This seems like the highest-level abstraction we have seen in this class so far. The programmer does not care about the hardware (cores/SIMD/GPUs, distributed computing clusters) or parallelism/synchronization (like worrying about independent loop iterations). The programmer specifies what samples to use, what functions to apply, and when to stop the program (that's what the untilconverged call does). It uses the map and groupby functions we introduced in the parallel-thinking lecture as well.
This seems like the highest-level abstraction we have seen in this class so far. The programmer does not care about the hardware (cores/SIMD/GPUs, distributed computing clusters) or parallelism/synchronization (like worrying about independent loop iterations). The programmer specifies what samples to use, what functions to apply, and when to stop the program (that's what the untilconverged call does). It uses the map and groupby functions we introduced in the parallel-thinking lecture as well.