US ATLAS HL-LHC computing discussion

["Malik,Abid" <amalik AT bnl.gov>]

Dear All,

Because of SC18, there will be no meeting today. I apologize for the late 
update.


Thanks,

Abid M. Malik
Computational Science Initiative (CSI)
Bld-725, 2-131
Phone#6313444657

________________________________________
From: Malik,Abid
Sent: Tuesday, November 6, 2018 10:42 AM
To: usatlas-hllhc-computing-l AT lists.bnl.gov
Cc: Dan Guest
Subject: Re: [Usatlas-hllhc-computing-l] Working Group meeting of the 
distributed training working group

Hello,

As decided in the last meeting, we will have biweekly meetings for this group.
Our next meeting is on Nov. 13th.

Thanks,

Abid M. Malik
Computational Science Initiative (CSI)
Bld-725, 2-131
Phone#6313444657

________________________________________
From: Malik,Abid
Sent: Monday, October 29, 2018 12:14 PM
To: Amir Farbin
Cc: Torre Wenaus; usatlas-hllhc-computing-l AT lists.bnl.gov; Dan Guest
Subject: Re: [Usatlas-hllhc-computing-l] Working Group meeting of the 
distributed training working group

Thanks, Amir.

Just a gentle reminder, we have a meeting tomorrow at 12:30 PM (EST). Blue 
Jean information is given below:

To join the meeting on a computer or mobile phone: 
https://bluejeans.com/984534274?src=calendarLink

Agenda is avaliable on the google doc.

https://docs.google.com/document/d/1w5CXzODmv5Z9HoEpmORTi8-sh6vChkLX4KCajCoFX-s/edit#heading=h.lsmlha7nuv1h


Thanks,

Abid M. Malik
Computational Science Initiative (CSI)
Bld-725, 2-131
Phone#6313444657

________________________________________
From: Amir Farbin <afarbin AT gmail.com> on behalf of Amir Farbin 
<afarbin AT uta.edu>
Sent: Monday, October 29, 2018 9:51 AM
To: Malik,Abid
Cc: Torre Wenaus; usatlas-hllhc-computing-l AT lists.bnl.gov; Dan Guest
Subject: Re: [Usatlas-hllhc-computing-l] Working Group meeting of the 
distributed training working group

Hi,

I’ve taken over as ATLAS ML coordinator and we just had a discussion about 
priorities. It appears that we have quiet a few people (mostly students) who 
are tasked with training and optimizing various DNNs that are being broadly 
used, for example in the flavor tagging group. We would like to work out how 
such training jobs can be submitted on PANDA to run on GPU resources. Dan 
Guest (the other coordinator) is going to identify one such person with 
enough skills to be the first person to try PANDA training submission. I’m 
not sure how ready PANDA is for such a task.

I would like to suggest that we discuss this topic at tomorrow’s meeting. If 
possible, can we make sure at least person is there to be able to answer 
questions about what needs to be done? We will like to have a contact person 
on the PANDA side to provide the first instructions and quick turn around 
support.

Also, can someone please add Dan Guest (daniel.hay.guest AT cern.ch>) to the 
hllhc computing list?

Thanks,
Amir