This is a new gadgeteer mainboard from Micromint. It uses the LPC43xx chip, which includes dual cores, an M4 for NETMF and M0 for possible native work. The company lists this as open-source so we look forward to messing with the code.
I like the master slave approach where some tasks are delegated to the other core and some even dynamically during compile. Downside for me, CAN is used in the demo but not available in the first release…
So based on this screen grab at 3:53, are they saying the firmware is taking care of the division of tasks across the CPUs (in .NETMF at least, looks like with mBed you can choose which CPU to execute a specific task)?
The way I interpret it (with my lack of h/w & firmware understanding) is:
I’m wondering what the advantage of such a system would be… wouldn’t it be simpler and just as fast to have the M4 do the ADC read? Is it doing the read asynchronously?
@ mhectorgato - I of course have no clue, but my guess would be. NETMF actually runs like normal on M4 and you can write native code for the M0 and just kick that off from the NETMF side. Analog is done by a sepearate peripheral any way, so I see no real win there, but something like SignalGenerator which is bit banging might be an ideal fit, since it would be using core CPU resources at the moment but can be off loaded to the M0. I do this with the DL40 at the moment for example so it is an appealing option to me.
hmmm…i think i’d much rather a multicore where code can be paralleled across multiple cpus. doesn’t sound like that is the case here. this seems as though its basically bringing two controllers into one board.