BEGIN:VCALENDAR VERSION:2.0 PRODID:-//CERN//INDICO//EN BEGIN:VEVENT SUMMARY:Alignment of the ATLAS Inner Detector DTSTART;VALUE=DATE-TIME:20110905T130000Z DTEND;VALUE=DATE-TIME:20110905T132500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-24@cern.ch DESCRIPTION:Speakers: Mr. WANG\, Jike (High Energy Group-Institute of Phys ics-Academia Sinica)\nAtlas is a multipurpose experiment that records the LHC collisions. In order to reconstruct the trajectories of charged partic les\, ATLAS is equipped with a tracking system built using disticnt techno logies: silicon planar sensors (both pixel and microstrips) and drift-tube s (the Inner Detector). The tracking system is embedded in a 2 T solenoida l field. In order to reach the track parameter accuracy requested by the p hysics goals of the experiment\, the ATLAS tracking system requires to det ermine accurately its almost 700\,000 degrees of freedom. The demanded pre cision for the alignment of the silicon sensors is below 10 micrometers.\n The implementation of the track based alignment within the ATLAS software framework unifies different alignment approaches and allows the alignment of all tracking subsystems together. The alignment software counts of cour se on the tracking information (track-hit residuals) but also includes the capability to set constraints on the beam spot and primary vertex for the global positioning\, plus constraints on the track parameters as the mome ntum measured by the Muon System or the E/p using the calorimetry informat ion. The assembly survey data can be used as constraint to the alignment c orrections.\nThe alignment chain starts at the trigger level where a strea m of high pT and isolated tracks is selected online. Also a cosmic ray tri gger is enabled while ATLAS is recording collision data\, but only during those short intervals where there are no LHC beams inside ATLAS. Thus a st ream of cosmic-ray tracks is recorded exactly with the same detector opera ting conditions as the normal collision tracks.\nAs the alignment algorith ms are based on the minimization of the track-hit residuals\, one needs to solve a linear system with large number of degrees of freedom. The solvin g involves the inversion or diagonalization of a large matrix that may be dense. The alignment jobs can be executed either at the CERN Analysis Faci lity or using the GRID infrastructure. The event processing is run in para llel in many jobs (for both collision data and cosmic ray tracks). Then al l output matrices and vectors are added together before the linear algebra solving. The alignment procedure can also be run either offline (to repro cess old data) or quasi-online at the Tier0 in the calibration loop. With the latter alignment constants are computed before the bulk reconstruction of the ATLAS data.\nWe will present results of the alignment of the ATLAS tracker using the 2011 collision data. The validation of the alignment is performed first using its own observables (track-hit residuals) as well a s using many other physics observables\, notably the resonance invariant m asses in a wide energy range (K0s\, J/ψ and Z decays in to μ+μ-) and th e effect of the detector systematic distortions on the reconstructed invar iant mass and on the μ momentum. Also the electrons E/p has been studied mainly in the W→ eν channel. The results of the alignment with real dat a reveals that the attained precision for the alignment parameters is appr oximately 5 micrometers.\n\nhttp://indico.cern.ch/contributionDisplay.py?c ontribId=24&sessionId=2&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=24&sessionId=2& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Visual Physics Analysis (VISPA) - From Desktop Towards Physics Ana lysis at Your Fingertips DTSTART;VALUE=DATE-TIME:20110905T155500Z DTEND;VALUE=DATE-TIME:20110905T162000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-27@cern.ch DESCRIPTION:Speakers: FISCHER\, Robert (RWTH Aachen University\, III. Phys ikalisches Institut A)\nVisual Physics Analysis (VISPA) is an analysis dev elopment environment with applications in high energy as well as astropart icle physics. VISPA provides a graphical steering of the analysis flow\, w hich is comprised of self-written C++ and Python modules. The advances pre sented in this talk extend the scope from prototyping to the execution of analyses. A novel concept of analysis layers has been integrated in VISPA. On top of a base layer\, it is possible to derive additional layers in wh ich options are adjustable and modules can be activated or deactivated. Th is enables the creation of different stages already within the design phas e of a single analysis\, e.g. the event selection and the statistical anal ysis\, or the optimization of settings for different types of input data s uch as electrons and muons which are to be processed within the same analy sis flow. Furthermore\, analysis execution in VISPA has been extended to i nclude a graphical interface for parameter sets that are handled within a back-end independent design. This allows for direct job submission from VI SPA to local computing clusters as well as to the LHC Computing Grid.\n\nh ttp://indico.cern.ch/contributionDisplay.py?contribId=27&sessionId=2&confI d=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=27&sessionId=2& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Online Measurement of LHC Beam Parameters with the ATLAS High Leve l Trigger DTSTART;VALUE=DATE-TIME:20110905T162000Z DTEND;VALUE=DATE-TIME:20110905T164500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-20@cern.ch DESCRIPTION:Speakers: STRAUSS\, Emanuel Alexandre (SLAC National Accelerat or Laboratory)\nWe present an online measurement of the LHC beam paramete rs in ATLAS using the High Level Trigger (HLT). When a significant change is detected in the measured beamspot\, it is distributed to the HLT. The re\, trigger algorithms like b-tagging which calculate impact parameters o r decay lengths benefit from a precise\,up-to-date set of beamspot paramet ers. Additionally\, online feedback is sent to the LHC operators in real t ime. The measurement is performed by an algorithm running on the Level 2 trigger farm\, leveraging the high rate of usable events. Dedicated algori thms perform a full scan of the silicon detector to reconstruct event vert ices from registered tracks. The distribution of these vertices is aggrega ted across the farm and their shape is extracted through fits every 60 sec onds to determine the beamspot position\, size\, and tilt. The reconstruct ed beam values are corrected for detector resolution effects\, measured in situ using the separation of vertices whose tracks have been split into t wo collections. Furthermore\, measurements for individual bunch crossings have allowed for studies of single-bunch distributions as well as the beha vior of bunch trains. This talk will cover the constraints imposed by the online environment and describe how these measurements are accomplished wi th the given resources. The algorithm tasks must be completed within the t ime constraints of the Level 2 trigger\, with limited CPU and bandwidth al locations. This places an emphasis on efficient algorithm design and the m inimization of data requests.\n\nhttp://indico.cern.ch/contributionDisplay .py?contribId=20&sessionId=1&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=20&sessionId=1& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Fractal dimension analysis in a highly granular calorimeter DTSTART;VALUE=DATE-TIME:20110905T153000Z DTEND;VALUE=DATE-TIME:20110905T155500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-22@cern.ch DESCRIPTION:Speakers: Dr. RUAN\, Manqi (Laboratoire Leprince-Ringuet (LLR) -Ecole Polytechnique)\nThe concept of "particle flow" has been developed t o optimise jet energy resolution by best separating the different componen ts of hadronic jets. A highly granular calorimetry is mandatory and provid es an unprecedented level of detail in the reconstruction of showers. This enables new approaches to shower analysis. Here the measurement and use o f of showers' fractal dimension is \ndescribed. \n The fractal dimensio n is a characteristic number that measures the global density of the showe r. This property is highly dependent on the type of interaction and the pa rticle energy. Its use in identifying particles and estimating their energ y is described \nin the context of the semi-digital hadron calorimeter for the ILD concept (International Large Detector for the International Linea r Collider)\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=22&s essionId=2&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=22&sessionId=2& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:The AAL project: Automated monitoring and intelligent AnaLysis for the ATLAS data taking infrastructure DTSTART;VALUE=DATE-TIME:20110905T153000Z DTEND;VALUE=DATE-TIME:20110905T155500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-23@cern.ch DESCRIPTION:Speakers: Mr. MAGNONI\, Luca (Conseil Europeen Recherche Nucl. (CERN))\nThe Trigger and Data Acquisition (TDAQ) system of the ATLAS expe riment at CERN is the infrastructure responsible for filtering and transfe rring ATLAS experimental data from detectors to the mass storage system. I t relies on a large\, distributed computing environment\, including thousa nds of computing nodes with thousands of application running concurrently. \nIn such a complex environment\, information analysis is fundamental for controlling applications behavior\, error reporting and operational monito ring. During data taking runs\, streams of messages sent by applications v ia the message reporting system together with data published from applicat ions via information services are the main sources of knowledge about corr ectness of running operations. The huge flow of data produced (with an ave rage rate of O(1-10KHz)) is constantly monitored by experts to detect pro blem or misbehavior. This require strong competence and experience in unde rstanding and discovering problems and root causes\, and often the meaning ful information is not in the single message or update\, but in the aggreg ated behavior in a certain time-line.\nThe AAL project is meant at reducin g the man power needs and at assuring a constant high quality of problem d etection by automating most of the monitoring tasks and providing real-tim e correlation of data-taking and system metrics.\nThis project combines te chnologies coming from different disciplines\, in particular it leverages on an Event Driven Architecture to unify the flow of data from the ATLAS i nfrastructure\, on a Complex Event Processing (CEP) engine for correlation of events and on a machine learning module to detect anomaly and problems that cannot be defined in advance. \nThe project is composed of 3 main co mponents: a core processing engine\, responsible for correlation of events through expert-defined queries\, a machine learning module to detect anom alies in an unsupervised manner and a web based front-end to present real- time information and interact with the system. All components works in a l oose-coupled event based architecture\, with a message broker to centraliz e all communication between modules.\nThe result is an intelligent system able to extract and compute relevant information from the flow of operatio nal data to provide real-time feedback to human experts who can promptly r eact when needed. The paper presents the design and implementation of the AAL project\, together with the results of its usage as automated monitori ng assistant for the ATLAS data taking infrastructure.\n\nhttp://indico.ce rn.ch/contributionDisplay.py?contribId=23&sessionId=1&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=23&sessionId=1& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Application of Symbolic Regression to Mass Measurement in H->WW Di lepton Channels DTSTART;VALUE=DATE-TIME:20110908T130000Z DTEND;VALUE=DATE-TIME:20110908T132500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-4@cern.ch DESCRIPTION:Speakers: CHOI\, Su Yong (Korea University)\nWe derive a kinem atic variable that is sensitive to the mass of the Standard Model Higgs bo son (M_H) in the H->WW*->l l nu nu-bar channel using symbolic regression m ethod. Explicit mass reconstruction is not possible in this channel due to the presence of two neutrinos which escape detection. Mass determination problem is that of finding a mass-sensitive function that depends on the m easured observables. We use symbolic regression\, which is an analytical a pproach to the problem of non-linear regression\, to derive an analytic fo rmula sensitive to M_H from the two lepton momenta and the missing transve rse momentum. Using the newly-derived mass-sensitive variable\, we expect Higgs mass resolutions between 1 to 4 GeV for M_H between 130 and 190 GeV at the LHC with 10 fb^-1 of data.\n\nhttp://indico.cern.ch/contributionDis play.py?contribId=4&sessionId=12&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=4&sessionId=12& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:SALAMI project DTSTART;VALUE=DATE-TIME:20110908T103000Z DTEND;VALUE=DATE-TIME:20110908T111000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-120@cern.ch DESCRIPTION:Speakers: Prof. DE ROURE\, David (Oxford e-Research Centre)\nh ttp://indico.cern.ch/contributionDisplay.py?contribId=120&sessionId=10&con fId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=120&sessionId=1 0&confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Building an Outsourcing Ecosystem for Science DTSTART;VALUE=DATE-TIME:20110905T101000Z DTEND;VALUE=DATE-TIME:20110905T105000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-121@cern.ch DESCRIPTION:Speakers: Dr. KEAHEY\, Kate (Argonne National Laboratory)\nInf rastructure-as-a-Service (IaaS) cloud computing is revolutionizing the way we acquire and manage computational and storage resources: by allowing on -demand resource leases and supporting user control over those resources i t enables us to treat resource acquisition as an operational consideration rather than capital investment. The emergence of this new model raises ma ny questions\, in particular for special requirements groups such as scien tific computing. Can cloud computing be used by scientific applications? D oes it\, or will it ever\, provide sufficient capabilities for high-perfor mance applications? How will it change our work patterns? What challenges need to be overcome\, and what is its overall potential for accelerating s cience?\n\nIn this talk\, I will give an overview of the challenges and po tential of cloud computing projects in scientific community. I will descri be what attracted various scientific communities to cloud computing\, give examples of how they integrated this new model into their work\, and desc ribe the challenges they encountered while doing so. I will then discuss h ow those challenges drove the development of Nimbus Infrastructure\, which allows users to provide cycle outsourcing via their clouds\, as well as t he Nimbus Platform\, which provides ecosystem tools allowing users to leve rage infrastructure cloud resources across different academic and commerci al platforms ranging from proporitary (Amazon Web Services) to open source (Nimbus\, OpenStack\, Eucalyptus and others). I will also discuss challen ges and issues – related to performance\, logistics\, utilization\, and privacy that need to be overcome to make the benefits of cloud computing a vailable to an ever larger set of scientific applications. Finally\, I wil l discuss the emerging technology trends and discuss how they can benefit science.\n\nBio: Kate Keahey is a Scientist in the Distributed Systems Lab at Argonne National Laboratory and a Fellow at the Computation Institute at the University of Chicago. Kate pioneered the use of cloud computing fo r scientific applications and created and leads the open source Nimbus pro ject which provides an Infrastructure-as-a-Service cloud computing impleme ntation as well as a set of higher-level services allowing users to build elastic application by combining on-demand commercial and scientific cloud resources.