Dear TI employees,
For a project I have to implement a piece of periodic fired software on the OMAP4460 MPSoC (Pandaboard ES) in such a way that hard real-time guarantees can be given. The piece of software will be running in parrallel with other hard and soft real-time applications and hence I need to come up with some reservation-based resource management strategy. In order to do so I need at least more information on the Arteris L3 interconnect which I was not able to find in both the TRM and on the internet. Questions regarding the interconnect are given below:
I have the following questions regarding the Arteris L3 interconnect:
* Which topology is used? E.g. a mesh network?
* Which arbitration algorithm is used? E.g. Round robin (RR) or first-come-first-serve (FCFS)?
* Which routing algorithm is used?
* How predictable are the transmission times? In other words, for each master network interface unit (master NIU) I want to be able to predict (bound) how much time it takes (minimum and maximum) to transfer a certain amount of data to a given slave NIU.
* In what extend do the different data flows influence each other? E.g. I would like to know in what extend the traffic from the A9 MPU subsystem to the L4_PER interconnect influences the traffic form the M3 MPU subsystem to the L4_PER interconnect.
* Which mechanisms from the Arteris L3 interconnect can be used in order to make certain connections between master NIUs and slave NIUs over the interconnect more predictable? E.g. is it possible to assign priorities?
Thanks in advance.
Kind regards,
Richard van Berkel.