Adapted from an article I found by Googlin'. There are many ways to define this, but simply put:
- Task parallelism is the simultaneous execution on multiple cores of many different functions across the same or different datasets.
- Data parallelism (aka SIMD) is the simultaneous execution on multiple cores of the same function across the elements of a dataset.