We didn't have time to analyze the performance through the lens of Amdahl's Law but this should be directly applicable because we have constant costs (for each worker) that limit the over performance of each worker.
Mentioned both in the paper. If we had more time we should have calculated the Karp-Flatt metric [1] (serial fraction as a function of core number), too.
A. H. Karp and H. P. Flatt. Measuring parallel processor performance. Commun. ACM, 33(5):539–543, May 1990. DOI: 10.1145/78607.78614
We didn't have time to analyze the performance through the lens of Amdahl's Law but this should be directly applicable because we have constant costs (for each worker) that limit the over performance of each worker.
Also see Gustafson-Barsis's Law.