[Hadstruct] Meeting Today

[Eloy Romero]

Hi folks, I won't be able to make it this time most likely. The runs at frontier are paused because we are consuming too much IO. I'm in conversation with the admins to see where the low bar is.

We know the final solution for this, and it's part of scidac. We are close, only two milestones ahead:
1) improve the performance of multigpu for easy to distribute contracting graphs (jie is working on this)
2) efficient partitioning of big graphs. Aydin buluç 's group is working on this but there's no algorithm fast enough for us. Although I believe we can hack our way out, at least for Kostas contractions, they are simple.

I believe 1) should be mostly in its way by the end of this week or next week, at least a usable version. And I hope we can go far with what we currently have for 2), but we'll see...


On September 18, 2023 4:32:05 PM GMT+02:00, Joseph Karpie via Hadstruct <hadstruct at jlab.org> wrote:
>Hey everyone,
>
>We will have our usual meeting at our usual time and place ( 230pm in F227 )
>
>Here is the link: https://jlab-org.zoomgov.com/j/1602975452?pwd=STN0b2hydnpzcE0xZWtpa2RwOGhsdz09<https://urldefense.proofpoint.com/v2/url?u=https-3A__jlab-2Dorg.zoomgov.com_j_1602975452-3Fpwd-3DSTN0b2hydnpzcE0xZWtpa2RwOGhsdz09&d=DwMFAg&c=CJqEzB1piLOyyvZjb8YUQw&r=9LwgpSkTsQPSpuio90w6T_IJLuZSEdqHpu00-KQ9Ojg&m=sDeVVStUMefq5p_6-wTjzKNV4BZbbI-HuMjWC39k7JhcMn4gaEr_uLJnaDmVp6Zb&s=I_NJ2YlAR9QHpplVtW9-WlHXkqtyAKBtxzLljktwEKQ&e=>
>ID: 160 297 5452
>Passcode: 163101
>
>Best,
>Joe
>
>--
>Joe Karpie
>Pronouns: He/His/Him
>Postdoctoral Fellow
>Theoretical and Computational Physics Center
>Jefferson Lab
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/hadstruct/attachments/20230918/355f4bd7/attachment.html>