Parallel Implementation of Parallel Implementation of the KFParticle Vertexing Package for the KFParticle Vertexing Package for the CBM and ALICE Experiments the CBM and ALICE Experiments Ivan Kisel Ivan Kisel 1,2,3 1,2,3 , Igor Kulakov , Igor Kulakov 1,4 1,4 , , Maksym Zyzak Maksym Zyzak 1,4 1,4 1 – Goethe-Universität Frankfurt, Frankfurt am Main, Germany 1 – Goethe-Universität Frankfurt, Frankfurt am Main, Germany 2 – Frankfurt Institute for Advanced Studies, Frankfurt am Main, Germany 2 – Frankfurt Institute for Advanced Studies, Frankfurt am Main, Germany 3 – 3 – GSI Helmholtzzentrum für Schwerionenforschung GmbH, Darmstadt, Germany GSI Helmholtzzentrum für Schwerionenforschung GmbH, Darmstadt, Germany 4 – Taras Shevchenko National University of Kyiv, Kyiv, Ukraine 4 – Taras Shevchenko National University of Kyiv, Kyiv, Ukraine Computing in High Energy and Nuclear Physics Computing in High Energy and Nuclear Physics 24 May 2012 24 May 2012 New York New York
14
Embed
Ivan Kisel 1,2,3 , Igor Kulakov 1,4 , Maksym Zyzak 1,4
Parallel Implementation of the KFParticle Vertexing Package for the CBM and ALICE Experiments. Ivan Kisel 1,2,3 , Igor Kulakov 1,4 , Maksym Zyzak 1,4 1 – Goethe- Universität Frankfurt, Frankfurt am Main, Germany - PowerPoint PPT Presentation
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Parallel Implementation of Parallel Implementation of the KFParticle Vertexing Package for the KFParticle Vertexing Package for
the CBM and ALICE Experimentsthe CBM and ALICE Experiments
Ivan KiselIvan Kisel1,2,31,2,3, Igor Kulakov , Igor Kulakov 1,41,4, , Maksym Maksym ZyzakZyzak1,41,4
1 – Goethe-Universität Frankfurt, Frankfurt am Main, Germany1 – Goethe-Universität Frankfurt, Frankfurt am Main, Germany2 – Frankfurt Institute for Advanced Studies, Frankfurt am Main, 2 – Frankfurt Institute for Advanced Studies, Frankfurt am Main, GermanyGermany3 – 3 – GSI Helmholtzzentrum für Schwerionenforschung GmbH, Darmstadt, GSI Helmholtzzentrum für Schwerionenforschung GmbH, Darmstadt, GermanyGermany4 – Taras Shevchenko National University of Kyiv, Kyiv, Ukraine4 – Taras Shevchenko National University of Kyiv, Kyiv, Ukraine
Computing in High Energy and Nuclear PhysicsComputing in High Energy and Nuclear Physics24 May 201224 May 2012
New YorkNew York
ContentContent
• Concept of the KFParticle package
• The block-diagram of KFParticle algorithm
• Functionality of the KFParticle package
• SIMDized KFParticle
• Particles finding with the SIMDized KFParticle package
24 May 201224 May 2012 Igor Kulakov, CHEP 2012, New YorkIgor Kulakov, CHEP 2012, New York 22/14/14
24 May 201224 May 2012 Igor Kulakov, CHEP 2012, New YorkIgor Kulakov, CHEP 2012, New York 33/14/14
Reconstruction of Vertices and Decayed ParticlesReconstruction of Vertices and Decayed Particles
DD00DD00
KK--
++
AliKFVertex PrimVtx( ESDPrimVtx ); // Set primary vertex// Set daughters
AliKFParticle K( ESDp1, -321 ), pi( ESDp2, 211 );
AliKFParticle D0( K, pi ); // Construct mother
PrimVtx += D0; // Improve the primary
vertex.
D0.SetProductionVertex( PrimVtx ); // m3 is fully
fitted
K.SetProductionVertex( D0 ); // K is fully fitted
pi.SetProductionVertex( D0 ); // pi is fully fitted
r = { x, y, z, px, py, pz, E }
Position, direction and momentum
State vector
• Mother and daughter particles have the same state vector and are treated in the same way
•
Geometry independent
•
Reconstruction of decay chains
•
Kalman filter (KF) based
x, y, z, px, y, z, pxx, p, pyy, p, pzz, E, m, L, c, E, m, L, cx, y, z, px, y, z, pxx, p, pyy, p, pzz, E, m, L, c, E, m, L, c
KFParticle: powerful tool for physics analysis
KFParticle: powerful tool for physics analysis
ExperimentsExperiments
24 May 201224 May 2012 Igor Kulakov, CHEP 2012, New YorkIgor Kulakov, CHEP 2012, New York 44/14/14
CBM (FAIR, Germany) – a fixed-target experiment
KFParticle is developed based on the ALICE and CBM experiments.
• Up to 1000 charged particles/collision
• Non-homogeneous magnetic field
• 107 AuAu collisions/sec
• Reconstruction of the full event topology is required in the first level trigger
• The speed and efficiency of the reconstruction is crucial
• Few 1000s charged particles/collision
• High statistic is collected – a speed of short-lived particles reconstruction is important for the physics analysis
ALICE (CERN, Switzerland) – a collider experiment
KFParticle AlgorithmKFParticle Algorithm
24 May 201224 May 2012 Igor Kulakov, CHEP 2012, New YorkIgor Kulakov, CHEP 2012, New York 55/14/14
The mathematics is described in:S. Gorbunov and I. Kisel, Reconstruction of decayed particles based on the Kalman filter. CBM-SOFT-note-2007-003, 7 May 2007
Initial approximatio
nr0, C0
Estimation of
parameters and their
errorsrn, Cn
Filtering a daughter particle
usingthe Kalman
Filter mathematics
Transport of a daughter to the
decay point
n p
artic
les
n p
art
icle
s
N ite
ratio
ns
N ite
rati
on
s
r0, C0 r0i, C0 rn
i, Cni rn, Cn
N iterations
rki, Ck
i
Structure of KFParticleStructure of KFParticle
24 May 201224 May 2012 Igor Kulakov, CHEP 2012, New YorkIgor Kulakov, CHEP 2012, New York 66/14/14
Intermediate level (for advanced users)
• Feasibility studies• Reconstruction of particles• Reconstruction of decay chains
Low level (for developers, basic functionality)
• Transport functions• Calculation of distances and
deviations between particles, a particle and a point
• KF mathematics• Constraints
High level or KFParticle-Light (for users and triggering)
• Reconstruction of standard decays (K0
s, hyperons, D0-mesons,…)
• Reconstruction of event topology• On-line selection of events
Functionality in ALICE and CBMFunctionality in ALICE and CBM
Functions ALICE CBMConstruct, SetMassConstraint, SetProductionVertex, SetVtxGuess + +