Closed jowens closed 4 years ago
@neoblizz : do we have any data for do_a and do_b that I can use?
Can you use the older JSONs for this for now? https://github.com/gunrock/io/tree/master/gunrock-output/topc/BFS.CentOS7.2_k40cx1_do_sweep2 Once I rerun the experiments the results would be approx. the same, just the schema would be different.
We already have this though:
https://gunrock.github.io/gunrock/doc/latest/md_stats_do_ab_random
What would be different?
Well no, it's the same thing. Do you have to update the scripts to reproduce that? I'll generate updated results this week for both this issue and issue #57.
I don't see that we have to update anything, do we? It's already published, it looks fine to me ...
Old one is fine. I would love viridis
color scheme on the new one when we generate that.
If you generate new data, I'll generate a new set of graphs, and use viridis. Probably not worth it to overhaul the old scripts just to put it in the new format.
Just to get things started, one dataset is finished for 1.x.x: https://github.com/gunrock/io/tree/master/gunrock-output/v1-0-0/dobfs_parameter_sweep
Rest are still running and I will add them to the same folder under their respective dataset name.
Unassigning myself until we get more data.
https://github.com/gunrock/io/tree/master/gunrock-output/v1-0-0/dobfs_parameter_sweep more data. I am having issues with the rmat generator, so the last 2 graphs will be added later.
Thank you, these are really cool to see. These can go live on the site.
rmats are very similar, so I am only doing one of those. Closing this.
Edit: we can close this after the merge.
do_a and do_b heatmap I want to regenerate a heatmap like this: http://gunrock.github.io/gunrock/doc/latest/md_stats_do_ab_random.html with v1+ code and have it be side-by-side to the older heatmap. This requires generating some data using a simple script.