Third INFN International School on: "Architectures, tools and methodologies for developing efficient large scale scientific computing applications" ESC11 - Bertinoro (Forlì-Cesena) Italy 23-29 October 2011

Oct 23, 2011, 4:00 PM → Oct 29, 2011, 3:00 PM Europe/Rome

Bertinoro

PRELIMINARY PROGRAM

Sun, October 23

4:00 PM → 8:30 PM Registration and Welcome

Class Photos

• Class_from_the_back.jpg
• class_front.jpg
• speaker_and_class.jpg
• students_at_work.jpg
• Tim.jpg

Group Photos gruppo_1.jpg gruppo_2.jpg

8:30 PM → 8:45 PM Welcome Dinner

Mon, October 24

8:30 AM → 8:00 PM Session 1

8:30 AM Welcome and Opening of the School

9:00 AM Concepts of performance and efficiency

Speaker: Dr Peter Elmer (Princeton University)

Slides esc11_elmer_intro_perf.pdf

9:50 AM Modern processors and related optimisation topics - Part 1

Speaker: Mr Sverre Jarp (CERN)

Slides Jarp_ESC11_Architecture_New.pdf

10:40 AM Coffee break

11:00 AM Modern processors and related optimisation topics - Part 2

Speaker: Mr Sverre Jarp (CERN)

11:50 AM Introduction to Performance tuning tools

Speakers: Lassi Tuura (Fermilab), Dr Peter Elmer (Princeton University)

Slides 20111023-ESC11-Exercises.pdf

12:40 PM Lunch break

2:15 PM Floating point computation: accuracy, optimization, vectorization (with exercises)

Speaker: Vincenzo Innocente (CERN)

Slides OptimalFloatingPoint2011.pdf

3:00 PM Floating point computation: accuracy, optimization, vectorization (with exercises)

Speaker: Vincenzo Innocente (CERN)

3:45 PM Coffee break

4:00 PM Floating point computation: accuracy, optimization, vectorization (with exercises)

Speaker: Vincenzo Innocente (CERN)

VI Blogs https://twiki.cern.ch/twiki/bin/view/LCG/MultiCoreBlogs

4:45 PM Floating point computation: accuracy, optimization, vectorization (with exercises)

Speaker: Vincenzo Innocente (CERN)

6:30 PM Evening Lecture: "GPU for scientific computing" - Alessandro Lonardo - INFN Roma 1

Slides lonardo_esc_11.pdf

8:30 PM → 8:30 PM Dinner

Tue, October 25

8:30 AM → 8:00 PM Session 2

8:30 AM Efficient C++ coding

Speakers: Dr Sebastien Binet (LAL/IN2P3), Vincenzo Innocente (CERN)

Slides 20111024-binet.pdf

9:20 AM Efficient C++ coding

Speakers: Dr Sebastien Binet (LAL/IN2P3), Vincenzo Innocente (CERN)

10:10 AM Coffee break

10:30 AM Exercises - Basic C++ optimisations

Speakers: Dr Sebastien Binet (LAL/IN2P3), Vincenzo Innocente (CERN)

Slides 20111023-ESC11-Git-Update.pdf

11:15 AM Exercises - Basic C++ optimisations

Speakers: Dr Sebastien Binet (LAL/IN2P3), Vincenzo Innocente (CERN)

12:40 PM Lunch break

2:15 PM The Memory Crisis

Speaker: Lassi Tuura (Fermilab)

Slides 20111023-ESC11-Memory-Lectures.pdf

3:00 PM How memory allocation works

Speaker: Lassi Tuura (Fermilab)

3:45 PM Coffee break

4:00 PM Exercises - Memory Allocations

Speaker: Lassi Tuura (Fermilab)

4:45 PM Exercises - Memory Allocations

Speaker: Lassi Tuura (Fermilab)

6:30 PM Evening Lecture: "Architectures of High Throughput DAQ systems" - Niko Neufeld (CERN)