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=121& sessionId=0&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=121&sessionId=0 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Summary - Computing Technology for Physics Research DTSTART;VALUE=DATE-TIME:20110909T080000Z DTEND;VALUE=DATE-TIME:20110909T084000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-122@cern.ch DESCRIPTION:Speakers: Dr. LAURET\, Jerome (BNL)\nhttp://indico.cern.ch/con tributionDisplay.py?contribId=122&sessionId=15&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=122&sessionId=1 5&confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Summary - Data Analysis – Algorithms and Tools DTSTART;VALUE=DATE-TIME:20110909T084000Z DTEND;VALUE=DATE-TIME:20110909T092000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-123@cern.ch DESCRIPTION:Speakers: BHAT\, Pushpalatha (Fermi National Accelerator Lab. (Fermilab))\nhttp://indico.cern.ch/contributionDisplay.py?contribId=123&se ssionId=15&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=123&sessionId=1 5&confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Summary - Computations in Theoretical Physics – Techniques and M ethods DTSTART;VALUE=DATE-TIME:20110909T095000Z DTEND;VALUE=DATE-TIME:20110909T103000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-124@cern.ch DESCRIPTION:Speakers: GLOVER\, Nigel (IPPP Durham)\nhttp://indico.cern.ch/ contributionDisplay.py?contribId=124&sessionId=15&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=124&sessionId=1 5&confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:ACAT 2011 - Summary and outlook DTSTART;VALUE=DATE-TIME:20110909T103000Z DTEND;VALUE=DATE-TIME:20110909T110000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-125@cern.ch DESCRIPTION:Speakers: Dr. PERRET-GALLIX\, Denis (CNRS/IN2P3)\nhttp://indic o.cern.ch/contributionDisplay.py?contribId=125&sessionId=15&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=125&sessionId=1 5&confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Welcome - Prof. Geoff Rodgers\, Pro-Vice Chancellor for Research\, Brunel University DTSTART;VALUE=DATE-TIME:20110905T084000Z DTEND;VALUE=DATE-TIME:20110905T090000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-126@cern.ch DESCRIPTION:Speakers: Prof. RODGERS\, Geoff (Brunel University)\nhttp://in dico.cern.ch/contributionDisplay.py?contribId=126&sessionId=0&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=126&sessionId=0 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Offloading peak processing to Virtual Farm by STAR experiment at R HIC DTSTART;VALUE=DATE-TIME:20110906T160000Z DTEND;VALUE=DATE-TIME:20110906T162500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-58@cern.ch DESCRIPTION:Speakers: Dr. BALEWSKI\, jan (MIT)\nIn recent years\, Cloud co mputing has become a very attractive “notion” and\npopular model for a ccessing distributed resources and has emerged as the next\nbig trend afte r the so-called Grid computing approach.\nThe onsite STAR computing resour ces amounting to about 3000 CPU slots have\nbeen extended by additional 10 00 slots using opportunistic resources from pilot\nDOE/Magellan and DOE/Ni mbus projects.\nThe Virtual Machine (VM) framework was used to assemble th e STAR-computing\nenvironment\, which is independent on specific hardware. STAR VM was\nvalidated once\, deployed on over 100 8-core VMs at NERSC an d Argon National\nLab\, and used as homogenous Virtual Farm processing in real time events\nacquired by STAR detector located at Brookhaven National Lab. To provide time\ndependent calibration constants to the large number of isolated VMs\, a database\nsnapshot scheme was devised and used for th is exercise. It allows periodic\nsynchronization of VM DB with the master DB without the need for frequent DB\nclient connections to the master DB f rom multiple jobs running on every VM. The\ntwo high capacity disks locali zed at the opposite coasts of US and interconnected\nvia Globus-Online pro tocol were used in this setup\, which resulted with highly\nscalable Cloud -based extension of STAR computing resources.\nThe STAR Virtual Farm scale d up between February and May of 2011 from 160\nto 1300 CPU slots. It has been used to process fraction of events STAR in real\ntime and later to re analyze past STAR events to providing key arguments for\nchanging the cour se of ongoing STAR data taking\n\nhttp://indico.cern.ch/contributionDispla y.py?contribId=58&sessionId=6&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=58&sessionId=6& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Challenges in using GPUs for the reconstruction of digital hologra m images. DTSTART;VALUE=DATE-TIME:20110906T153500Z DTEND;VALUE=DATE-TIME:20110906T160000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-55@cern.ch DESCRIPTION:Speakers: Prof. HOBSON\, Peter R (Brunel University)\nIn-line holography has recently made the transition from silver-halide based recor ding media\, with laser reconstruction\, to recording with large-area pixe l detectors and computer-based reconstruction. This form of holographic im aging is used for small particulates\, such as cloud or fuel droplets\, ma rine plankton and alluvial sediments\, and enables a true 3D object field to be recorded at high resolution over a considerable depth. To reconstruc t a digital hologram a 2D FFT must be calculated for every depth slice des ired in the replayed image volume. A typical hologram of ~100 micrometre p articles over a depth of a few hundred millimetres will require O(1000) 2D FFT operations to be performed on an hologram of typically a few million pixels. With the growing use of video-rate recording and the desire to rec onstruct fully every frame the computational challenge becomes considerabl e. In previous work (http://bura.brunel.ac.uk/handle/2438/2823) we have re ported on our experiences with reconstruction on a computational grid. In this paper we discuss the technical challenges in converting our reconstru ction code to make efficient use of the NVIDIA CUDA based GPU cards and sh ow how near real-time video slice reconstruction can be obtained with holo grams as large as 4K by 4K pixels. We also discuss the issues surrounding the reconstruction of holograms which are larger than 50% of the GPU memor y where a different approach to reconstruction will be needed. Finally we consider the implications for grid and cloud computing\, and the extent t o which GPU can replace these approaches\, when the important step of loca ting focussed objects within a reconstructed volume is included.\n\nhttp:/ /indico.cern.ch/contributionDisplay.py?contribId=55&sessionId=6&confId=938 77 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=55&sessionId=6& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Gibbs sampler for background discrimination in particle physics DTSTART;VALUE=DATE-TIME:20110906T151000Z DTEND;VALUE=DATE-TIME:20110906T153500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-54@cern.ch DESCRIPTION:Speakers: Dr. COLECCHIA\, Federico (University College London) \nBackground properties in experimental particle physics are typically est imated from large collections of events. This usually provides precise kno wledge of average background distributions\, but inevitably hides fluctuat ions. To overcome this limitation\, an approach based on statistical mixtu re model decomposition is presented. Events are treated as heterogeneous p opulations comprising particles originating from different processes\, and individual particles are mapped to a process of interest on a probabilist ic basis. When used to discriminate against background\, the proposed tech nique based on the Gibbs sampler allows some features of the background di stributions to be estimated directly from the data without training on hig h-statistics samples. A feasibility study on Monte Carlo is presented\, to gether with a comparison with existing techniques. Finally\, the prospects for the development of the Gibbs sampler into a tool for intensive offlin e analysis of interesting events at the Large Hadron Collider are discusse d.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=54&sessionId= 7&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=54&sessionId=7& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Track finding using GPUs DTSTART;VALUE=DATE-TIME:20110906T151000Z DTEND;VALUE=DATE-TIME:20110906T153500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-51@cern.ch DESCRIPTION:Speakers: Dr. SCHMITT\, Christian (Institut fuer Physik-Johann es-Gutenberg-Universitaet Mainz)\nThe reconstruction and simulation of co llision events is a major task\nin modern HEP experiments involving severa l ten thousands of \nstandard CPUs. On the other hand the graphics process ors (GPUs) have \nbecome much more powerful and are by far outperforming t he standard\nCPUs in terms of floating point operations due to their massi ve\nparallel approach. The usage of these GPUs could therefore\nsignifican tly reduce the overall reconstruction time per event or allow for the\nusa ge of more sophisticated algorithms.\n\nIn this contribution the track fin ding in the ATLAS experiment will be\nused as an example on how the GPUs c an be used in this context: the\nseed finding alone shows already a speed increase of one order of magnitude\ncompared to the same implementation on a standard CPU. On the other\nhand the implementation on the GPU requires a change in the\nalgorithmic flow to allow the code to work in the rather limited\nenvironment on the GPU in terms of memory\, cache\, and transfer speed\nfrom and to the GPU.\n\nhttp://indico.cern.ch/contributionDisplay. py?contribId=51&sessionId=6&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=51&sessionId=6& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:One-loop integrations with Hypergeometric functions DTSTART;VALUE=DATE-TIME:20110908T141500Z DTEND;VALUE=DATE-TIME:20110908T144000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-52@cern.ch DESCRIPTION:Speakers: Prof. KANEKO\, Toshiaki (KEK)\nNumerically stable an alytic expression of a one-loop integration\nis one of the most important elements of the accurate\ncalculations of one-loop corrections to the phys ical processes.\nIt is known that these integrations are expressed by some \ngeneralized classes of Gauss hypergeometric functions. Power\nseries ex pansions\, differential equations\, contiguous and many\nother identities are known for them. For Lauricella $F_D$\nfunctions\, analytic properties are studied in detail\, which\nprovide useful information for the numeric al stabilities.\n\nWe show that two- and three-point functions are exactly expressed\nin terms of $F_D$ for arbitrary combinations of mass parameter s\nin any space-time dimensions. We also show the relation between\nfour- point functions and Aomoto-Gelfand hypergeometric functions.\n\nhttp://ind ico.cern.ch/contributionDisplay.py?contribId=52&sessionId=13&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=52&sessionId=13 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:NFS 4.1/pNFS\, the final step DTSTART;VALUE=DATE-TIME:20110906T080000Z DTEND;VALUE=DATE-TIME:20110906T084000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-115@cern.ch DESCRIPTION:Speakers: FUHRMANN\, Patrick (DESY)\nWith the introduction of clustered storage\, combining a set of hosts to a single storage system\, a very successful standard data access protocol\, NFS2/3 became obsolete. One of the reasons was that NFS 2/3 assumes the name service part of the p rotocol being severed from the same host as the actual data\, which is of course no longer true for clustered systems. As a result\, high performanc e storage systems e.g. Panasas\, GPFS\, Lustre and many more\, designed th eir own file system network protocols\, with the obvious advantage of an e xtremely optimized use of the underlying network and storage resources\, a s the server and client software are provided by the same source. The draw backs however were that proprietary software had to be installed on all cl ient machines\, with the hassle of kernel and driver dependencies and main tenance issues\, particularly annoying when operating large compute farms. In order to catch up on that development\, well-known storage providers d ecided to invest into a standard network file system protocol supporting c lustered storage services\, the Parallel Network File System (pNFS). The a ctivity is organized by the Center for Information Technology Integration (CITI) at the University of Michigan. At the time being\, all partners in this group have the NFS 4.1/pNFS server software integrated into their st orage systems\, however\, except for dCache.org\, companies seem to be rel uctant making it available to customers. NFS 4.1/pNFS client drivers are a vailable for the Linux 2.6.38 kernel and are slowly approaching standard Linux distributions. This presentation will elaborate on the advantages of NFS 4.1/pNFS as well as on the availability of the different components a nd possibly on missing bits and pieces. Furthermore it will provide detail s on the stability and performance evaluation done in the context of the E uropean Middleware Initiative (EMI) and at dCache.org.\n\nhttp://indico.ce rn.ch/contributionDisplay.py?contribId=115&sessionId=5&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=115&sessionId=5 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:New approaches for numerical techniques in higher order calculatio ns DTSTART;VALUE=DATE-TIME:20110908T095000Z DTEND;VALUE=DATE-TIME:20110908T103000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-114@cern.ch DESCRIPTION:Speakers: Prof. CZAKON\, Michal (RWTH Aachen)\nIt has become c ustomary to think of higher order calculations as analytic\, in the sense that the result should be presented in the form of known functions or cons tants. If such a result is obtained\, numerical evaluation for practical a pplications or expansion in asymptotic regimes should not pose any problem . There are\, however\, many problems of interest\, where the analytic str ucture\, due to the number of involved variables\, does not make it possib le to express predictions through known functions. One strategy is to ext end the class of functions\, as for example in the case of harmonic and ge neralized harmonic polylogarithms. On the other hand\, if the aim is to pr ovide results quickly and with moderate effort\, then there are much more efficient approaches\, which involve numerical methods at earlier stages o f the calculation. In this talk\, I will review methods for the evaluation of virtual corrections\, such as contour deformation in Feynman-parametri c and Mellin-Barnes representations\, as well as the method of differentia l equations. Subsequently\, I will present recent advances in the calculat ion of real radiation contributions with non-analytic evaluation of integr als over the unresolved phase space.\n\nhttp://indico.cern.ch/contribution Display.py?contribId=114&sessionId=10&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=114&sessionId=1 0&confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:GPU Linear algebra extensions for GNU/Octave DTSTART;VALUE=DATE-TIME:20110908T153500Z DTEND;VALUE=DATE-TIME:20110908T160000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-89@cern.ch DESCRIPTION:Speakers: Dr. SANTOCCHIA\, Attilio (Universita e INFN Perugia) \nOctave is one if the most used open source tools for numerical analysis\ nand liner algebra. Our project wants to improve Octave introducing the su pport for GPU computing\, in order to speed up some linear algebra operati ons. The core of our work is a C library that executes on GPU some BLAS op erations concerning vector-vector\, vector-matrix and matrix-matrix functi ons. OpenCL functions are used to program GPU kernels\, which are bind wit hin the GNU/octave framework. We report the project implementation desing and some preliminary results about performances.\n\nhttp://indico.cern.ch/ contributionDisplay.py?contribId=89&sessionId=13&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=89&sessionId=13 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Polynomial Algebra in Form 4 DTSTART;VALUE=DATE-TIME:20110906T141500Z DTEND;VALUE=DATE-TIME:20110906T144000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-111@cern.ch DESCRIPTION:Speakers: KUIPERS\, Jan (Nikhef)\nNew features of the symbolic algebra package Form 4 are\ndiscussed. Most importantly\, these features include polynomial\nfactorization and polynomial GCD computation. Examples of\ntheir use are shown. One of them is an exact version of Mincer which\ ngives answers in terms of rational polynomials and 5 master integrals.