I am wondering if building a logical model in Groovy is the right way to go about this.
Perhaps what I should do is build a “distributed malloc” – in other words something that manages memory allocation from a pool that resembles the “memory map” of a NoC system.
It would be a lot simpler to test I think – though this is just me thinking as I type this – we could have sixteen threads of execution, calling malloc(thread_number, block_size) and then test if the computation was handled correctly.
Sounds too simple to be true to me though.
Advertisements
