The Curious Engineer

All blogs / Performance / optimization

Georg Hager's Blog

blogs.fau.de/hager

Half-day tutorial “Performance Engineering for Sparse Linear Solvers” to be featured again at ISC26

Performance / optimizationCPU / microarchitectureDev tools / editors / IDEsScience

’10’15’20’25
Posts
106 from 12 Nov 2006 to 14 Jun 2026
Posting
now and then (1 in the last year)
Kind
personal blog
Feed
https://blogs.fau.de/hager/feed

2026 1

  1. Half-day tutorial “Performance Engineering for Sparse Linear Solvers” to be featured again at ISC26

2025 1

  1. Get a full dose of performance engineering with our two half-day tutorials at ISC25!

2024 2

  1. Paper on Write-Allocate Evasion is Best Paper Candidate at IPDPS 2024
  2. New tutorial “Performance Engineering for Linear Solvers” at ISC High Performance 2024

2023 4

  1. SIAM Parallel Processing 24 (PP24) Minisymposium on “Advancements in Sparse Linear Algebra: Hardware-Aware Algorithms and Optimization Techniques”
  2. Node-Level Performance Engineering tutorial to be featured again at SC23
  3. PERMAVOST Workshop submission deadline approaching
  4. New tutorial on “Core-Level Performance Engineering” accepted for ICPE 2023

2022 2

  1. ISC 2022 has started
  2. Gprofng is the next-generation GNU profiler

2021 7

  1. SIAM Parallel Processing 2022 Minisymposium on “Advances in Performance Modeling of Parallel Code”
  2. John McCalpin (“Dr. Bandwidth”): Memory Bandwidth and System Balance in HPC Systems – 2021 Update
  3. LIKWID 5.2.1 is out!
  4. A riddle in D/A conversion – and its solution
  5. Write-allocate evasion has finally arrived at Intel – or has it?
  6. IACS Stony Brook seminar talk available
  7. Upcoming: 38th VI-HPS Online Tuning Workshop, March 1-3, 2021

2020 9

  1. Tutorial: Empirical Roofline model with LIKWID
  2. LIKWID 5.1 released
  3. EoCoE webinar on A64FX
  4. SC20 tutorial “Node-Level Performance Engineering”
  5. LIKWID 5.0.2 released
  6. Fujitsu’s A64FX demystified. Well, somewhat.
  7. Introducing the MachineState reproducibility tool
  8. What is this LIKWID, anyway?
  9. PMACS 2020 Workshop Call for Papers

2019 11

  1. Intel’s port 7 AGU blunder
  2. Node-Level Performance Engineering Course at LRZ
  3. PMBS19 Workshop Best Late-Breaking Paper Award
  4. LIKWID 5.0 is here
  5. SC19 incoming!
  6. ISC 2020 call for research papers: publish with Gold Open Access – deadline extended
  7. PPAM 2019 Workshops Best Paper Award for Dominik Ernst
  8. Intel Architecture Code Analyzer (IACA) R.I.P.! All hail OSACA!
  9. Christie Alappat comes second place in ACM Student Research Competition Grand Finals
  10. LIKWID 4.3.4 released
  11. The McCalpin STREAM benchmark: How to do it right and interpret the results

2018 8

  1. Christie Louis Alappat wins the SC18 ACM Student Research Competition Award
  2. SC18 incoming!
  3. A LIKWID bouquet
  4. And the 2018 Gauss Award goes to: Erlangen!
  5. LIKWID marker overhead and “Meltdown” patches
  6. Himeno stencil benchmark: ECM model, SIMD, data layout
  7. Fun with likwid-pin and shepherd threads
  8. LIKWID 4.3 is out – and it has a quick reference, too!

