3 ms·
It's using GraphLab 2.0, not 2.3 (no idea why), and they also forced synchronous mode on GraphLab, hence the less than stellar scaling. On at least the problem
by erichocean 12y ago
It's using GraphLab 2.0, not 2.3 (no idea why), and they also forced synchronous mode on GraphLab, hence the less than stellar scaling.
On at least the problems I use GraphLab for, synchronous mode is not only a lot slower, it converges WAY more slowly as well.
It's important when doing comparisons to not rig the game, and it's not clear to me that the authors took pains to keep things fair.
- supersonick 12y agoWe are big fans of GraphLab. We used the version that was available when we collected the performance data on GraphLab - last summer. The asynchronous engine was seg-faulting which is why we used the synchronous engine. Lots of details about this in the SIGMOD paper: http://mapgraph.io/papers/MapGraph-SIGMOD-2014.pdf http://mapgraph.io/papers/MapGraph-SIGMOD-2014.pdf And of course MapGraph is open-source, so anyone can make their own comparisons. :-)