Using HEP experiment workflows for the benchmarking and accounting of WLCG computing resources

Benchmarking of CPU resources in WLCG has been based on the HEP-SPEC06 (HS06) suite for over a decade. It has recently become clear that HS06, which is based on real applications from non-HEP domains, no longer describes typical HEP workloads. The aim of the HEP-Benchmarks project is to develop a new benchmark suite for WLCG compute resources, based on real applications from the LHC experiments. By construction, these new benchmarks are thus guaranteed to have a score highly correlated to the throughputs of HEP applications, and a CPU usage pattern similar to theirs. Linux containers and the CernVM-FS filesystem are the two main technologies enabling this approach, which had been considered impossible in the past. In this paper, we review the motivation, implementation and outlook of the new benchmark suite.

[1]  Johannes Elmsheuser,et al.  ATLAS Grid Workflow Performance Optimization , 2019 .

[2]  Andrzej Nowak.  An update on perfmon and the struggle to get into the Linux kernel , 2010 .

[3]  D Gregori,et al.  LHCb Computing TDR , 2005 .

[4]  Brian A. Wichmann,et al.  A Synthetic Benchmark , 1976, Comput. J..

[5]  G Benelli,et al.  The CMSSW benchmarking suite: Using HEP code to measure CPU performance , 2010 .

[6]  John L. Henning.  SPEC CPU2006 benchmark descriptions , 2006, CARN.

[7]  S. Eranian.  Perfmon2: a flexible performance monitoring interface for Linux , 2010 .

[8]  Vanessa Sochat,et al.  Singularity: Scientific containers for mobility of compute , 2017, PloS one.

[9]  M. Giovannozzi,et al.  SixTrack Version 5: Status and New Developments , 2019, Journal of Physics: Conference Series.

[10]  E McIntosh.  Benchmarking computers for HEP , 1992 .

[11]  Jack Dongarra,et al.  Computer benchmarking: paths and pitfalls , 1987 .

[12]  Franco Brasolin,et al.  Benchmarking the ATLAS software through the Kit Validation engine , 2010 .

[13]  Jim Gray,et al.  The Benchmark Handbook for Database and Transaction Systems , 1993 .

[14]  Rolf Seuster,et al.  Running ATLAS workloads within massively parallel distributed applications using Athena Multi-Process framework (AthenaMP) , 2015 .

[15]  Reinhold Weicker,et al.  Dhrystone: a synthetic systems programming benchmark , 1984, CACM.

[16]  M. Michelotto,et al.  Next Generation of HEP CPU benchmarks , 2019, Journal of Physics: Conference Series.

[17]  Ahmad Yasin,et al.  A Top-Down method for performance analysis and counters architecture , 2014, 2014 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS).

[18]  J. L. Martin,et al.  Computer benchmarking: Paths and pitfalls: The most popular way of rating computer performance can confuse as well as inform; avoid misunderstanding by asking just what the benchmark is measuring , 1987, IEEE Spectrum.

[19]  Michael Wong.  C++ benchmarks in SPEC CPU2006 , 2007, CARN.

[20]  Peter Wegner,et al.  A comparison of HEP code with SPEC1 benchmarks on multi-core worker nodes , 2010 .

[21]  Servesh Muralidharan,et al.  Trident : An Automated System Tool for Collecting and Analyzing Performance Counters , 2019 .

[22]  Jack J. Dongarra,et al.  The LINPACK Benchmark: past, present and future , 2003, Concurr. Comput. Pract. Exp..

[23]  P Charpentier.  Benchmarking worker nodes using LHCb productions and comparing with HEPSpec 06 , 2017 .

[24]  Predrag Buncic,et al.  Distributing LHC application software and conditions databases using the CernVM file system , 2011 .

[25]  Kaivalya M. Dixit,et al.  Overview of the SPEC Benchmarks , 1993, The Benchmark Handbook.

[26]  Elizabeth Sexton-Kennedy,et al.  Implementation of a Multi-threaded Framework for Large- scale Scientific Applications , 2015 .

[27]  Walter Lampl,et al.  A Roadmap for HEP Software and Computing R&D for the 2020s , 2019 .

[28]  Philippe Charpentier.  IOP : Benchmarking worker nodes using LHCb productions and comparing with HEPSpec06 , 2017 .