\n\ nhttp://indico.cern.ch/contributionDisplay.py?contribId=111&sessionId=8&co nfId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=111&sessionId=8 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Automated one-loop calculations with Golem/Samurai DTSTART;VALUE=DATE-TIME:20110908T151000Z DTEND;VALUE=DATE-TIME:20110908T153500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-110@cern.ch DESCRIPTION:Speakers: HEINRICH\, Gudrun (Max Planck Institute Munich)\nA p rogram package will be presented which aims at the automated calculation o f \none-loop amplitudes for multi-particle processes.\nThe program offers the possibility to optionally use either unitarity cuts \nor traditional t ensor reduction of Feynman diagrams\, or a combination of both.\nIt can be used to calculate one-loop corrections to both QCD and electro-weak theor y. \nBeyond the Standard Model theories can be interfaced using FeynRules or LanHep.\nA standard interface to programs calculating real radiation is also included. \nIt will further be described how the program detects and deals with numerical instabilities\, \nand how the rational terms can be computed efficiently.\n\nhttp://indico.cern.ch/contributionDisplay.py?cont ribId=110&sessionId=13&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=110&sessionId=1 3&confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Reduze 2 DTSTART;VALUE=DATE-TIME:20110906T135000Z DTEND;VALUE=DATE-TIME:20110906T141500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-113@cern.ch DESCRIPTION:Speakers: Dr. STUDERUS\, Cedric (University of Bielefeld)\nRed uze is a computer program for reducing Feynman Integrals to master integra ls employing the Gauss/Laporta algorithm. Reduze is written in C++ and use s the GiNaC library to perform simplifications of the algebraic prefactors in the system of equations.\nIn this talk\, the new version\, Reduze 2\, is presented. The program supports fully parallelised computations with MP I and allows to resume aborted reductions with the use of the Berkeley dat abase. The user inputs are standardized with the YAML file format. Reduze 2 also provides an interface to use the computer algebra system Fermat.\n\ nhttp://indico.cern.ch/contributionDisplay.py?contribId=113&sessionId=8&co nfId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=113&sessionId=8 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:An analytical solution for a non-planar massive double box diagram DTSTART;VALUE=DATE-TIME:20110905T141500Z DTEND;VALUE=DATE-TIME:20110905T144000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-112@cern.ch DESCRIPTION:Speakers: Mr. VON MANTEUFFEL\, Andreas (University of Zurich)\ nAn analytical calculation of a non-planar 2-loop box diagram is presented .\nThis diagram appears in the computation of higher order corrections to top-\nquark pair production and contains one internal massive line. The \n corresponding integrals are solved with differential equation and Mellin-B arnes \ntechniques.\n\nhttp://indico.cern.ch/contributionDisplay.py?contri bId=112&sessionId=3&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=112&sessionId=3 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Multicore in Production: Advantages and Limits of the Multi-proces s Approach. DTSTART;VALUE=DATE-TIME:20110908T130000Z DTEND;VALUE=DATE-TIME:20110908T132500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-83@cern.ch DESCRIPTION:Speakers: TSULAIA\, Vakhtang (LBL)\nThe shared memory architec ture of multicore CPUs provides HENP developers with the opportunity to re duce the memory footprint of their applications by sharing memory pages be tween the cores in a processor. ATLAS pioneered the multi-process approach to parallelizing HENP applications. Using Linux fork() and the Copy On Wr ite mechanism we implemented a simple event task farm which allows to shar e up to 50% memory pages among event worker processes with negligible CPU overhead. \n\nBy leaving the task of managing shared memory pages to the o perating system\, we have been able to run in parallel large reconstructio n and simulation applications originally written to be run in a single thr ead of execution with little to no change to the application code. In spit e of this\, the process of validating athena multi-process for production took ten months of concentrated effort and is expected to continue for sev eral more months. In general terms\, we had two classes of problems in the multi-process port: merging the output files produced by the event worke rs\, and assuring the reproducibility of the results\, especially of Monte carlo simulations\, when running with different configurations\, in partic ular with different number of event workers.\n\nBesides validating the sof tware itself\, an important and time-consuming aspect of running multicore applications in production is to configure the production system to handl e multicore jobs. This entails defining multicore batch queues\, where the unit resource is not a core\, but a whole computing node\; monitoring the output of many event workers\; and adapting the job definition layer to h andle computing resources with very different event throughputs (depending on the number of cores used).\n\nTo conclude\, we will present scalabilit y and memory usage studies\, based on data gathered both on dedicated hard ware and on ATLAS production nodes. From these it should become apparent t hat the most promising development to improve performance will be to trans ition from a simple\, flat\, event task farm in which all processes handle events independently to a task farm with specialized worker processes\, w hich will be in charge of event I/O. This approach will further reduce the memory footprint of our multicore applications\, and at the same time add ress the issue of merging event worker outputs\, at the cost of some incre ase in the complexity of the ATLAS core software.\n\nhttp://indico.cern.ch /contributionDisplay.py?contribId=83&sessionId=11&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=83&sessionId=11 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Using machine learning techniques in classification problems in As trophysics DTSTART;VALUE=DATE-TIME:20110906T103000Z DTEND;VALUE=DATE-TIME:20110906T111000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-119@cern.ch DESCRIPTION:Speakers: Dr. RAYCHAUDHURY\, Somak (University of Birmingham)\ nMultivariate datasets in astrophysics can be large\, with the\nincreasing volume of information now becoming available from a range\nof observation s\, from ground and Space\, across the electromagnetic\nspectrum. The obs ervations are in the form of raw images and/or\nspectra\, and tables of de rived quantities\, obtained at multiple epochs\nin time. Large archives of images\, spectra and catalogues are now\nbeing assembled into publicly-av ailable databases: one example is the\nemerging global effort towards the Virtual Observatory. This\nnecessitates the development of techniques that will allow fast\,\nautomated classification and extraction of key physica l properties for\nvery large datasets\, and the ability to visualise the s tructure of\nhighly multi-dimensional data\, for extracting and studying\n substructures in a flexible way. Automated algorithms for clustering and o utlier\ndetection are necessary for a wide range of Astrophysical\nproblem s involving these growing datasets.\nThe applicability of commercial data mining tools is\nlimited\, since these do not incorporate the handling of errors in a\nprincipled manner\, which is central to the analysis of Astro nomical\ndata\, as it is in other branches of Physics. I will review how\n techniques used in the field of machine learning are being adapted for\nus e in classification and clustering problems. Examples will include\nthe us e of topographic mapping to classify light curves of eclipsing\nbinary sta rs\, showing that this is an efficient way of searching for\ntransiting ex trasolar planets in large datasets\, and robust density\nmodelling for det ermining clusters and outliers\, resulting in finding\nhigh-redshift quasa rs.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=119&sessionI d=5&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=119&sessionId=5 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Online Particle Detection by Neural Networks Based on Topologic Ca lorimetry Information DTSTART;VALUE=DATE-TIME:20110908T141500Z DTEND;VALUE=DATE-TIME:20110908T144000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-87@cern.ch DESCRIPTION:Speakers: Mr. DE SEIXAS\, José Manoel (Univ. Federal do Rio d e Janeiro (UFRJ))\nElectrons and photons are among the most important sign atures in ATLAS. Their identification against jets background by the onlin e trigger system relies very much on calorimetry information. The ATLAS on line trigger comprises three cascaded levels and the Ringer is an alternat ive set of algorithms that uses calorimetry information for electron detec tion at the second trigger level (L2). It is split into two parts: the fea ture extraction algorithm (FEX)\, which represents particle interaction as a set of concentric ring sums\, and the hypothesis test (HYPO)\, which im plements a multilayer perceptron neural network to perform final particle identification. The neural network may also be used to implement a Fisher discriminant\, in case linear processing is desired in this stage.\n\nThe Ringer FEX starts by searching the most energetic cell (hot cell) in each calorimeter layer from the Region of Interest (RoI) previously selected by the ATLAS level-1 trigger. The hot cell energy becomes the first ring and it is also considered the center of all further rings\, which are formed as the sum of the energies from the outer cells of the inner ring. A total of 100 rings are computed. The Ringer HYPO normalizes the ring values in order to fit them to the neural network dynamic range. After propagating t he rings through the network\, a single output node provides the incoming event classification.\n\nOptimizations\, guided by detailed time performan ce analysis\, were made at the Ringer algorithm core\, in order to make it prepared for operation in ATLAS. Studies showed that the execution time w as improved by a factor of 50\, while its payload necessary to store the R inger information represents only 1.2% of the present total HLT amount. Al so\, Monte Carlo simulations of 14 TeV proton-proton collisions at 2x10^34 luminosity were used to evaluate the Ringer performance over pile-up.\n\n http://indico.cern.ch/contributionDisplay.py?contribId=87&sessionId=12&con fId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=87&sessionId=12 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:The toolbox of modern multi-loop computations: novel analytic and semi-analytic techniques DTSTART;VALUE=DATE-TIME:20110905T105000Z DTEND;VALUE=DATE-TIME:20110905T113000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-84@cern.ch DESCRIPTION:Speakers: Dr. PAK\, Alexey (TTP KIT Karlsruhe)\nAfter a short introduction\, sketching the structure of a typical calculation of higher- order quantum corrections\, I will discuss a few examples illustrating ide as that were instrumental in obtaining some recent novel results. Attentio n will be given to the tools facilitating those techniques and the techn ical challenges. In particular\, the talk will cover the progress in secto r decomposition method\, gluing relations\, and dimensional recurrence r elations. Finally\, I will mention some very promising theoretical develop ments in understanding the mathematical structure of Feynman integrals th at are yet to yield new results.\n\nhttp://indico.cern.ch/contributionDisp lay.py?contribId=84&sessionId=0&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=84&sessionId=0& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:One-loop tensor Feynman integral reduction with signed minors DTSTART;VALUE=DATE-TIME:20110908T135000Z DTEND;VALUE=DATE-TIME:20110908T141500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-85@cern.ch DESCRIPTION:Speakers: RIEMANN\, Tord (DESY)\nThe algebraic tensor reductio n of one-loop Feynman integrals with signed minors has been further develo ped.\nThere is now available the\nC++ package PJFry by V. Yundin for the r eduction of 5-point 1-loop tensor integrals up to rank 5.\nSpecial care is devoted to vanishing or small Gram determinants.\nFurther\, we derived \n extremely compact expressions for the contractions of the tensor integrals with external momenta. They are based on sums over signed minors weighted with scalar products of the external momenta.\n\nhttp://indico.cern.ch/co ntributionDisplay.py?contribId=85&sessionId=13&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=85&sessionId=13 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Multiloop calculations in supersymmetric theories with the higher covariant derivative regularization DTSTART;VALUE=DATE-TIME:20110905T132500Z DTEND;VALUE=DATE-TIME:20110905T135000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-3@cern.ch DESCRIPTION:Speakers: Dr. STEPANYANTZ\, konstantin (Moscow State Universit y)\nMost calculations of quantum correction in the supersymmetric theories are made with the dimensional reduction\, which is a modification of the dimensional regularization. However\, it is well known that the dimensiona l reduction is not self-consistent. A consistent regularization\, which do es not break the supersymmetry is the higher covariant derivative regulari zation. However\, the integrals obtained with this regularization can not be usually calculated analytically. We discuss application of this regular ization to the calculations in supersymmetric theories. In particular\, it is demonstrated that integrals defining the beta-function are possibly in tegrals of total derivatives. This feature allows to explain the origin of the exact NSVZ beta-function\, relating the beta-function with the anomal ous dimensions of the matter superfields. However\, integrals for the anom alous dimension should calculated numerically.\n\nhttp://indico.cern.ch/co ntributionDisplay.py?contribId=3&sessionId=3&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=3&sessionId=3&c onfId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:HYPERDIRE: HYPERgeometric DIfferential REduction HYPERDIRE: HYPE Rgeometric DIfferential REduction Mathematica based programs for differe ntial reduction of hypergeometric functions and its application to F eynman Diagrams Calculation. DTSTART;VALUE=DATE-TIME:20110906T130000Z DTEND;VALUE=DATE-TIME:20110906T132500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-7@cern.ch DESCRIPTION:Speakers: Dr. VLADIMIR\, Bytev (JINR)\nThe differential reduct ion algorithm allows to change the values \n of parameters of any Horn-typ e hypergeometric functions on arbitrary \n integers numbers. The descripti on of mathematical part of algorithm \n have been presented on ACAT08 by M.Kalmykov [6]. \n We will describe the status of project and will present a new version\n of MATHEMATICA based package including a several importa nt hypergeometric \n functions of one and two variables.\n\n Interrelation between Differential Reduction algorithm \n and Integration-by-Parts tech nique is discussed.\n We illustrate the procedure in the context of \n gen eralized hypergeometric functions\, and give an \n example for a type of b ubble and propagator type diagram. \n Another application of HYPERDIRE is the construction of \n epsilon-expansion of Horn-type Hypergeometric Funct ions.\n\n\n Talk is based on the following publications:\n\n1. \n"HYPERDIR E: HYPERgeometric functions DIfferential REduction MATHEMATICA \nbased pac kages for differential reduction of generalized hypergeometric \nfunctions : now with pFq\, F1\,F2\,F3\,F4"\nby V.V.Bytev\, M.Yu.Kalmykov\,B.A.Kniehl \, \n[arXiv:1105.3565]\n\n2. \n"Differential Reduction Techniques for the Evaluation of Feynman Diagrams"\nby S.A. Yost\, V.V. Bytev\, M.Yu. Kalmyko v\, B.A. Kniehl\, B.F.L. Ward\nPoS ICHEP2010:135\,2010\n[arXiv:1101.2348]\ n\n\n3.\n"Differential reduction of generalized hypergeometric functions f rom \nFeynman diagrams: One-variable case"\, \nby V.V.Bytev\, M.Yu.Kalmyko v\,B.A.Kniehl\, \nNucl.Phys.B836:129-170\, 2010 [arXiv:0904.0214]\n\n4.\n" Counting master integrals: integration by parts vs. differential reduction "\nby Mikhail Yu. Kalmykov\, Bernd A. Kniehl\n[arXiv:1105.5319]\n\n5.\n"Di fferential Reduction Algorithms for Hypergeometric Functions \nApplied to Feynman Diagram Calculation"\nby V.V.Bytev\, M.Kalmykov\, B.A.Kniehl\, B.F .L.Ward\, S.A.Yost\n[arXiv:0902.1352]\n\n[6]\n"Feynman Diagrams\, Differen tial Reduction\, and Hypergeometric Functions"\nby M. Yu. Kalmykov\, V. V. Bytev\, Bernd A. Kniehl\, B.F.L. Ward\, S.A.Yost \nPoS ACAT08:125\,2009 [ arXiv:0901.4716]\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId =7&sessionId=8&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=7&sessionId=8&c onfId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Progress in Automated Next-to-Leading Order calculations DTSTART;VALUE=DATE-TIME:20110907T084000Z DTEND;VALUE=DATE-TIME:20110907T092000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-108@cern.ch DESCRIPTION:Speakers: TRAMONTANO\, Francesco (CERN)\nWith the beginning of the experimental programs at the LHC\, the need of\ndescribing multi part icle scattering events with high accuracy becomes\nmore pressing. On the t heoretical side\, perturbative calculation within\nleading order precision cannot be sufficient\, therefore accounting for\neffects due to Next-to-L eading Order (NLO) corrections becomes mandatory.\nIn the last few years w e observed a tremendous progress in the computation\nof one-loop virtual c orrections for processes involving many particles.\nThe new ideas based on the universal four-dimensional decomposition for \nthe numerator of the i ntegrand for any one-loop scattering amplitudes\,\nthe four-dimensional un itarity-cuts\, and unitarity-cuts in $d$-dimension\, \nyielding the comple te determination of dimensionally regulated one-loop \namplitudes\, give t he possibility to develop automated multi-process \nevaluators for scatter ing amplitudes at NLO.\n\nhttp://indico.cern.ch/contributionDisplay.py?con tribId=108&sessionId=9&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=108&sessionId=9 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Modern actions\, algorithms\, and computers for Lattice QCD DTSTART;VALUE=DATE-TIME:20110906T095000Z DTEND;VALUE=DATE-TIME:20110906T103000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-109@cern.ch DESCRIPTION:Speakers: BOYLE\, Peter (University of Edinburgh)\nI discuss r ecently developed formulations of lattice Fermions possessing \nnear-exact chiral symmetry. These are particularly appropriate for the\nsimulation o f complex weak matrix elements. I also discuss the state\nof the art of su percomputing for Lattice simulation\n\nhttp://indico.cern.ch/contributionD isplay.py?contribId=109&sessionId=5&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=109&sessionId=5 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:One click dataset transfer: toward efficient coupling of distribut ed storage resources and CPUs. DTSTART;VALUE=DATE-TIME:20110905T141500Z DTEND;VALUE=DATE-TIME:20110905T144000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-102@cern.ch DESCRIPTION:Speakers: Mr. ZEROLA\, Michal (Academy of Sciences\, Czech Rep ublic)\nThe massive data processing in a multi-collaboration environment w ith geographically spread diverse facilities will be hardly "fair" to user s and hardly using network bandwidth efficiently unless we address and dea l with planning and reasoning related to data movement and placement. The needs for coordinated data resource sharing and efficient plans solving th e data transfer paradigm in a dynamic way are being more required. We will present the work which purpose is to design and develop an automated plan ning system acting as a centralized decision making component with emphasi s on optimization\, coordination and load-balancing.\n\nWe will describe t he most important optimization characteristic and\nmodeling approach based on "constraints". Constraint-based approach\nallows for a natural declara tive formulation of what must be satisfied\, without expressing how. The a rchitecture of the system\, communication between components and execution of the plan by underlying data transfer tools will be shown. We will emph asize the separation of the planner from the "executors" and explain how t o keep the proper balance between being deliberative and reactive. The ext ension of the model covering full coupling and reasoning about computing r esources will be shown.\n \nThe system has been deployed within STAR exper iment over several Tier sites and has been used for data movement in the f avour of user analyses or production processing. We will present several real use-case scenario and performance of the system with a comparison to the "traditional" - solved by hands methods. The benefits in terms of indi spensable shorter data delivery time due to leveraging available network p aths and intermediate caches will be revealed. \nFinally\, we will outline several possible enhancements and avenues for future work.\n\nhttp://indi co.cern.ch/contributionDisplay.py?contribId=102&sessionId=1&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=102&sessionId=1 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Modeling Fake Missing Transverse Energy with Bayesian Neural Netwo rks DTSTART;VALUE=DATE-TIME:20110906T160000Z DTEND;VALUE=DATE-TIME:20110906T162500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-101@cern.ch DESCRIPTION:Speakers: Dr. TENTINDO\, Silvia (Department of Physics-Florida State University)\nNeural networks (NN) are universal approximators. Ther efore\, in principle\, it should be possible to use them to model any reas onably smooth probability density such as the probability density of fake missing transverse energy (MET). The modeling of fake MET is an important experimental issue in events such as\n$Z \\rightarrow l^+ l^-$+jets\, whic h is an important background in high-mass Higgs searches at the Large Hadr on Collider. We describe how Bayesian neural networks (BNN) can be used to model the MET in $\\gamma$+jets events and how\, in turn\, the resulting BNN function can be used to model the missing transverse energy distribut ion in samples other than $\\gamma$+jets in which the MET is largely due t o instrumental effects.\n\nhttp://indico.cern.ch/contributionDisplay.py?co ntribId=101&sessionId=7&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=101&sessionId=7 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:The five dimensions of the genome DTSTART;VALUE=DATE-TIME:20110908T080000Z DTEND;VALUE=DATE-TIME:20110908T084000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-106@cern.ch DESCRIPTION:Speakers: ROUGEMONT\, Jacques (EPFL)\nThanks to large sequenci ng initiatives of the last 10 years we now have access to full genome sequ ences in digital form\, in particular\nfor laboratory species such as the mouse whose genome is about 3.5 billion letters in size.\nRecent high-thro ughput technologies allow to then probe the function of this genome in man y different experimental conditions by\nsampling the genome at the rate of 2-3 billion letters per experiment\, distributed with strong bias towards particular \nregions of the genome sharing a given biochemical property. \nThe analysis of these large datasets is a fascinating challenge. \nI wi ll illustrate this with two situations where time\, space and chemical sta te of the DNA are interrelated:\nI will first present data on the circadia n (24h) rhythms in the mouse liver: \nmany biological functions must be ac tivated synchronously at certain times of the day\nand are coupled to an i nternal (biochemical) clock within each cell. \nThe second example comes f rom embryonic development\, where the correct\nbody patterning relies on a complex network of interactions within the genome and in particular on a tight control of the 3D folding of the DNA molecule within the cell's nucl eus.\nI will show how we reconstruct such 5D configurations from the stati stical analysis of the genome samples relative to the known full genome se quence\, \nand how we can make inferences about cellular machineries from these data.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=106& sessionId=10&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=106&sessionId=1 0&confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Status of parallelization of FORM DTSTART;VALUE=DATE-TIME:20110906T151000Z DTEND;VALUE=DATE-TIME:20110906T153500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-107@cern.ch DESCRIPTION:Speakers: UEDA\, Takahiro (Karlsruhe Institute of Technology)\ nWe report on the current status of the development of parallel versions\n of the symbolic manipulation system FORM. Currently there are two\nparalle l versions of the FORM: one is TFORM which is based on the POSIX\nthreads and for running on multicore machines\, and the other is ParFORM\nwhich us es the MPI and can run on computer clusters. By using these\nversions\, mo st of existing FORM programs can benefit from the\nparallelization without any modifications.\n\nhttp://indico.cern.ch/contributionDisplay.py?contri bId=107&sessionId=8&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=107&sessionId=8 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:The alignment of the CMS Silicon Tracker DTSTART;VALUE=DATE-TIME:20110905T132500Z DTEND;VALUE=DATE-TIME:20110905T135000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-104@cern.ch DESCRIPTION:Speakers: FLUCKE\, Gero (DESY (Hamburg))\nThe CMS all-silicon tracker consists of 16588 modules. In 2010 it has been successfully aligne d using tracks from cosmic rays and pp-collisions\, following the time dep endent movements of its innermost pixel layers. Ultimate local precision i s now achieved by the determination of sensor curvatures\, challenging the algorithms to determine about 200000 parameters. Remaining alignment unce rtainties are dominated by systematic effects that can bias track paramete rs by an amount relevant for physics analyses. These effects are controlle d by adding further information\, e.g. the mass of decaying resonances. Th e orientation of the TK respect to the magnetic field of CMS is determined with a stand-alone chi^2 minimization procedure. The geometries are final ly validated with several tools\, the monitored quantities include the bas ic track quantities (for both tracks from colliisons and cosmics) and phys ics resonances.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId= 104&sessionId=2&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=104&sessionId=2 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Three-Loop Calculation of the Higgs Boson Mass in Supersymmetry DTSTART;VALUE=DATE-TIME:20110905T130000Z DTEND;VALUE=DATE-TIME:20110905T132500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-39@cern.ch DESCRIPTION:Speakers: Dr. KANT\, Philipp (Humboldt-Universität zu Berlin) \nA Key feature of the minimal supersymmetric extension of the Standard\nM odel (MSSM) is the existence of a light Higgs boson\, the mass of\nwhich i s not a free parameter but an observable that can be predicted\nfrom the t heory. Given that the LHC is able to measure the mass of a\nlight Higgs w ith very good accuracy\, a lot of effort has been put into\na precise theo retical prediction.\n\nWe present a calculation of the SUSY-QCD correction s to this\nobservable to three-loop order. We perform multiple asymptotic \nexpansions in order to deal with the multi-scale three-loop diagrams\,\n making heavy use of computer algebra and keeping a keen eye on the\nnumeri cal error introduced.\n\nWe provide a computer code in the form of a Mathe matica package that\ncombines our three-loop SUSY-QCD calculation with the literature of\none- and two-loop corrections to the Higgs mass\, providin g a\nstate-of-the-art prediction for this important observable.\n\nhttp:// indico.cern.ch/contributionDisplay.py?contribId=39&sessionId=3&confId=9387 7 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=39&sessionId=3& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:SecDec: a tool for numerical multi-loop/leg calculations DTSTART;VALUE=DATE-TIME:20110905T151000Z DTEND;VALUE=DATE-TIME:20110905T153500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-38@cern.ch DESCRIPTION:Speakers: CARTER\, Jonathon (University of Durham)\nSector dec omposition is a method to extract singularities from \nmulti-dimensional polynomial parameter integrals in a universal way . \nIntegrals of this type arise in perturbative higher order calculations \nin multi-loop integrals as well as \nin phase space integrals involv ing unresolved massless particles. \n \nThe program 'SecDec' will be pres ented\, \nwhich applies iterated sector decomposition in an automated way\ , \nto produce a Laurent series in the regularisation parameter. \nT he coefficients of this series are finite parameter integrals\nwhich are i ntegrated numerically by Monte Carlo techniques. \n\nThe power of the prog ram is illustrated by presenting results and timings \nfor a number of cut ting edge multi-loop integrals\, \ne.g. 2-loop box integrals entering top quark pair production at NNLO or \n4-loop propagators. Applications to integrals occurring \nin calculations of real radiation at higher pert urbative orders\nwill also be presented.\n\nhttp://indico.cern.ch/contribu tionDisplay.py?contribId=38&sessionId=3&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=38&sessionId=3& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Dynamic deployment of a PROOF-based analysis facility for the ALIC E experiment over virtual machines using PoD and OpenNebula DTSTART;VALUE=DATE-TIME:20110905T130000Z DTEND;VALUE=DATE-TIME:20110905T132500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-33@cern.ch DESCRIPTION:Speakers: Dr. DARIO\, Berzano (Sezione di Torino (INFN)-Univer sita e INFN)\nThe conversion of existing computing centres to cloud facili ties is becoming popular also because of a more optimal usage of existing resources. Inside a medium to large cloud facility\, many specific virtual computing facilities might concur for the same resources based on their u sage and destination elastically\, i.e. by expanding or reducing allocated resources for currently running VMs\, or by turning them on and off. In t he ALICE experiment PROOF\, a parallel processing infrastructure\, has bec ome very popular for interactive analysis. The locality of PROOF-based ana lysis facilities forces site admins to scavenge enough resources to dedica te\, yet the chaotic nature of user-written analysis tasks would deem thes e resources to be unstable and used intensively only in small bursts typic ally during working hours\, making PROOF a typical use-case for HPC cloud computing. Currently\, a solution named PROOF-on-Demand (PoD) does exist t o dynamically and quickly provide a PROOF-enabled cluster by enqueuing age nts to a job scheduler. In a medium-sized computing centre\, namely a Tier -2\, sharing a queue between PROOF and ordinary Grid jobs is not viable du e to the very large time to wait in order to get enough workers ready: how ever\, an elastic cloud approach will enable existing machines currently r unning Grid jobs to temporarily make room for many personal PoD-provided P ROOF clusters on the same hardware in near-real-time\, with no stability i ssues for long-running Grid jobs\, through the perfect sandboxing intrinsi cally offered by virtual machines. We will show a usable prototype of a dy namically-deployed PROOF-based analysis facility by using existing tools\, such as PoD and OpenNebula\, orchestrated by a simple and lightweight con trol daemon.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=33& sessionId=1&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=33&sessionId=1& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Monitoring the Grid at local\, national\, and global levels DTSTART;VALUE=DATE-TIME:20110906T130000Z DTEND;VALUE=DATE-TIME:20110906T132500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-32@cern.ch DESCRIPTION:Speakers: Mr. GRONBECH\, Peter (Particle Physics-University of Oxford)\nMonitoring the Grid at local\, national\, and global levels\nThe GridPP Collaboration\n\nThe World-wide LHC Computing Grid is the computin g infrastructure setup to process the experimental data coming from the ex periments at the Large Hadron Collider located at CERN. \nGridPP is the pr oject that provides the UK part of this infrastructure across 19 sites in the UK. To ensure that these large computational resources are available a nd reliable requires many different monitoring systems. These range from l ocal site monitoring of\, for example\, the hardware and of batch system u tilization\, to UK-wide monitoring of Grid functionality and ultimately th e worldwide monitoring of resource provision and usage. In this paper we d escribe the monitoring systems used for the many different aspects of the system\, and how some of them are being integrated together.