But then many aspects to SMP

But then many aspects to SMP

* Decomposition of a single task
* This is where you take a single algorithm and parallelize it to share its workload between the
multiple processors.
* Execution of multiple different algorithms concurrently
* This looks more at how the bigger system, including the OS, handles the workloads from
multiple concurrent algorithms.
* Execution of a single algorithm over multiple data sets
* This looks again at the bigger system, but concentrates more on the data throughput, and how
a system can handle multiple 'channels' of data.

Conclusion, We’d apply these three SMP techniques to individual and groups of existing EEMBC consumer benchmarks