% faster) the model inserts only one wait state (33% quicker).

A mismatch between processor clock-speed and memory map file can sometimes lead to faster processor speeds having worse performance. For example, a 100MHz processor (10ns period) takes five wait states to access 60ns memory (a total access time of 60ns). At 110MHz, the map memory model must insert six wait states (a total access time of 63ns). So the 100MHz-processor system is faster than the 110MHz processor. (This does not apply to cached processors, where the 110MHz processor would be faster.)

Note

For accurate simulation of the real hardware, access times specified in the memory map file must include propagation delays and memory controller decode time as well as the access time of the memory devices. For example, for 70ns RAM, if there is a 10ns propagation delay, configure the map file as 80ns.

Copyright © 1999-2001 ARM Limited. All rights reserved.ARM DUI0058D
Non-Confidential