\nLocal site m onitoring covers\, cluster load\, batch system utilization\, network bandw idth monitoring and fault condition monitoring. The most common software u sed to monitor a cluster is Ganglia \, this system can be easily installed on all clients allowing data to be collected on a master node and display ed via a web server. Monitoring specific to the batch system used at a si te is also typically used. Many GridPP sites use the torque batch system ( developed from PBS). This can be monitored with pbswebmon \, which provide s a graphical way to monitor the occupancy of the cluster\, and the differ ent user’s job shares and efficiencies. Another tool is Nagios\, which provides a very powerful frame work that can be used to monitor the status of systems. The Nagios system can be configured to run tests at intervals and carry out actions dependant on the results. This can be emailing a wa rning message or running an event handler that takes remedial action to so lve a problem. One of the advantages of Nagios is that if all is well it does not bother you and there is no need to actually look at a status Web page. It can let you know (via email\, web or SMS) when there is a problem . Network health\, usage and bandwidth is monitored at many sites with cac ti and/ or Network Weathermap. Available bandwidth between sites in the UK is monitored by each site having a dedicated ‘Gridmon’ test box that performs a matrix of iperf and other tests between the UK sites. The resu lts are stored on a central database with a web frontend. \nOther UK wide testing includes a GridPP developed summation of relevant WLCG tests coup led with dedicated UK tests developed by Prof. S. Lloyd at QMUL and the UK regional Nagios based Service Availability Monitoring (SAM). This service queries a central database (GOCDB) and Grid information services to crea te a list of sites and systems to be tested. The services offered are tes ted and the results of the tests are sent via an active MQ message bus to the EGI Central Operations Dashboard. Each region has an operator on duty that can raise alarm tickets against sites that have failed critical tests . \nSystems Administrators are often overwhelmed by the number of differen t web sites and monitoring systems they should track. Attempts to integrat e output from several systems into a site dashboard have been made at the Tier 1 and some of the larger sites. These systems will be described.\n\nh ttp://indico.cern.ch/contributionDisplay.py?contribId=32&sessionId=6&confI d=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=32&sessionId=6& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:A Linear Iterative Unfolding Method DTSTART;VALUE=DATE-TIME:20110908T160000Z DTEND;VALUE=DATE-TIME:20110908T162500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-37@cern.ch DESCRIPTION:Speakers: LASZLO\, Andras (CERN\, Geneva (on leave of absence from KFKI Research Institute for Particle and Nuclear Physics\, Budapest)) \nA freqently faced task in experimental physics is to measure the probabi lity distribution of some quantity. Often this quantity to be measured is smeared by a non-ideal detector response or by some physical process. The procedure of removing this smearing effect from the measured distribution is called unfolding\, and is a delicate problem in signal processing. Due to the numerical ill-posedness of this task\, various methods were invente d which\, given some assumptions on the initial probability distribution\, try to regularize the problem. Most of these methods definitely introduce bias on the estimate of the initial probability distribution. We propose a linear iterative method (motivated by the Neumann series known in functi onal analysis)\, which has the advantage that no assumptions on the initia l probability distribution is needed. Since it is a linear scheme\, statis tical error propagation can be performed in an exact manner. Convergence i s proved under certain quite general conditions\, and in that case the met hod can be seen to be asymptotically unbiased. On the other hand\, as a pr ice\, the approach is relatively statistics demanding. We provide a numeri cal C and C++ library for the implementation of the method.\n\nhttp://indi co.cern.ch/contributionDisplay.py?contribId=37&sessionId=12&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=37&sessionId=12 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Continuous simulation of Beyond-Standard-Model processes with mult iple parameters DTSTART;VALUE=DATE-TIME:20110908T132500Z DTEND;VALUE=DATE-TIME:20110908T135000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-36@cern.ch DESCRIPTION:Speakers: ZHONG\, Jiahang (Institute of Physics-Academia Sinic a)\nWe present a new approach to simulate Beyond-Standard-Model (BSM) proc esses which are defined by multiple parameters. In contrast to the traditi onal grid-scan method where a large number of events are simulated at each point of a sparse grid in the parameter space\, this new approach simulat es only a few events at each of a selected number of points distributed ra ndomly over the whole parameter space. In subsequent analysis\, we rely on the fitting by the Bayesian Neural Network (BNN) technique to obtain accu rate estimation of the acceptance distribution. With this new approach\, t he signal yield can be estimated continuously\, while the required number of simulation events is greatly reduced.\n\nhttp://indico.cern.ch/contribu tionDisplay.py?contribId=36&sessionId=12&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=36&sessionId=12 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Unparametrized multi-dimensional kernel density- and likelihood ra tio estimator DTSTART;VALUE=DATE-TIME:20110908T153500Z DTEND;VALUE=DATE-TIME:20110908T160000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-35@cern.ch DESCRIPTION:Speakers: Mr. KOEVESARKI\, Peter (Physikalisches Institut-Univ ersitaet Bonn)\nA novel method to estimate probability density functions\, suitable for multivariate analyses will be presented. The implemented alg orithm can work on relatively large samples\, iteratively finding a non-pa rametric density function with adaptive kernels. With increasing number of sample points the resulting function converges to the real probability de nsity. Specifically\, we discuss a classification example\, showing the op timal separation of signal and background events based on likelihood ratio s. Unlike traditional classification methods\, such as neural networks\, t his method is free from classical overtraining effects. Furthermore\, as i t is possible to calculate likelihood ratios depending on signal and backg round cross section\, the method is suitable for small signal searches at LHC.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=35&sessionI d=12&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=35&sessionId=12 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:10 Years of Object-Oriented Analysis on H1 DTSTART;VALUE=DATE-TIME:20110905T135000Z DTEND;VALUE=DATE-TIME:20110905T141500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-60@cern.ch DESCRIPTION:Speakers: Dr. LAYCOCK\, Paul (University of Liverpool)\nOver a decade ago\, the H1 Collaboration decided to embrace the\nobject-oriented paradigm and completely redesign its data analysis\nmodel and data storag e format. The event data model\, based on the\nRooT framework\, consists of three layers - tracks and calorimeter\nclusters\, identified particles and finally event summary data -\nwith a singleton class providing unified access. This original\nsolution was then augmented with a fourth layer c ontaining\nuser-defined objects.\nThis contribution will summarise the his tory of the solutions used\,\nfrom modifications to the original design\, to the evolution of\nthe high-level end-user analysis object framework whi ch is used\nby H1 today. Several important issues are addressed - the\npo rtability of expert knowledge to increase the efficiency of\ndata analysis \, the flexibility of the framework to incorporate\nnew analyses\, the per formance and ease of use\, and lessons\nlearned for future projects.\n\nht tp://indico.cern.ch/contributionDisplay.py?contribId=60&sessionId=2&confId =93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=60&sessionId=2& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Advances in Service and Operations for ATLAS Data Management DTSTART;VALUE=DATE-TIME:20110905T135000Z DTEND;VALUE=DATE-TIME:20110905T141500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-61@cern.ch DESCRIPTION:Speakers: Dr. STEWART\, Graeme Andrew (CERN)\nATLAS has record ed almost 5PB of RAW data since the LHC started\nrunning at the end of 200 9. Many more derived data products and\ncomplimentary simulation data have also been produced by the\ncollaboration and\, in total\, 55PB is current ly stored in the Worldwide\nLHC Computing Grid by ATLAS. All of this data is managed by the ATLAS\nDistributed Data Management system\, called Don Q uixote 2 (DQ2).\n\nDQ2 has evolved rapidly to help ATLAS Computing operati ons to manage\nthese large quantities of data across the many grid sites a t which\nATLAS runs and to help ATLAS physicists get access to this data. In\nthis paper we describe new and improved DQ2 services:\n\n- Popularity service\, which measures usage of data across ATLAS.\n- Space monitoring a nd accounting at sites.\n- Automated blacklisting service.\n- Cleaning age nts\, which trigger deletion of unused data at sites.\n- Deletion agents\, to reliably delete unwanted data from sites.\n\nWe describe the experienc e of data management operation in ATLAS\ncomputing\, showing how these ser vices enable management of petabyte\nscale computing operations.\n\nWe ill ustrate the coupling of data management services to other parts\nof the AT LAS computing infrastructure\, in particular showing how\nfeedback from th e distributed analysis system in ATLAS has enabled\ndynamic placement of t he most popular data\, helping users and groups\nto analyse the increasing data volumes on the grid.\n\nhttp://indico.cern.ch/contributionDisplay.py ?contribId=61&sessionId=1&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=61&sessionId=1& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:An Alternative Method for Tilecal Signal Detection and Amplitude E stimation DTSTART;VALUE=DATE-TIME:20110908T151000Z DTEND;VALUE=DATE-TIME:20110908T153500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-63@cern.ch DESCRIPTION:Speakers: Mr. SOTTO-MAIOR\, Peralva (Universidade Federal do R io de Janeiro (UFRJ))\nThe Barrel Hadronic calorimeter of ATLAS (Tilecal) is a detector used in the reconstruction of hadrons\, jets\, muons and mis sing transverse energy from the proton-proton collisions at the Large Hadr on Collider (LHC). It comprises 10\,000 channels in four readout partition s and each calorimeter cell is made of two readout channels for redundancy . The energy deposited by the particles produced in the collisions is read out by the several readout channels and its value is estimated by an opti mal filtering algorithm\, which reconstructs the amplitude and the time of the digitized signal pulse sampled every 25 ns.\nThis work deals with sig nal detection and amplitude estimation for the Tilecal under low signal-to -noise ratio (SNR) conditions. It explores the applicability (at the cell level) of a Matched Filter (MF)\, which is known to be the optimal signal detector in terms of the SNR. Moreover\, it investigates the impact of sig nal detection when summing both signals from the same cell before estimati ng the amplitude\, instead of performing it afterwards as it is currently done. The signal of interest is electronically conditioned to have a well- defined shape (the Tilecal reference pulse shape) and the electronic noise distribution is a Gaussian-like\, for which decorrelation can be handled by estimating the whitening transformation of the process. As a result\, t he MF method implements a finite impulse response (FIR) filter whose coeff icients are the Tilecal reference pulse shape.\nThe MF method is compared to the Optimal Filter (OF) algorithm currently implemented in the Tilecal DSP\, which performs the signal reconstruction online. To this end\, two c lasses of data have been used: the noise dataset\, which comprises noise s ignals taken from a pedestal run during nominal Tilecal operation\, and th e signal dataset\, which is constructed from Tilecal reference pulse shape in convolution with added noise. In order to simulate realistic condition s\, amplitude and time shifting distributions were taken into account to g enerate the signal dataset. The results showed that for conditions where t he signal pedestal could be considered stationary\, the MF filter techniqu e achieves a better SNR performance compared to the OF technique for the t ested simulated signals. Current studies include analyzing the behavior of the MF method in conditions where the signal pulse is distorted by the pi le-up from additional interactions to the primary collision.\n\nhttp://ind ico.cern.ch/contributionDisplay.py?contribId=63&sessionId=12&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=63&sessionId=12 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:An adaptive Monte-Carlo Markov chain algorithm for counting muons in Auger water Cherenkov detector signals DTSTART;VALUE=DATE-TIME:20110908T135000Z DTEND;VALUE=DATE-TIME:20110908T141500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-69@cern.ch DESCRIPTION:Speakers: Mr. KÉGL\, Balázs (Linear Accelerator Laboratory)\ nAdaptive Metropolis (AM) is a powerful recent algorithmic tool in numeric al Bayesian data analysis. AM builds on a well-known Markov Chain Monte Ca rlo (MCMC) algorithm but optimizes the rate of convergence to the target d istribution by automatically tuning the design parameters of the algorithm on the fly. In our data analysis problem of counting muons in the water C herenkov signal of the surface detectors in the Pierre Auger Experiment\, the signal is modeled by a mixture distribution. Label switching is a majo r problem in inference on such models because of the invariance to symmetr ies. The simplest (non-adaptive) solution is to modify the prior in order to make it select a single permutation of the variables\, introducing an i dentifiability constraint. This solution is known to cause artificial bias es by not respecting the topology of the posterior. In this paper we desci be a new online relabeling procedure which can be incorporated into the AM algorithm. We state the convergence of the algorithm and identify the lin k between its modified target measure and the original posterior distribut ion of interest. \n\nOur long-term goal in the Pierre Auger Experiment is to develop a comprehensive generative model for the surface detector sign al and use MCMC techniques to estimate the parameters. The first step of t his program is the development of a generative model of the response of an Auger water tank and an adaptive reversible jump MCMC algorithm that can deal with the unknown number of muonic components in the signal. In the se cond part of this paper we discuss the algorithmic and computational issue s of implementing MCMC techniques for large-scale data analysis.\n\nhttp:/ /indico.cern.ch/contributionDisplay.py?contribId=69&sessionId=12&confId=93 877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=69&sessionId=12 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:DRA method: Powerful tool for the calculation of the loop integral s. DTSTART;VALUE=DATE-TIME:20110906T132500Z DTEND;VALUE=DATE-TIME:20110906T135000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-2@cern.ch DESCRIPTION:Speakers: Dr. LEE\, Roman (Budker Institute of Nuclear Physics )\nThe method of calculation of the loop integrals based on the dimensiona l recurrence relation and analyticity of the integrals as functions of $d$ is reviewed. Special emphasis is made on the possibility to automatize ma ny steps of the method. New results obtained with this method are presente d.