Difference: AtlasEdinburghGPUComputing (19 vs. 20)

Revision 202011-11-21 - AndrewWashbrook

Line: 1 to 1
 
META TOPICPARENT name="AtlasEdinburghGroupUpgradeSoftware"
Changed:
<
<

ATLAS Edinburgh GPU Computing

>
>
AtlasEdinburghGPUComputing
 
Changed:
<
<

Porting the Z-finder algorithm to GPU (Chris Jones)

>
>

High Level Trigger Studies

Porting the HLT Z-finder algorithm to GPU

 
Changed:
<
<

Overview

>
>
Chris Jones, Andrew Washbrook
 
Changed:
<
<
ATLAS is a general-purpose detector in the Large Hadron Collider (LHC), at CERN. Its purpose is to investigate the physics of the Standard Model, including the search for the Higgs boson, and possible new physics beyond this model.
>
>

Project Overview

 
Changed:
<
<
With around a billion particle collisions occurring every second within the detector, it is impossible to store the data collected in its entirety. It is responsibility of the trigger to reduce this data, selecting out events of potential scientific interest for storage and further analysis. The works in three main stages: level 1 running on customised hardware, and levels 2 and 3 in software on server farms.
>
>
With around a billion particle collisions occurring every second within the detector, it is impossible to store the data collected by ATLAS in its entirety. It is responsibility of the trigger to reduce this data, selecting out events of potential scientific interest for storage and further analysis. The works in three main stages: level 1 running on customised hardware, and levels 2 and 3 in software on server farms.
  IDScan is a track-reconstruction program running in level 2 of the trigger, and taking its input from the inner detector. The first algorithm used by IDScan is the Z-finder algorithm, the purpose of which is to produce a good estimate for the z-coordinate of the collision event along the axis of the detector. This reduces the execution time of the track-reconstruction algorithms later in the chain.

The aim of this project will be to port the Z-finder algorithm to GPU using NVIDIA's C for CUDA, and investigate the performance.

Changed:
<
<

Project Writeup

>
>

Project Report

 
  • GPUs and the ATLAS experiment (pdf)
Changed:
<
<

SIMT design of the High Level Trigger Kalman Fitter (Maria Rovatsou)

>
>

Project Code

 
Changed:
<
<

Overview

>
>
  • Z-Finder GPU version (tar.gz)
 
Changed:
<
<
The ATLAS detector is one of the two biggest detectors of the Large Hadron Collider (LHC) at the European Center of Nuclear Physics (CERN) dedicated for research of the origin of mass, extra dimensions, and dark matter. The detector at design performance works at luminosity 1034 cm^2 s^2 with beam bunch rate of 40MHz. The rate of the collisions taking place in the accelerator is the input rate of the software that stores and processes them, the ATLAS Trigger. The Trigger is divided into three levels, the first level is a custom electronics trigger and the other two levels compose the High-Level trigger that is responsible for combining the data from the detector and reconstructing the tracks of the particles in order to store interesting data. The track reconstruction is achieved with the utilization of the Kalman filter technique. The current implementation of the filter is a distributed sequential implementation. An acceleration of the filter would boost the whole process and would allow even more significant data to be available for processing. Apart from that, the proposed upgrade of the LHC in order to increase its performance, results in luminosity increase of one order of magnitude higher and would require a more efficient implementation of the Kalman filter. A possible way for acceleration of the filter is by porting it on a Graphics Processing Unit (GPU), which has immense computational power that can be exploited through programming models as CUDA. The proposed approach is a parallelization on thread level of the Kalman filter for execution on an NVIDIA GPU. The proposed programming model is CUDA and the analysis of its performance will be done with the CUDA profiling tools and with benchmarking tools.
>
>

SIMT design of the High Level Trigger Kalman Fitter

 
Changed:
<
<

Project Writeup

>
>
Maria Rovatsou, Andrew Washbrook
 
Changed:
<
<
  • Analysis Acceleration on GPUs for the ATLAS Experiment at CERN (pdf)
>
>

Project Overview

The track reconstruction in the ATLAS trigger is achieved with the utilization of the Kalman filter technique. The current implementation of the filter is a distributed sequential implementation. An acceleration of the filter would boost the whole process and would allow even more significant data to be available for processing. Apart from that, the proposed upgrade of the LHC in order to increase its performance, results in luminosity increase of one order of magnitude higher and would require a more efficient implementation of the Kalman filter. A possible way for acceleration of the filter is by porting it on a Graphics Processing Unit (GPU), which has immense computational power that can be exploited through programming models as CUDA. The proposed approach is a parallelization on thread level of the Kalman filter for execution on an NVIDIA GPU. The proposed programming model is CUDA and the analysis of its performance will be done with the CUDA profiling tools and with benchmarking tools.

