3 ms·
The experiments were conducted by invoking spark with flags --master local[1] This means running on a single core. Not really a fair comparison with the Julia
by pathsjs 10y ago
The experiments were conducted by invoking spark with flags --master local[1]
This means running on a single core. Not really a fair comparison with the Julia multithreaded or multiprocess version
- ViralBShah 10y agoI am pretty sure that is a typo in the blog post, since the performance of Spark improves as more cores are used, as does Julia's and they both show similar scaling characteristics. The performance plot would probably be more readable if the y axis was on log scale. The typo is also now fixed.