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=2&sessionId=8 &confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=2&sessionId=8&c onfId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Efficient Pseudo-Random Number Generation for Monte-Carlo Simulati ons Using Graphic Processors DTSTART;VALUE=DATE-TIME:20110908T160000Z DTEND;VALUE=DATE-TIME:20110908T162500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-98@cern.ch DESCRIPTION:Speakers: Dr. CARMINATI\, Federico (CERN)\nThe future of high power computing is evolving towards the efficient use of highly parallel computing environment. The class of devices that has been designed havin g parallelism features in mind is the Graphics Processing Units (GPU) whi ch are highly parallel\, multithreaded computing devices. One application where the use of massive parallelism comes instinctively is Monte-Carlo s imulations where a large number of independent events have to be simulated . At the core of the Monte-Carlo simulation lies the random number genera tors. For GPU programming\, the random number generator should have (a) good statistical properties (b) high computational speed (c) low memory u se\, and (d) a large period . The most commonly used Mersenne Twister gene rator has very good statistical properties with a long period of 2^(19937) -1 \, but not suitable for implementation in the GPU as it has a large st ate that must be updated serially. Each GPU thread must have an individual state in global RAM and requires multiple access per generator. The rela tively large number of computation per generated number makes the generato r too slow for GPU programming except in cases where the ultimate in quali ty is needed. In this paper\, we have used a hybrid approach as used in N VIDIA CUDA library. The suggestion is to use a combination of three Tauswo rthe generator with different parameters along with a simple Linear Congr uential Generator (LCG) where the mod operation is not performed explicitl y. The period of these combinations is quite high (2^121) and has good st atistical properties as the defects of one generator gets compensated by o ther. This hybrid generator requires four random seeds which can be suppli ed using a CPU-side random number generator. We have carried out alias Mo nte-Carlo sampling using this hybrid generator where each GPU thread is us ed to generate random variable in parallel. This would mean each thread ne eds to be provided a random seed independently. In the present work\, we have implemented alias sampling with NVIDIA GeForce GTX 480 GPU card usin g both CUDA and OpenCL kernels. It is noticed that the kernel execution in both cases is about 1000 times faster as compared to the CPU whereas th e total code execution is only 10 times faster. This is due to the fact th at memory copy from host to device or vice-versa is very slow. Therefore\, we try to minimise memory access time and implement a simple scheme to ge nerate random seed per thread on the fly from the formulae seed=1099087573 *id where id is the thread index. This is known as quick and dirty LCG whi ch has a period of 232 and mod operation is not explicitly needed due to o verflow of unsigned integer. It is shown that this hybrid generator which takes seed on the fly is quite fast\, reproduces the statistical properti es reasonably well and can easily be implemented on each thread of GPU as well as CPU in an efficient way.\n\nhttp://indico.cern.ch/contributionDisp lay.py?contribId=98&sessionId=11&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=98&sessionId=11 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Lessons from Static Analysis on HEP Software DTSTART;VALUE=DATE-TIME:20110908T135000Z DTEND;VALUE=DATE-TIME:20110908T141500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-91@cern.ch DESCRIPTION:Speakers: NAUMANN\, Axel (CERN)\nCoverity's static analysis to ol has been run on most of the LHC experiments' frameworks\, as well as se veral of the packages provided to them (e.g. ROOT\, Geant4). I will presen t how static analysis works and why it is complimentary to dynamic checker s like valgrind or test suites\; typical issues discovered by static analy sis\; and lessons learned.\n\nhttp://indico.cern.ch/contributionDisplay.py ?contribId=91&sessionId=11&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=91&sessionId=11 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:An Exploration of SciDB in the Context of Emerging Technologies fo r Data Stores in Particle Physics and Cosmology DTSTART;VALUE=DATE-TIME:20110908T132500Z DTEND;VALUE=DATE-TIME:20110908T135000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-90@cern.ch DESCRIPTION:Speakers: Dr. MALON\, David (High Energy Physics Division-Argo nne National Laboratory (ANL))\nTraditional relational databases have not always been well matched to the needs of data-intensive sciences\,\nbut ef forts are underway within the database community to attempt to address man y of the requirements of large-scale\nscientific data management. One suc h effort is the open-source project SciDB. Since its earliest incarnation s\,\nSciDB has been designed for scalability in parallel and distributed e nvironments\, with a particular emphasis\nupon native support for array co nstructs and operations. Such scalability is of course a requirement of a ny strategy\nfor large-scale scientific data handling\, and array construc ts are certainly useful in many contexts\, but these\nfeatures alone do no t suffice to qualify a database product as an appropriate technology for h osting particle physics\nor cosmology data. In what constitutes its 1.0 r elease in June 2011\, SciDB has extended its feature set\nto address addit ional requirements of scientific data\, with support for user-defined type s and functions\,\nfor data versioning\, and more.\n\nThis paper describes an evaluation of the capabilities of SciDB for two very different kinds o f physics data:\nevent-level metadata records from proton collisions at th e Large Hadron Collider\, and the output of cosmological\nsimulations run on very-large-scale supercomputers. This evaluation exercises the spectru m of SciDB capabilities\nin a suite of tests that aim to be representative and realistic\, including\, for example\, definition of four-vector\ndata types and natural operations thereon\, and computational queries that mat ch the natural use cases for\nthese data.\n\nhttp://indico.cern.ch/contrib utionDisplay.py?contribId=90&sessionId=11&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=90&sessionId=11 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Moving ROOT Forward. DTSTART;VALUE=DATE-TIME:20110908T141500Z DTEND;VALUE=DATE-TIME:20110908T144000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-92@cern.ch DESCRIPTION:Speakers: RADEMAKERS\, Fons (CERN)\nNow that the LHC has start ed the LHC experiments crave for stability in ROOT\, however progress in c omputing technology is not stopping and to keep ROOT up to date and compat ible with new technologies requires a lot of work. In this presentation we will show what we are currently working on and what new technologies we t ry to exploit.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=9 2&sessionId=11&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=92&sessionId=11 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Evaluation of likelihood functions on CPU and GPU devices DTSTART;VALUE=DATE-TIME:20110908T151000Z DTEND;VALUE=DATE-TIME:20110908T153500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-95@cern.ch DESCRIPTION:Speakers: Mr. SNEEN LINDAL\, Yngve (Norges Teknisk-Naturvitens . Univ. (NTNU) and CERN openlab)\nIn this work we present the parallel imp lementations of an algorithm used to evaluate the likelihood function of t he data analysis. The implementations run on CPU and GPU\, respectively\, and both devices cooperatively (hybrid). Therefore the execution of the al gorithm can take full advantage from users commodity systems\, like deskto ps and laptops\, using entirely the hardware at disposal. CPU and GPU impl ementations are based on OpenMP and OpenCL\, respectively. For the hybrid case\, we implemented a scheduler of the tasks so that the workload can be split and balanced in the two devices. Initially the scheduler determines the workloads for each device\, so that the corresponding execution time s are balanced. From this phase a ratio of the workloads is obtained. Then it starts the likelihood function evaluations\, keeping fixed the previou sly determined ratio of the workloads. We show the results of the scalabil ity when running on CPU. Then we show the comparison of the performance of the GPU implementation on different hardware systems from different vendo rs\, and the performance when running in the hybrid case. The tests are ba sed on likelihood functions from real data analysis carried out in the hig h energy physics community.\n\nhttp://indico.cern.ch/contributionDisplay.p y?contribId=95&sessionId=11&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=95&sessionId=11 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Multivariate analysis and data mining: statistics in the computer age DTSTART;VALUE=DATE-TIME:20110906T084000Z DTEND;VALUE=DATE-TIME:20110906T092000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-94@cern.ch DESCRIPTION:Speakers: HAND\, David (Imperial College London)\nFor very sou nd reasons\, including the central limit theorem and mathematical tractabi lity\, classical multivariate statistics was heavily based on the multivar iate normal distribution. However\, the development of powerful computers \, as well as increasing numbers of very large data sets\, has led to a dr amatic blossoming of research in this area\, and the development of entire ly new tools for multivariate analysis. The talk will present an overview of such developments\, illustrating with ideas\, tools\, and methods such as empirical Bayes\, false discovery rate\, bootstrap methods\, anomaly d etection methods\, and streaming data analysis.\n\nhttp://indico.cern.ch/c ontributionDisplay.py?contribId=94&sessionId=5&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=94&sessionId=5& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Multivariate Correlated Sampling Using Extended Alias Techniques DTSTART;VALUE=DATE-TIME:20110905T141500Z DTEND;VALUE=DATE-TIME:20110905T144000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-97@cern.ch DESCRIPTION:Speakers: Dr. CARMINATI\, Federico (CERN)\nMonte-Carlo techni que enables one to generate random samples from distributions with known characteristics and helps to make probability based inferences of the unde rlying physical processes. Fast and efficient Monte-Carlo particle trans port code particularly for high energy nuclear and particle physics exper iments has become an important tool starting from the design and fabrica tion of detectors to the modelling of the physics outcome as close as the reality. Quite often Monte-Carlo simulations require multivariate rando m numbers to be generated from correlated data both from normal and non-no rmal distributions. Although several techniques exist for multivariate co rrelated samplings of varying degrees of success\, the most elegant method is the technique that uses the principal component analysis of the given correlation matrix R for generating multivariate random numbers with spec ified inter-correlations. While the component analysis is suitable for mul tivariate normal distribution\, it may not work always particularly when the distribution is non Gaussian. In this work\, we propose an extended al ias sampling which was originally proposed by A. J. Walker in 1977 to samp le from an one dimensional distribution. This method is quite fast \, e fficient and reproduces the original distributions quite accurately (verif ied through chi-square as well co-variance test). It may be mentioned her e that this method is quite robust and is applicable to all type of multiv ariate distribution irrespective of whether the distribution is Gaussian o r Non-Gaussian.\nAlthough this method is quite general and can be applied to any dimensions\, in this work we have restricted sampling only from a two dimensional correlated distribution. The motivation behind this study has been to develop a ROOT based Monte-Carlo application package for l ow energy neutron transport down in energy to a few keV using the evaluat ed nuclear data file (ENDF) which is available in ROOT format. Work is in progress to apply this new method of alias technique to the ENDF data se t where the angle and energy distributions are strongly correlated.\n\nhtt p://indico.cern.ch/contributionDisplay.py?contribId=97&sessionId=2&confId= 93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=97&sessionId=2& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Do regions of ALICE matter? (Social relationships and data exchang es in the Grid) DTSTART;VALUE=DATE-TIME:20110908T153500Z DTEND;VALUE=DATE-TIME:20110908T160000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-96@cern.ch DESCRIPTION:Speakers: Mr. CARMINATI\, Federico (CERN\, Geneva\, Switzerlan d)\nFollowing a previous publication\, this study aims at investigating th e impact of regional affiliations of centres on the organisation of collab oration within the Distributed Computing ALICE infrastructure\, based on s ocial networks methods. A self-administered questionnaire was sent to all centre managers about support\, email interactions and wished collaboratio ns in the infrastructure. Several additional measures\, stemming from tech nical observations were produced\, such as bandwidth\, data transfers and Internet Round Trip Time (RTT) were also included. Information for 50 cent res were considered (60\\% response rate). Empirical analysis shows that d espite the centralisation on CERN\, the network is highly organised by reg ions. The results are discussed in the light of policy and efficiency issu es.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=96&sessionId =11&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=96&sessionId=11 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Regularization Schemes and Higher Order Corrections DTSTART;VALUE=DATE-TIME:20110905T135000Z DTEND;VALUE=DATE-TIME:20110905T141500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-11@cern.ch DESCRIPTION:Speakers: KILGORE\, William (Brookhaven National Lab)\nI apply commonly used regularization schemes to a multiloop\ncalculation to exami ne the properties of the schemes at higher orders.\nI find complete consis tency between the conventional dimensional\nregularization scheme and dime nsional reduction\, but I find that the\nfour-dimensional helicity scheme produces incorrect results at\nnext-to-next-to-leading order and singular results at\nnext-to-next-to-next-to-leading order. It is not\, therefore\ , a\nunitary regularization scheme.\n\nhttp://indico.cern.ch/contributionD isplay.py?contribId=11&sessionId=3&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=11&sessionId=3& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Application of Remote Debugging Techniques in User-Centric Job Mon itoring DTSTART;VALUE=DATE-TIME:20110905T155500Z DTEND;VALUE=DATE-TIME:20110905T162000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-10@cern.ch DESCRIPTION:Speakers: Dr. DOS SANTOS\, Tim (Bergische Universitaet Wuppert al)\nWith the Job Execution Monitor\, a user-centric job monitoring softwa re developed at the University of Wuppertal and integrated into the Pilot- based "PanDA" job brokerage system of the WLCG\, job progress and grid wor ker node health can be supervised in real time. Imminent error conditions can thusly be detected early by the submitter and countermeasures taken. G rid site admins can access aggregated data of all monitored jobs to infer the site status and to detect job misbehaviour. To remove the last "blind spot" from this monitoring\, a remote debugging technique based on the GNU C compiler suite was developed and integrated into the software\; its des ign concept and architecture will be described and its application discuss ed.