[Hadstruct] Minutes of 7/21/2020

[David Richards]

LaMET

There was a lot of discussion regarding the best strategy for submitting abstracts, given there are essentially only 16 slots. The feeling was that we should not constrain ourselves, submit what we are doing, and then leave the organizers to constrain the talk. We noted that LaMET was now in the workshop title, which may be indicative of the tone. We would not submit a review talk, since they are well aware of the approach. Please submit abstracts to the group by the next meeting.

<https://github.com/JeffersonLab/HadStruc/wiki/Minutes-of-21st-July,-2020#computational-time>Computational Time:

FW has implemented the hybrid 3D/4D code - this has been tested on the Intel MKL library, but GPUs have not yet been tested.

  *   NERSC

KO will test the hybrid implementation.

  *   Frontera

CE and ZS has been running the GPDs, using the same 16 interpolating operators using "Redstar", using the 32^3 lattices. We've used around 50% of our allocation.

  *   USQCD

KNL/JLab. CE and DR will work with ER to implement the momentum smearing on the 48^4 lattices, using, say, 1 or 2 boosting factors, with the aim of reaching momenta of the order of ap around 8. It would be useful to know precisely how many lattices of the a091m170 ensemble there are. BJ will look at ORNL.

SKX. This would be a good place for two-point functions for gluon distributions where we have very high statistics.

  *   ALCC

FW is running Anisotropic (in z) gauge generation. Also matrix-element time.

  *   Paris

SZ running 48^3 physical point for PDFs with distillation, set up by CK. Running in earnest since Friday, once chroma modified to exploit omnipath rather than infiniband. Total have 250K node hours - trying to get more once used up.

  *   (XSEDE)

DR submitted request on Bridges, and a small amount on Stampede, for the gluon distribution.

  *   (INCITE)

A set of computational-readiness questions had been received, and CM had set up a document in the overleaf page.

<https://github.com/JeffersonLab/HadStruc/wiki/Minutes-of-21st-July,-2020#aob>AOB

  *   RS asked about gluon distributions, in particular the polarized distributions.

  *   LOI for Snowmass, in preparation for the upcoming P5 report (HEP equivalent to NSAC long-range plan). Kronfeld had circulated an email following extensive discussion within USQCD. CM is an active participant. In general, the more LOI the better, and in particular we should try to involve ourselves in several "frontiers", such as theory, computation, PDFs.

  *   Nikhil Karthik will be joining the in the Fall, and would like to start collaborating as soon as possible. There was anonymous approval.

<https://github.com/JeffersonLab/HadStruc/wiki/Minutes-of-21st-July,-2020#next-meeting>Next Meeting
Tuesday 28th at 9am. Agenda: status of analysis.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/hadstruct/attachments/20200721/0ed59a90/attachment-0001.html>