Project Report

 
Changed:
<
<

Resources

>
>
  • Analysis Acceleration on GPUs for the ATLAS Experiment at CERN (pdf)
 
Changed:
<
<

GPU-based Code

>
>

Project Code

 
Changed:
<
<
  • Z-Finder GPU - Chris Jones (tar.gz)
  • Kalman Fitter SIMT Design - Maria Rovatsou (tar.gz)
  • Kalman Fitter for CUDA - Dimitry Emeliyanov (tar.gz)
>
>
  • Kalman Fitter SIMT Design (tar.gz)
  • Kalman Fitter for CUDA (Dimitry Emeliyanov) (tar.gz)
 
Changed:
<
<

Presentations

>
>

HLT Project Presentations

 
  • Processing Petabytes per Second with the ATLAS Experiment at the Large Hadron Collider in CERN, Nvidia GPU Technology Conference September 2010 (stream, pdf)
  • Algorithm Acceleration from GPGPUs for the ATLAS Upgrade, Computing in High Energy and Nuclear Physics October 2010 (pdf)
  • GPU-based tracking for ATLAS Level 2 Trigger Parallelism in Experimental Nuclear Physics Workshop January 2011 (pdf)
Changed:
<
<

ATLAS Trigger Software

>
>

HLT Resources

Software

 
Changed:
<
<

HLT resources

>
>

Articles and Preprints

 
  • Determination of the z position of primary interactions in ATLAS prior to track reconstruction (pdf) Nikos Konstantinidis and Hans Drevermann ATLAS-DAQ-2002-014
Line: 68 to 76
  Detectors and Experimental Techniques, 17 Mar 2009. - 19 p.
  • A Probabilistic Data Association Filter for fast tracking in the ATLAS Transition Radiation Tracker (pdf) D. Emeliyanov Nucl. Instrum. Meth. A566 (2006) 50-53.
Deleted:
<
<

Articles and Preprints

 
  • ATLAS Technical Design Report: High-Level Trigger, Data Acquisition and Controls (pdf) ATLAS HLT/DAQ/DCS Group ATLAS-TDR-016 2 October, 2003
  • Commissioning the ATLAS Inner Detector Trigger (pdf)
Line: 85 to 90
  Nikos Konstantinidis ATL-UPGRADE-PROC-2010-003
  • ATLAS Track Trigger for SLHC - Ideas & Plans (pdf) Nikos Konstantinidis ATL-UPGRADE-SLIDE-2009-243
Deleted:
<
<

CERN Articles and Preprints

 
Deleted:
<
<

Other Trigger Links

 
Added:
>
>

Tracking Studies

 
Changed:
<
<
>
>

Tracking of particles in Electromagnetic Field using Runge-Kutta method

 
Changed:
<
<
<!--
Related topics: TWikiUsers, TWikiGroups, TWikiAccessControl    
-->
>
>
James Henderson, Phil Clark
 
Added:
>
>

Project Overview

 
Added:
>
>
Over recent years Graphical Processing Units (GPUs) have demonstrated their great ability in scientific calculations. They have the capability to carry out tasks in parallel, enabling huge speed-ups of calculations. This project investigates the potential speed-up using GPUs when calculating a particle's trajectory through a magnetic field. The results show that, when considering many particles, a speed-up of 32x can be achieved on the GPU versus a serial processor.
 
Added:
>
>

Project Report

 
Added:
>
>
  • An Investigation Into Particle Tracking and Simulation Algorithms Using GPUs (short version) (pdf)
  • An Investigation Into Particle Tracking and Simulation Algorithms Using GPUs (long version) (pdf)
 
Added:
>
>

Project Code

 
Added:
>
>
  • RK4 stepper code used to compare timings (.cpp)
  • CUDA code for RK4 stepper (.cu)
 
Added:
>
>

Tracking Resources

Analysis Tools Studies

Analysis Acceleration in TMVA using GPU Computing

Hazel McKendrick, Andrew Washbrook

Project Overview

This project forms a feasibility study into the use of GPU (Graphics Processing Unit) devices to parallelise TMVA, the Toolkit for Multi-Variate Analysis, to determine whether such techniques might lead to future performance gains for the framework.

In particular, the Multi-layer perceptron, a class of neural network, is ported to the GPU programming platform CUDA. Performance when training single networks is generally comparable to CPU performance but show promise for future improvement. However, this parallelism of the GPU also allows multiple networks to be trained simultaneously, and this is leveraged to show significant performance gains over under- taking such a task in serial. The challenges and potential for these results to be applied across the TMVA framework is then considered and discussed.

Project Report

  • Analysis Acceleration in TMVA for the ATLAS Experiment at CERN using GPU Computing (pdf)

General GPU programming resources

Links to CUDA and other GPU parallisation resources:

This university course link has some very helpful video lectures, linked in 'Syllabus/ Lectures' and under 'Materials'.

<!--
Related topics: TWikiUsers, TWikiGroups, TWikiAccessControl    
-->
 
META FILEATTACHMENT attachment="ATLAS-Upgrade-GPGPU.pdf" attr="" comment="GPGPU and HEP" date="1260470432" name="ATLAS-Upgrade-GPGPU.pdf" path="ATLAS-Upgrade-GPGPU.pdf" size="4081455" stream="ATLAS-Upgrade-GPGPU.pdf" tmpFilename="/usr/tmp/CGItemp43556" user="wash" version="1"
META FILEATTACHMENT attachment="AlanRichardsonDissertation.pdf" attr="" comment="Alan Richardson" date="1260471141" name="AlanRichardsonDissertation.pdf" path="AlanRichardsonDissertation.pdf" size="595403" stream="AlanRichardsonDissertation.pdf" tmpFilename="/usr/tmp/CGItemp43462" user="wash" version="1"
 
This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback