HPC resources LHC experiments are using (as seen from ATLAS - CMS lens) erhtjhtyhy Doug Benjamin Argonne National Lab.

Slides:



Advertisements
Similar presentations
NSF Experience with Management of Research Infrastructure
Advertisements

LinkSCEEM-2: A computational resource for the Eastern Mediterranean.
Exception for Informed Consent for Emergency Research and the IRB: Operationalizing It Ronald F. Maio DO, MS Director, Office of Human Research Compliance.
Office of Science U.S. Department of Energy SLAC Users Organization Meeting July 6, 2004 Dr. Robin Staffin, Associate Director Office of High Energy Physics.
Plans for Exploitation of the ORNL Titan Machine Richard P. Mount ATLAS Distributed Computing Technical Interchange Meeting May 17, 2013.
Grants.gov – What to Expect Presented by the Office for Research Cindy Hope, Director Office for Sponsored Programs.
Global Science experiment Data hub Center Oct. 13, 2014 Seo-Young Noh Status Report on Tier 1 in Korea.
Grid Computing Oxana Smirnova NDGF- Lund University R-ECFA meeting in Sweden Uppsala, May 9, 2008.
1 What is PRACE? Hank Nussbacher PRACE Winter School, Tel Aviv, Feb 10, 2014.
On the organization and conduct of expert examination in science and technology in the USA and the European Union Scientific Research.
How to Give a Good Technical Presentation Dean F. Hougen REAL Lab (Robotics, Evolution, Adaptation, and Learning Laboratory) School of Computer Science.
BNL Tier 1 Service Planning & Monitoring Bruce G. Gibbard GDB 5-6 August 2006.
Population of Ukraine: 45 million people Area: km2 Crimea: 1,9 million people Occupied territories of Donetsk and Luhansk regions ~ 3, 5 mln >
ISTeC Research Computing Open Forum: Using NSF or National Laboratory Resources for High Performance Computing Bhavesh Khemka.
DOE/HENP Networking Update Vicky White October 26th, 2001 HENP Network Users Group Internet 2, Ann Arbor.
Considering Time in Designing Large-Scale Systems for Scientific Computing Nan-Chen Chen 1 Sarah S. Poon 2 Lavanya Ramakrishnan 2 Cecilia R. Aragon 1,2.
PEER 2003 Meeting 03/08/031 Interdisciplinary Framework Major focus areas Structural Representation Fault Systems Earthquake Source Physics Ground Motions.
Training Course on Data Management for Information Professionals and In-Depth Digitization Practicum September 2011, Oostende, Belgium Concepts.
The Helix Nebula marketplace 13 May 2015 Bob Jones, CERN.
Benefits. CAAR Project Phases Each of the CAAR projects will consist of a: 1.three-year Application Readiness phase ( ) in which the code refactoring.
May 23, 2007ALICE DOE Review - Computing1 ALICE-USA Computing Overview of Hard and Soft Computing Resources Needed to Achieve Research Goals 1.Calibration.
J. Templon Nikhef Amsterdam Physics Data Processing Group Large Scale Computing Jeff Templon Nikhef Jamboree, Utrecht, 10 december 2012.
High Performance Computing Activities at Fermilab James Amundson Breakout Session 5C: Computing February 11, 2015.
Ian Bird LHCC Referees; CERN, 2 nd June 2015 June 2,
A Statistical Analysis of Job Performance on LCG Grid David Colling, Olivier van der Aa, Mona Aggarwal, Gidon Moont (Imperial College, London)
Rome, Sep 2011Adapting with few simple rules in glideinWMS1 Adaptive 2011 Adapting to the Unknown With a few Simple Rules: The glideinWMS Experience by.
Scientific Data Processing Portal and Heterogeneous Computing Resources at NRC “Kurchatov Institute” V. Aulov, D. Drizhuk, A. Klimentov, R. Mashinistov,
First collisions in LHC
ICPSR Data Fair November 8, 2010 Katherine McNeill, MIT Libraries
Finding Your Faculty Job
A new model for volunteer computing
Review of the WLCG experiments compute plans
Status of WLCG FCPPL project
Past research work and research work in progress on elephant flows
A Brief Introduction to NERSC Resources and Allocations
BigPanDA Workflow Management on Titan
A Dutch LHC Tier-1 Facility
U.S. ATLAS Tier 2 Computing Center
Working Group 4 Facilities and Technologies
The LHC Computing Grid Visit of Mtro. Enrique Agüera Ibañez
Jarek Nabrzyski Director, Center for Research Computing
Defining EOSC Rules of Engagement Damien Lecarpentier (CSC)
CERN presentation & CFD at CERN
LinkSCEEM-2: A computational resource for the Eastern Mediterranean
HPC DOE sites, Harvester Deployment & Operation
SA1-NA4 Meeting 15 September 2004
Christos Markou Institute of Nuclear Physics NCSR ‘Demokritos’
DIRAC services.
Status and Prospects of The LHC Experiments Computing
Scientific LNGS: users point of view
PanDA in a Federated Environment
Accessing LinkSCEEM resources
The Shifting Landscape of CI Funding
UK Status and Plans Scientific Computing Forum 27th Oct 2017
CyberShake Study 16.9 Discussion
ATLAS Sites Jamboree, CERN January, 2017
Introduction to XSEDE Resources HPC Workshop 08/21/2017
Univ. of Texas at Arlington BigPanDA Workshop, ORNL
CernVM Status Report Predrag Buncic (CERN/PH-SFT).
SA1 ROC Meeting Bologna, October 2004
This work is supported by projects Research infrastructure CERN (CERN-CZ, LM ) and OP RDE CERN Computing (CZ /0.0/0.0/1 6013/ ) from.
Data Issues Julian Borrill
New strategies of the LHC experiments to meet
Year 10 Computer Science Hardware - CPU and RAM.
This work is supported by projects Research infrastructure CERN (CERN-CZ, LM ) and OP RDE CERN Computing (CZ /0.0/0.0/1 6013/ ) from.
Coming Out as a Transgender Woman at a Major HPC Center
A Possible OLCF Operational Model for HEP (2019+)
Alexei Klimentov BNL Jun 25, 2019
Workflow and HPC erhtjhtyhy Doug Benjamin Argonne National Lab.
Welcome to HTCondor Week #17 (year 32 of our project)
Presentation transcript:

