CERN openlab/Intel Workshop on Numerical Computing (3rd instance)

27 May 2013, 08:40 → 28 May 2013, 17:00 Europe/Zurich

Building 593, Room 11 (CERN)

The Workshop

CERN openlab and Intel are pleased to announce a third instance of the Numerical Computing workshop, scheduled for May 27^th and 28^th at CERN. The course will be taught by external experts and several CERN staff. The workshop is targeted at attendees with good programming experience who would like to understand and control the floating-point calculations in their scientific programs.

Please note the agenda is still under construction.

Topics

•             Advanced Floating point computing on x86
•             Pitfalls in High Energy Physics software
•             Floating point control in Intel Software
•             Exploration and future directions in accuracy control
•             Hands-on tutorials

Speakers

Jeff Arnold, Intel SSG
Martyn Corden, Intel SSG
Florent de Dinechin, ENS Lyon
Vincenzo Innocente, CERN
Lorenzo Moneta, CERN
Danilo Piparo, CERN

Attendance is free, as for the regular openlab workshops, but but requires registration in EDH. If you don’t have EDH access, please let us know by e-mail that you’re planning to participate (openlab.workshops@cern.ch). A laptop is needed for accessing the servers that will be used for the exercises.
 
 

Monday 27 May

Introduction to the workshop

Andrzej Nowak

Convener: Andrzej Nowak (CERN)

slides NC-AN-Intro.pdf

Floating-point issues and pitfalls seen in High Energy Physics calculations (PH/SFT)

• Reconstruction (30 m) - V. Innocente
• Data analysis (30 m) – L. Moneta

Conveners: Lorenzo Moneta (CERN), Vincenzo Innocente (CERN)

slides LM-FPDataAnalysis_May2013.pdf VI-FPinEHEP.pdf

1 Floating Point in Experimental HEP Data Processing

Speakers: Lorenzo Moneta (CERN), Vincenzo Innocente (CERN)

Slides FPinEHEP.pdf

2 Floating point issues in Data Analysis

Slides FPDataAnalysis_May2013.pdf

10:00 Break

Understand floating-point arithmetic (x86-64) Part - 1 (Jeff Arnold, Intel Corporation)

Convener: Jeffrey Arnold (Intel Corporation (US))

paper FPWorkshop-Exercises.pdf

slides FPWorkshop-print.pdf

3 Understand floating-point arithmetic

Speaker: Jeffrey Arnold (Intel Corporation (US))

11:00 Break

Exercises (Day 1)

Please bring your own laptop

4 Exercises

Speaker: Jeffrey Arnold (Intel Corporation (US))

Opportunities for target accuracy in HEP software: focus on mathematical functions

Convener: Danilo Piparo (CERN)

slides DPiparo-OpenLabFPWorkshop_5_2013.pdf

5 Opportunities for target accuracy in HEP software: focus on mathematical functions

Speaker: Danilo Piparo (CERN)

Slides OpenLabFPWorkshop_5_2013.pdf

12:30 Lunch (on your own)

Understand floating-point arithmetic (x86-64) Part - 2 (Jeff Arnold, Intel Corporation)

Convener: Jeffrey Arnold (Intel Corporation (US))

6 Understand floating-point arithmetic (x86-64) part 2

Speaker: Jeffrey Arnold (Intel Corporation (US))

Exercises (Day 1)

Please bring your own laptop

Tuesday 28 May

Floating-point control in the Intel C/C++ compiler and relevant libraries (Martyn Corden, Intel Corporation)

Convener: Martyn Corden (Intel)

slides MC_FP_control.pdf

7 Floating-point control in the Intel C/C...

Speaker: Martyn Corden

10:00 Break

From CRLibm to Metalibm: assisting the production of high-performance proven floating-point code

Convener: Florent de Dinechin

slides FdD-2013-CERN.pdf

11:00 Break

Alice TPC online tracking on GPUs and CPUs

Convener: David Michael Rohr (Johann-Wolfgang-Goethe Univ. (DE))

slides DRohr_cern_numerical_stability_2013_05_28.pdf

Floating Point on FPGAs

Convener: Florent de Dinechin (ENS Lyon)

slides FdD-2013-CERN-PFGA.pdf

12:15 Lunch (on your own)

Exercises (Day 2)

Bring your own laptop