Slides High_Throughput_DAQ-esc11.pptx

8:30 PM → 8:30 PM Dinner

Wed, October 26

8:30 AM → 4:30 PM Session 3

8:30 AM Exercises (Floating Point, Memory use, C++)

Exercise list http://lat.web.cern.ch/lat/esc11/

9:20 AM Exercises (Floating Point, Memory use, C++)

10:10 AM Coffee break

10:30 AM Exercises (Floating Point, Memory use, C++)

11:20 AM Exercises (Floating Point, Memory use, C++)

12:20 PM Lunch break

2:00 PM Exercises (Floating Point, Memory use, C++)

2:50 PM Exercises (Floating Point, Memory use, C++)

3:40 PM Exercises (Floating Point, Memory use, C++)

5:30 PM → 7:45 PM Social tour

Guided tour to CESENA and visit of the city
Guided tour to Casa Artusi (centre of gastronomic culture dedicated to Italian home cooking).

7:45 PM → 7:45 PM Social dinner

Casa Artusi Restaurant

Thu, October 27

8:30 AM → 8:00 PM Session 4

8:30 AM Parallel programming theory

Speaker: Dr Tim Mattson (Intel)

Slides par-comp-omp-hands-on-ESC11.pdf

9:20 AM An introduction to OpenMP

Speaker: Dr Tim Mattson (Intel)

10:10 AM Coffee break

10:30 AM Synchronization in OpenMP

Speaker: Dr Tim Mattson (Intel)

11:15 AM Work sharing constructs

Speaker: Dr Tim Mattson (Intel)

12:05 PM The OpenMP data environment

Speaker: Dr Tim Mattson (Intel)

12:50 PM Lunch break

2:30 PM I/O Efficiency

Speaker: Vincenzo Maria Vagnoni (BO)

Slides ioefficiency.pdf ioefficiency.pptx

4:00 PM Coffee break

4:15 PM I/O Efficiency

Speaker: Vincenzo Maria Vagnoni (BO)

6:30 PM Evening Lecture: "OpenCL and the quest for Performance Portability" - Timothy G. Mattson (Intel)

Slides ocl-port-perf.pdf

8:30 PM → 8:30 PM Dinner

Fri, October 28

8:30 AM → 7:30 PM Session 5

8:30 AM OpenMP tasks

Speaker: Dr Tim Mattson (Intel)

9:20 AM OpenMP Memory model

Speaker: Dr Tim Mattson (Intel)

10:10 AM Coffee break

10:30 AM A survey of programming models

Speaker: Dr Tim Mattson (Intel)

11:30 AM A survey of programming models

Speaker: Dr Tim Mattson (Intel)

12:50 PM Lunch break

2:30 PM Lecture on low-level CPU profiling

Speaker: Mr Sverre Jarp (CERN)

Slides SJAN-Understanding_performance_tuning_v3.pdf

3:15 PM Physical design of SW

Speaker: Benedikt Hegner (CERN)

Slides ESC_swdesign.pdf

4:00 PM Coffee break

4:15 PM Physical Design Examples/HEP Frameworks

Speaker: Benedikt Hegner (CERN)

Slides ESC_swdesignexample.pdf

5:00 PM Software Development Tools

Speaker: Benedikt Hegner (CERN)

Slides ESC_swtools.pdf

5:45 PM Announcements

Speaker: Federico Ruggieri (ROMA3)

Slides friday_28102011.ppt

6:30 PM Evening Lecture: "Linux on Multicore: challanges and perspective" - Andrea Arcangeli (RedHat)

Slides numa_bertinoro11.pdf

8:30 PM → 8:30 PM Dinner

Sat, October 29

8:30 AM → 2:50 PM Session 6

8:30 AM Students feedback

9:00 AM Final examination

11:00 AM Coffee break

11:30 AM Delivery of certificates of attendance

12:00 PM Lunch

2:30 PM Shuttle departure