HPC resources LHC experiments are using (as seen from ATLAS - CMS lens) erhtjhtyhy Doug Benjamin Argonne National Lab

Thanks and apologies Talk is seen through ATLAS – CMS lens Thanks : based on feed back and information from Andrej Filipcic (ATLAS) Dirk Hufnagel (CMS) Alexei Klimentov (ATLAS) Daniele Spiga (CMS) Apologies to ALICE and LHCb. ALICE successful in using several HPC centers for a while NERSC and OLCF for example. LHCb – HPC talk from 2018. https://indico.cern.ch/event/651338/contributions/2871262/attachments/1594114/2523997/preGDB-HPC_at_LHCb.pdf

ATLAS HPC Sites/ PanDA queues

CMS HPC sites

ALOCATION Process

European HPC’s (non – pledged) Typically negotiations directly with HPC center Centers allocate most of their time through the local allocation process HPC’s provide a fraction of resources to PRACE for HPC scientific communities, and the approval procedure is much more complicated and stricter when a project is approved, the PRACE assigns the allocation to the HPC center which best fits the requirements of the application. For the LHC experiments this would create additional technical issues.

US DOE LCF process Source: Talk given by Katherine Riley – ALCF science director. In 2018-2109 ATLAS ALCC: 80M hrs OLCF 80M hrs ALCF CMS: startup allocation at ALCF

US DOE – NERSC allocation program Most time comes through DOE production DOE Production time – managed through DOE program managers- Total ATLAS CMS time essentially constant 2018 ATLAS: 175 M MPP hours CMS: 25M MPP hours Total: 200 M hrs. 2019 ATLAS: 120 M MPP hours CMS: 82M MPP hours Total: 202 M hrs.

National Science Foundation (NSF) XSEDE allocation program https://portal.xsede.org/allocations/research CMS’ success Q2 2019 – Q1 2020 PSC/Bridges - 4.78 M hours TACC/Stampede2 - 0.73 (node) hours SDSC/Comet - 3.3M hours

Going forward Next few months - LHC experiments should think if ALCF – ADSP program would really help us in short term. As more HPC’s with accelerators come on-line – challenge will be to get time if our needs are predominantly CPU only

Acknowledgement Policy - CSCS NERSC Acknowledgements Users must quote and acknowledge the use of CSCS resources in all publications related to their production and development projects as follows: "This work was supported by a grant from the Swiss National Supercomputing Centre (CSCS) under project ID ###"

Conclusions ATLAS and CMS both use a variety of HPC sites CMS has done a lot of work to cvmfs available at sites they use. ATLAS should make use of all CMS’ hard work Should come up with common criteria for cvmfs where possible use containers a places that can not or will not use cvmfs. We will discuss the Operational model evolution later in this meeting. Need to keep applying for yearly allocations through various programs. Need multi-year allocations. NERSC ATLAS allocation - 2018 (175 M MPP hrs) to 2109 (120 M MPP hrs) Notified just before the yearly allocation started. If HPC’s are to be come part of the pledged resources then we need have multi-year allocations. This is a discussion that should happen at the Funding Agency levels.