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=10&sessionId =1&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=10&sessionId=1& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Integrating Amazon EC2 with the CMS Production Framework DTSTART;VALUE=DATE-TIME:20110905T132500Z DTEND;VALUE=DATE-TIME:20110905T135000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-15@cern.ch DESCRIPTION:Speakers: MELO\, Andrew Malone (Vanderbilt University)\nAs clo ud middleware (and cloud providers) have become more robust\, various expe riments with experience in Grid submission have begun to investigate the p ossibility of taking previously Grid-Enabled applications and making them compatible with Cloud Computing\, which will allow for dynamic scaling of the available hardware resources on a dynamic basis\, providing access to peak-load handling capabilities and possibly resulting in lower costs to t he experiment. Here we discuss current work within the CMS collaboration a t the LHC to both perform computation on EC2\, both for production and ana lysis use-cases.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId =15&sessionId=1&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=15&sessionId=1& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Advanced event reweighting for MVA training. DTSTART;VALUE=DATE-TIME:20110906T141500Z DTEND;VALUE=DATE-TIME:20110906T144000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-14@cern.ch DESCRIPTION:Speakers: MARTSCHEI\, Daniel (Inst. für Experimentelle Kernph ys.-Universitaet Karlsruhe-KIT)\nTitle: Advanced event reweighting for MVA training.\n\n\nMultivariate discrimination techniques\, such as Neural Ne tworks\, are key ingredients to modern data analysis and play an important role in high energy physics. They are usually trained on simulated Monte Carlo (MC) samples to discriminate signal from background and are then app lied to data. This has in general some side effects which we address in th is talk.\n\nOne is that the discriminator behaviour on real data depends o n the agreement between the MC training sample and data. We present ways o f re-weighting MC samples on a per event basis to make them more look like data. \nIn some cases it is even possible to become completely independen t from MC simulations by using the sPlot technique\, which also makes exte nsive use of weights during the training and is a sort of advanced backgro und subtraction procedure.\n\nAnother issue is that a cut on the discrimin ator can change the distribution of variables which discriminate signal fr om background themselves. This becomes an issue if one wants to see and fi t a clear signal peak in this distribution on data as a final result\, e.g . in the invariant mass of decay particles. Our approach uses a neural net work which is trained to discriminate between signal and background while explicitely disallowing any influence on the distribution the variable of interest to be used for template fits in the end.\n\nWe will give examples of the application of these three techniques performed with the NeuroBaye s package in different physics analysis.\n\nhttp://indico.cern.ch/contribu tionDisplay.py?contribId=14&sessionId=7&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=14&sessionId=7& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Tau identification using multivariate techniques in ATLAS DTSTART;VALUE=DATE-TIME:20110906T132500Z DTEND;VALUE=DATE-TIME:20110906T135000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-16@cern.ch DESCRIPTION:Speakers: Prof. O'NEIL\, Dugan (Simon Fraser University (SFU)) \nTau leptons will play an important role in the physics program at the \nLHC. They will be used in electroweak measurements and in detector \nrelated studies like the determination of the missing transverse \nenergy s cale\, but also in searches for new phenomena like the Higgs \nboson or Supersymmetry. \n \nDue to the huge b ackground from QCD processes\, efficient tau \nidentification techniques with large fake re jection are essential. Tau \nobject appear as collimated jets with low track multiplicity and \nsingle variable criteria a re not enough to efficiently separate them \nfrom jets and electrons. This can be achieved using mo dern \nmultivar iate techniques which make optimal use of all the information \navailable. They are particularly use ful when the discriminating \nvariables are not independent and no single variable provides go od \nsignal and backgro und separation. \n\nIn ATLAS several advanced algorithms are applied to identify taus\, in \np articular a projective likelihood estimator and boosted decision trees. \nAll multivariate methods appl ied to the ATLAS simulated data perform \nbetter than the baseline cut analysis. Their performance is shown \nusing high energy data collected at the ATLAS experiment. The \nstrengths and weaknesses of each techni que are also discussed.\n\nhttp://indico.cern.ch/contributionDisplay.py?co ntribId=16&sessionId=7&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=16&sessionId=7& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Computing On Demand: Analysis in the Cloud DTSTART;VALUE=DATE-TIME:20110908T084000Z DTEND;VALUE=DATE-TIME:20110908T092000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-117@cern.ch DESCRIPTION:Speakers: Dr. MANAFOV\, Anar (GSI - Helmholtzzentrum fur Schwe rionenforschung GmbH)\nConstant changes in computational infrastructure li ke the current interest in Clouds\, imply conditions on the design of appl ications. We must make sure that our analysis infrastructure\, including s ource code and supporting tools\, is ready for the on demand computing (OD C) era.\n\n\nThis presentation is about a new analysis concept\, which is driven by users needs\, completely disentangled from the computational res ources\, and scalable.\n\n\nWhat does it take for an analysis code to be p erformed on any resource management system?\nHow can one achieve goals of on demand analysis\, using PROOF on Demand (PoD)?\nThese questions and suc h topics as preferable location of data files as well as tools and softwar e development techniques for on demand data analysis are covered. Also ana lysis implementation requirements and comparisons of traditional and “on demand” facilities will be discussed during this talk.\n\nhttp://indico .cern.ch/contributionDisplay.py?contribId=117&sessionId=10&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=117&sessionId=1 0&confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Where do we go from here? - The next phase of computing in HEP DTSTART;VALUE=DATE-TIME:20110905T090000Z DTEND;VALUE=DATE-TIME:20110905T094000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-116@cern.ch DESCRIPTION:Speakers: JARP\, Sverre (CERN)\nThe speaker will start by revi ewing the dominant technologies chosen for \nthe LHC Computing Grid and br iefly discuss their suitability. He will \nthen go on to look at technolog ies that have emerged since\, but are not \nbeing seriously used. Some of these technologies are being or have been \nevaluated by the CERN openlab. In the last part of the talk the speaker \nwill argue for the adoption of certain of these technologies for the \ndirect benefit of the LCG/HEP com munity.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=116&sess ionId=0&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=116&sessionId=0 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Can 'Go' address the multicore issues of today and the manycore pr oblems of tomorrow ? DTSTART;VALUE=DATE-TIME:20110906T141500Z DTEND;VALUE=DATE-TIME:20110906T144000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-49@cern.ch DESCRIPTION:Speakers: Dr. BINET\, Sebastien (Laboratoire de l'Accelerateur Lineaire (LAL)-Universite de Pari)\nCurrent HENP libraries and frameworks were written before multicore\nsystems became widely deployed and used.\n From this environment\, a 'single-thread' processing model naturally\nemer ged but the implicit assumptions it encouraged are greatly\nimpairing our abilities to scale in a multicore/manycore world.\n\nWhile parallel progra mming - still in an intensive phase of R&D\ndespite the 30+ years of liter ature on the subject - is an obvious\ntopic to consider\, other issues (bu ild scalability\, code clarity\, code deployment and ease of coding) are w orth investigating when preparing for the manycore era.\nMoreover\, if one wants to use another language than C++\, a language\nbetter prepared and tailored for expressing concurrency\, one also\nneeds to ensure a good and easy reuse of already field-proven\nlibraries.\n\nWe present the work res ulting from such investigations applied to the 'Go' programming language.\ nWe first introduce the concurrent programming facilities 'Go' is\nprovidi ng and how its module system addresses the build scalability and dependenc y hell issues.\nWe then describe the process of leveraging the many (wo)ma n-years put into scientific Fortran/C/C++ libraries and making them availa ble to the Go ecosystem.\nThe ROOT data analysis framework\, the C-BLAS li brary and the Herwig-6 MonteCarlo generator will be taken as examples.\nFi nally\, performances of a small analysis written in Go and using\nFortran and C++ libraries will be discussed.\n\nreferences:\nGo: http://gola ng.org\nROOT: http://root.cern.ch\nC-BLAS: http://www.netlib.org/cla pack/cblas/\nHerwig-6: http://hepwww.rl.ac.uk/theory/seymour/herwig/\n\nht tp://indico.cern.ch/contributionDisplay.py?contribId=49&sessionId=6&confId =93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=49&sessionId=6& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Strange Bedfellows: Quantum Mechanics and Data Mining DTSTART;VALUE=DATE-TIME:20110907T080000Z DTEND;VALUE=DATE-TIME:20110907T084000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-46@cern.ch DESCRIPTION:Speakers: Dr. WEINSTEIN\, Marvin (SLAC National Accelerator La boratory)\nAll fields of scientific research have experienced an explosion of data. Analyzing this data to extract unexpected patterns presents a c omputational challenge that requires new\, advanced methods of analysis. DQC (Dynamic Quantum Clustering)\, invented by David Horn (Tel Aviv Unive rsity)\, is a novel\, interactive and highly visual approach to this prob lem. Studies are already underway at SLAC to apply this technology to\, a mong other things\, discovering hard-to-find events in particle physics da ta\, analyzing Fermi/Glast data and implementing large scale SSRL XAF stud ies of the in-situ chemistry of macroscopic heterogeneous samples. The m ethod has also been applied to problems in medicine\, bio-informatics and even the stock market. My talk will provide a brief introduction to the d istinction between supervised and unsupervised methods in data mining (clu stering in particular). Then\, I will\, very briefly\, discuss the theory of DQC and show a simple application. Finally I will review some of the problems that have been studied to date. This part of the discussion will \, as an aside\, present a very simple visualization technique that makes it possible to see very small features in two-dimensional data (think Dali tz plots).\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=46&se ssionId=9&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=46&sessionId=9& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Status of TMVA\, the toolkit for multivariate analysis DTSTART;VALUE=DATE-TIME:20110906T130000Z DTEND;VALUE=DATE-TIME:20110906T132500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-86@cern.ch DESCRIPTION:Speakers: VON TOERNE\, Eckhard (University of Bonn)\nThe toolk it for multivariate analysis\, TMVA\, provides a large set of advanced mul tivariate analysis techniques for signal/background classification and reg ression problems. These techniques are embedded in a framework capable of handling input data preprocessing and the evaluation of the results\, thus providing a simple and convenient tool for multivariate techniques. The a nalysis techniques implemented in TMVA can be easily invoked and the direc t comparison of their performance allows the user to choose the most appro priate for a particular data analysis. This talk presents recently develop ed features\, such as improved preprocessing\, option tuning and an exte nded unit test framework to ensure code stability. We also discuss the per formance of our most important multivariate techniques on example data and a comparison with theoretical performance limits.\n\nhttp://indico.cern.c h/contributionDisplay.py?contribId=86&sessionId=7&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=86&sessionId=7& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Numerical evaluation of one-loop QCD amplitudes DTSTART;VALUE=DATE-TIME:20110908T130000Z DTEND;VALUE=DATE-TIME:20110908T132500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-44@cern.ch DESCRIPTION:Speakers: Mr. BIEDERMANN\, Benedikt (Humboldt Universität zu Berlin)\nWe present the publicly available program NGLUON allowing the\nnu merical evaluation of colour-ordered amplitudes at one-loop \norder in mas sless QCD.\nThe program allows the evaluation of one-loop amplitudes\nfor an arbitrary number of gluons. We discuss in detail the speed as\nwell as the numerical stability. In addition the packages allows the\nevaluation o f one-loop scattering amplitudes using extended floating\npoint precision. \nFurthermore we discuss the extension to one-loop amplitudes including ma ssless\nquarks and show some phenomenological applications.\n\nhttp://indi co.cern.ch/contributionDisplay.py?contribId=44&sessionId=13&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=44&sessionId=13 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Feynman integrals\, polylogarithms and symbols DTSTART;VALUE=DATE-TIME:20110907T095000Z DTEND;VALUE=DATE-TIME:20110907T103000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-45@cern.ch DESCRIPTION:Speakers: Dr. DEL DUCA\, Vittorio (Laboratori Nazionali di Fra scati (INFN))\nWe suppose that a solution to a given Feynman integral is k nown in terms of multiple polylogarithms\, and address the question of how to find another solution which is equivalent to the former\, but with a s impler analytic structure.\n\nhttp://indico.cern.ch/contributionDisplay.py ?contribId=45&sessionId=9&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=45&sessionId=9& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:GELATIO - The GERDA framework for digital signal analysis DTSTART;VALUE=DATE-TIME:20110905T150500Z DTEND;VALUE=DATE-TIME:20110905T153000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-43@cern.ch DESCRIPTION:Speakers: Mr. AGOSTINI\, Matteo (Munich Technical University)\ nWe present the concept\, the implementation and the performance of a new software framework developed to provide a flexible and user-friendly envir onment for advanced analysis and processing of digital signals. The softwa re has been designed to handle the full data analysis flow of GERDA\, a lo w-background experiment which searches for the neutrinoless double beta de cay of Ge-76 by using high-purity germanium detectors at the INFN Gran Sas so underground Laboratory. The framework organizes the data into a multi-t ier structure\, from the raw traces of the Ge detectors up to the condense d analysis parameters\, and includes tools and utilities to handle the dat a stream between the different tiers. It supports a multi-channel modular and flexible analysis\, widely customizable by the user either via human-r eadable initialization files or via a graphical interface. The framework i s designed to be solid\, maintainable over a long lifetime and scalable to the future phases of the experiment. To ensure flexibility and good compu tational performances\, the framework includes both compiled and interpret ed code (C++\, Python and Bash). It relies upon ROOT and its extension TAM \, which provides compatibility with PROOF\, enabling the software to run in parallel on clusters of computers or multi-core machines. The software was tested on different platforms and benchmarked in several GERDA-related applications. A stable version is presently available for the collaborati on and it is used to provide the reference analysis of the GERDA data. A f ew applications of the framework to real GERDA data are presented and disc ussed.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=43&sessio nId=2&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=43&sessionId=2& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:FormCalc 7 DTSTART;VALUE=DATE-TIME:20110905T162500Z DTEND;VALUE=DATE-TIME:20110905T165000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-40@cern.ch DESCRIPTION:Speakers: HAHN\, Thomas (MPI f. Physik)\nThe talk presents the new features in FormCalc 7 (and some in LoopTools)\, such as analytic ten sor reduction\, inclusion of the OPP method\, and the interface to FeynHig gs.