2017 7

  1. Himeno stencil benchmark: Roofline performance modeling and validation
  2. Open Source Architecture Code Analyzer (OSACA) has been released
  3. Why IPC (or CPI) is not a good performance metric
  4. New HPC@RRZE website is now online
  5. Node-Level Performance Engineering tutorial to be featured again at SC17
  6. Benchmarking the memory hierarchy of the Intel “Skylake” Xeon E3-1275 v5 using the vector triad
  7. Benchmarking the memory hierarchy of the new AMD Ryzen CPU using the vector triad

2016 2

  1. Stepanov test faster than light?
  2. Energy vs. performance: Introducing the Z-plot

2015 6

  1. WPMVP 2016: 3rd Workshop on Programming Models for SIMD/Vector Processing
  2. Fooling the masses – Stunt 16: Worship the God of Automation!
  3. Preview for SC15 tutorial on “Node-Level Performance Engineering” now available
  4. ISC15 Workshop on “Performance Modeling:Methods and Applications”
  5. Fooling the masses – Stunt 15: Play mysterious!
  6. Attendee survey results from the SC14 “Code Optimization War Stories” BoF

2014 6

  1. SC14 Intel Parallel Universe Computing Challenge: The Squad fails in the semi-final
  2. SC14 Parallel Universe Computing Challenge Match #1: The Squad smashes the “Invincible Buckeyes”
  3. Team members for “Gaussian Elimination Squad” now official
  4. “Gaussian Elimination Squad” to fight again at SC14
  5. Blaze 2.1 has been released
  6. Bill Gropp on “Engineering for Performance in High Performance Computing”

2013 11

  1. Gaussian Elimination Squad wins Intel Parallel Universe Computing Challenge at SC13!
  2. SC13 Intel Parallel Universe Challenge: The Squad moves to the final round!
  3. SC13 Intel Parallel Universe Challenge: China eliminated!
  4. SC13 Intel Parallel Universe Computing Challenge: Teams disclosed
  5. Gaussian Elimination Squad to fight in “SC13 Intel Parallel Universe Computing Challenge”
  6. Intel vs. GCC for the OpenMP vector triad: Barrier shootout!
  7. Fooling the masses – Stunt 14: Secretly use fancy hardware setups and software tricks!
  8. Fooling the masses – Stunt 13: If they get you cornered, blame it all on OS jitter!
  9. Fooling the masses – Stunt 12: Redefine “performance” to suit your needs!
  10. Fooling the masses – Stunt 11: Show data! Plenty. And then some.
  11. Fooling the masses – Stunt 10: Always emphasize the “interesting” part of your work!

2012 4

  1. Fooling the masses – Stunt 9: Boast massive speedups with accelerators!
  2. Fooling the masses – Stunt 8: Impress your audience with awe-inspiring accuracy
  3. Fooling the masses – Stunt 7: Be creative when comparing scaled performance!
  4. Fooling the masses – Stunt 6: Ignore affinity and topology issues!

2010 7

  1. The 400x GPU speedup baloney
  2. Fooling the masses – Stunt 5: Instead of performance, plot absolute runtime versus CPU count!
  3. Fooling the masses – Stunt 4: Quietly employ weak scaling to show off!
  4. Fooling the masses – Stunt 3: The log scale is your friend!
  5. Fooling the masses – Stunt 2: Slow down code execution!
  6. Fooling the masses – Stunt 1: Report speedup, not absolute performance!
  7. Fooling the masses with performance results on parallel computers – prelude

2009 1

  1. LIKWID tools online

2008 10

  1. A case for the non-temporal store
  2. The “roofline model” for kernel performance assessment
  3. Floating point computations demystified
  4. Allocatable array alignment with Intel Fortran compilers
  5. Energy efficient HPC?
  6. Lecture “Programming Techniques for Supercomputers” in SS 08
  7. Intel Threading Building Blocks (TBB 2.0) on RRZE systems
  8. ccNUMA and Linux buffer cache
  9. Restricting member function calls by numeric template parameters
  10. News about Sun UltraSPARC T2

2007 6

  1. Sun SPARC Enterprise T5120 at RRZE
  2. Sun UltraSPARC T2 under test
  3. IMB multi-mode Ping-Pong demystified?
  4. OpenMP, ccNUMA and C++
  5. Array summation benchmark
  6. Benchmarking fun with calloc() and zero pages

2006 1

  1. Gauss Centre for Supercomputing founded