LPC2103 vs. LPC2104 - MAM Bug?


I am experiencing a weird problem when running the same code on the two chips running at the same speed. 

  • with MAM mode = 0, both cycle counts are identical (13680 cycles)
  • with MAM mode = 2, the LPC2103 is slightly slower than the LPC2104 (4892 vs. 4820 cycles), this is expected behavior because the latter uses two (instead of just one) Flash banks and has two buffers for prefetch and branches 
  • with MAM mode = 1, the LPC2103 is massively faster than the LPC2104 (5793 vs. 9472 cycles). I would have expected that instead the LPC2104 is slightly faster, just like in mode 2.

The description of MAM mode 1 is identical for both. They both buffer only sequential code access. 

So it looks like either the LPC2103 or the LPC2104 have a bug in MAM mode 1. Or that for one of these, MAM mode 1 works differently.

Is that a known problem? Does there maybe an updated documentation exist or an errata sheet which explain what I am experiencing?

More questions in this forum
There are no posts to show. This could be because there are no posts in this forum or due to a filter.