Hi Guys,
Iam using IAR with LPC 2142 core. I had downloaded the code to RAM and executed the code. I had also executed the same code after downloading it into the flash. The flash code executed was very slow, something five times slower that code run from the RAM. I have not enabled MAM. What could be the problem. Does Code from flash run so slow?
robinjacob:
Hi Guys,Iam using IAR with LPC 2142 core. I had downloaded the code to RAM and executed the code. I had also executed the same code after downloading it into the flash. The flash code executed was very slow, something five times slower that code run from the RAM. I have not enabled MAM. What could be the problem. Does Code from flash run so slow?
Try enabling MAM.
Leon
Hi,
I enabled MAM. The core is running at 60 Mhz. Now the speeds are nearly equal. Can somebody say at what MIPS a LPC 2142 can work, if MAM is set at 3 and CCLK is 60 Mhz?
code from flash will always run slower than in RAM
there is no 60 mhz flash (17 ns), Philips make accelerating the process by reading 128 bit (4 words) at once but if you code have branches and is not linear execution this advantage will not be seen
Tsvetan
Hi,
I understand that the code from RAM runs faster than flash. But there should be some value of the number of instruction that it can execute in a sec. How can we calculate that value. This value will help us to write a optimized code.
this is ARM with 3 stage pipeline, not PIC 
there is no fixed instruction execution time on this microcontroller i.e. in PIC you can assume X instructions x 4 us per instruction = this time delay
here you use timers and nothing else as time reference
Tsvetan