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=40&sessionId =3&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=40&sessionId=3& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:A Validation System for Data Preservation in HEP DTSTART;VALUE=DATE-TIME:20110906T132500Z DTEND;VALUE=DATE-TIME:20110906T135000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-41@cern.ch DESCRIPTION:Speakers: KEMP\, Yves (Deutsches Elektronen-Synchrotron (DESY) )\nPreserving data from past experiments and preserving the ability to\npe rform analysis with old data is of growing importance in many\ndomains of science\, including High Energy Physics (HEP). A study group on this issue \, DPHEP\, has been established in this field to provide guidelines and a structure\nfor international collaboration on data preservation projects i n HEP.\n\nThis contribution presents a framework that allows experimentali sts to\nvalidate their software against a previously defined set of tests in\nan automated way. The framework has been designed with a special\nfoc us for longevity\, as it makes use of open protocols\, has a modular\ndesi gn and is based on simple communication mechanisms. On the fabrics\nside\, tests are carried out in a virtual environment using a cloud\ninfrastruct ure. Within the framework\, it is easy to run validation\ntests on differe nt hardware plattforms\, or different major or minor\nversions of operatin g systems. Experts from IT or the experiments can\nautomatically detect f ailures in the test procedure by the help of\nreporting tools. Hence\, app ropriate actions can be taken in a timely\nmanner. The design and importa nt implementation aspects of the\nframework are shown and first experience s from early-bird-users will\nbe presented.\n\nhttp://indico.cern.ch/contr ibutionDisplay.py?contribId=41&sessionId=6&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=41&sessionId=6& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Full Reconstruction based on NeuroBayes at the Belle Experiment DTSTART;VALUE=DATE-TIME:20110906T135000Z DTEND;VALUE=DATE-TIME:20110906T141500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-9@cern.ch DESCRIPTION:Speakers: ZANDER\, Daniel (Karlsruhe Institute of Technology)\ nFull Reconstruction is an important analysis technique utilized at B fact ories where B mesons are produced in e+e- -> Y(4S) -> BBbar processes. By reconstructing one of the two B mesons in an event fully in a hadronic fin al state\, the properties of the other B meson are determined using moment um conservation. Therefore\, it allows to measure or perform searches for rare B meson decays involving one or more neutrinos in the final state. \n \nThis ansatz is complicated in practice by huge combinatorics and large a mounts of background. With over 1000 exclusively reconstructed B decay cha nnels the Full Reconstruction utilizes a hierachical reconstruction proced ure and probabilistic calculus instead of classical selection cuts. In thi s approach\, the decision to accept or reject a candidate is delayed to a later stage in order to make the most use of all available information. Th e multivariate analysis software package NeuroBayes was used extensively t o hold the balance between highest possible efficiency and acceptable cons umption of CPU time.\n\nAs a result of applying this ansatz\, the number o f fully reconstructed B mesons was increased by a factor of 2 after 10 yea rs of successful data taking. The new full reconstruction algorithm will t hus allow for more precise measurements of rare B meson decays.\n\nhttp:// indico.cern.ch/contributionDisplay.py?contribId=9&sessionId=7&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=9&sessionId=7&c onfId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Different forms of the generalized Crewther relation in QCD and QE D: concrete consequences of analytical multiloop calculations DTSTART;VALUE=DATE-TIME:20110905T160000Z DTEND;VALUE=DATE-TIME:20110905T162500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-77@cern.ch DESCRIPTION:Speakers: Dr. KATAEV\, Andrei (INR\, Moscow\, Russia)\nDiffere nt forms of the generalized Crewther relation in QED and QCD \nare discuss ed. They follow from applyication of the method of OPE to the AVV trian gle amplitude in the limit when conformal symmetry is valid and broken by the prosedure of renormalizations in the \nvarious variants of MS scheme\ , including 't Hooft prescription for \ndefining beta-function. Special fe atures of the conseuences of the \nadvanced alpha_s^4 -order analytical ca lculations of the Bjorken polarized sum rule and non-singlet contribution to the Adler D-function are discussed. The results of application of conf ormal symmetry and the original Crewther relation for getting QED-type a nalytical contributions to the Ellis-Jaffe sum rule in the 4-th order of P T is also demonstrated.\n\nhttp://indico.cern.ch/contributionDisplay.py?co ntribId=77&sessionId=3&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=77&sessionId=3& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Self-Organizing Maps Parametrization of Deep Inelastic Structure F unctions with Error Determination DTSTART;VALUE=DATE-TIME:20110906T160000Z DTEND;VALUE=DATE-TIME:20110906T162500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-76@cern.ch DESCRIPTION:Speakers: Prof. LIUTI\, simonetta (university of virginia)\nWe will present a method to extract parton distribution functions from hard scattering processes based on an alternative type of neural networks\, the Self-Organizing Maps (SOMs). Quantitative results including a detailed tr eatment of uncertainties will be presented within a Next to Leading Order analysis of both unpolarized and polarized inclusive deep inelastic scatte ring data. With a fully working procedure in hand\, we are capable to exte nd our analysis to the Generalized Parton Distribution (GPD) case\, thus e xploiting the “classification” and “visualization” properties of t he SOMs.\n\nWork supported by US D.O.E. grant DE-FG02-01ER41200. We thank for computer time the University of Virginia Alliance for Computational Sc ience and Engineering\, and the HPC group at Jefferson Lab.\n\nhttp://indi co.cern.ch/contributionDisplay.py?contribId=76&sessionId=8&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=76&sessionId=8& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:The EOS disk storage system at CERN DTSTART;VALUE=DATE-TIME:20110905T150500Z DTEND;VALUE=DATE-TIME:20110905T153000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-75@cern.ch DESCRIPTION:Speakers: Mr. PETERS\, Andreas Joachim (CERN)\nEOS was designe d to fulfill generic requirements on disk storage scalability and IO sched uling performance for LHC analysis use cases following the strategy to dec ouple disk and tape storage as individual storage systems.\n\nThe project was setup in April 2010. Since October 2010 EOS was evaluated by ATLAS as a disk only storage pool at CERN for analysis use cases in the context of various WLCG demonstrator projects.\n\nSince May 2011 analysis data has be en migrated to the EOSCMS and EOSATLAS production instances. Each instance contains several thousand disks and provides few petabytes of storage cap acity individually managed by EOS.\n\nIn this paper we summarize features available in the first release version of EOS and highlight some of the be nefits as a user analysis disk pool in comparison with other storage solut ions.\nIn the second part we will describe the current deployment and oper ation model of EOS in the CERN computer centre and it's usage by the CMS a nd ATLAS experiments. We will conclude with a roadmap and future direction s of EOS development and operations at CERN.\n\nhttp://indico.cern.ch/cont ributionDisplay.py?contribId=75&sessionId=1&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=75&sessionId=1& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Regularization of IR-divergent loop integrals DTSTART;VALUE=DATE-TIME:20110905T153500Z DTEND;VALUE=DATE-TIME:20110905T160000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-73@cern.ch DESCRIPTION:Speakers: Prof. DE DONCKER\, Elise (Western Michigan Universit y)\nWe report results of a new regularization technique for infrared (IR) divergent loop integrals using dimensional regularization\, where a positi ve regularization parameter (epsilon\, satisfying that the dimension d = 4 +2*epsilon) is introduced in the integrand to keep the integral from diver ging as long as epsilon > 0. \n Based on an asymptotic expansion of the integral we construct a linear system of equations\, which incorporate s values of the integral for varying epsilon in the right hand side of the system. The linear system is extended by one equation at a time for decre asing epsilon\, and solved for the leading coefficients of the Laurent exp ansion of the integral. This gives rise to an extrapolation as epsilon ten ds to zero. The solutions can be obtained by solving the systems directly or by a recursive method. \n We will outline the computations and t he evaluation of the integrals for various problems. An analysis involves the condition and truncation error of the method. All computations are kep t numerical and performed with automatic code\, including a possible reduc tion of the integral to a form without entangled singularities.\n The b asic technique can be applied to IR divergent integrals without (threshold ) singularities in the interior of the domain. For non-IR divergent integr als with threshold singularities\, the same method reduces to a linear ext rapolation for a calculation of the integral. We outline an extension of t he technique for integrals which have both types of singularities by resor ting to a double extrapolation or regularization.\n\nhttp://indico.cern.ch /contributionDisplay.py?contribId=73&sessionId=3&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=73&sessionId=3& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Progress on the Direct Computation Method DTSTART;VALUE=DATE-TIME:20110908T132500Z DTEND;VALUE=DATE-TIME:20110908T135000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-72@cern.ch DESCRIPTION:Speakers: Dr. YUASA\, Fukuko (KEK)\nWe report our progress on the development of the\nDirect Computation Method (DCM)\, which is a fully \nnumerical method for the computation of Feynman diagrams.\nBased on a co mbination of a numerical integration tool\nand a numerical extrapolation t echnique\, all steps in\nthe computation are carried out in a fully numeri cal\nway. The combined method is applicable to one-\, two-\nand multi-loop diagrams with arbitrary masses including\ncomplex masses.\n In this talk we show numerical results of a scalar\none-loop pentagon and hexagon witho ut any analytical\ntreatment\, neither reducing to a sum of box diagrams n or\nsector decomposition. Further we discuss the possibility\nof handling ultraviolet divergence using DCM.\n\nhttp://indico.cern.ch/contributionDis play.py?contribId=72&sessionId=13&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=72&sessionId=13 &confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:PROOF Perfomance Measurements Using PROOF Benchmark Suite DTSTART;VALUE=DATE-TIME:20110906T162500Z DTEND;VALUE=DATE-TIME:20110906T165000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-71@cern.ch DESCRIPTION:Speakers: Dr. RYU\, Sangsu (KiSTi Korea Institute of Science & Technology Information (KiS)\, Dr. GANIS\, Gerardo (CERN)\nPROOF (Paralle l ROOT Facility) is an extention of ROOT enabling interactive analysis in parallel on clusters of computers or a many-core machine. PROOF has been a dopted and successfully utilized as one of main analysis models by LHC exp eriments including ALICE and ATLAS. ALICE has seen growing number of PROOF clusters around the world\, CAF at CERN\, SKAF in Slovakia\, GSIAF at Dar mstadt being the main ALICE PROOF service farms. KIAF at KISTI is also pla nning on PROOF farm service in 2011.\n The PROOF benchmark suite is a ne w utility suite of PROOF to measure the performance and scalability of PRO OF. The primary goal of benchmark suite is to determine the optimal config uration parameters for a set of machines to be used as PROOF cluster.\n The suite measures the performance of the cluster for a set of standard ta sks\, CPU-intensive task and IO-intensive task which are 2 distintive styl es of analysis in typical HEP application\, as a function of the number of effective processes. From these results\, indications about the optimal n umber of concurrent processes can be derived. For large facilities\, the s uite should also give indications about the optimal number of sub-masters into which the cluster should be partitioned.\n Site administrators of P ROOF cluster can use the suite to measure the performance of the cluster a nd optimize the configuration of their cluster. PROOF developers can also utilize the suite to help them measure\, identify problems and improve the ir software.\n Performance of PROOF cluster measured with the benchmark suite will be presented including real use cases at ALICE experiment.\n\nh ttp://indico.cern.ch/contributionDisplay.py?contribId=71&sessionId=6&confI d=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=71&sessionId=6& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Reweighting NNPDFs. DTSTART;VALUE=DATE-TIME:20110906T153500Z DTEND;VALUE=DATE-TIME:20110906T160000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-70@cern.ch DESCRIPTION:Speakers: Mr. CERUTTI\, Francesco (Universitat de Barcelona)\n I present a method\, elaborated within the NNPDF Collaboration\, that allo ws the inclusion of the information contained in new datasets into an exis ting set of parton distribution functions without the need for refitting.\ nThe method exploits bayesian inference in the space of PDF replicas\, com puting for each replica a chisquare with respect to the new dataset and a weight associated to this. These weights are then applied to the ensemble of parton densities\, producing a reweighted set of replicas.\nThe reweig hting method may be used to assess the impact of any new data or pseudodat a on parton densities and thus on their predictions.\n\nhttp://indico.cern .ch/contributionDisplay.py?contribId=70&sessionId=8&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=70&sessionId=8& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:Semi-Supervised Anomaly Detection - Towards Model-Independent Sear ches of New Physics DTSTART;VALUE=DATE-TIME:20110906T153500Z DTEND;VALUE=DATE-TIME:20110906T160000Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-78@cern.ch DESCRIPTION:Speakers: Mr. KUUSELA\, Mikael (Helsinki Institute of Physics (HIP))\nMost classification algorithms used in high energy physics fall un der the category of supervised machine learning. Such methods require a tr aining set containing both signal and background events and are prone to c lassification errors should this training data be systematically inaccurat e for example due to the assumed MC model. To complement such model-depend ent searches\, we propose an algorithm based on anomaly detection techniqu es\, which does not require a MC training sample for the signal data. We f irst model the MC background using multivariate mixtures of Gaussians. We then search for deviations from the background model by fitting to the obs ervations a mixture of the background model and a number of additional Gau ssians using a variant of the EM algorithm. This allows us to perform patt ern recognition of any excess over the background. We show by comparison t o neural networks that such a semi-supervised approach is a lot more robus t against misspecification of the signal MC than supervised classification . In cases where there is an unexpected signal\, a neural network fails to correctly identify it while anomaly detection does not suffer from such a limitation. On the other hand\, when there are no systematic errors in th e signal MC\, both methods perform comparably. Due to its fully probabilis tic nature\, the anomaly detection model has a number of additional advant ages as well. Firstly\, the mixing proportion of the anomalous excess imme diately gives an estimate for its cross section and secondly\, the statist ical significance of the excess can easily be estimated using a bootstrapp ing-based likelihood-ratio test.\n\nhttp://indico.cern.ch/contributionDisp lay.py?contribId=78&sessionId=7&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=78&sessionId=7& confId=93877 END:VEVENT BEGIN:VEVENT SUMMARY:The LHCb DIRAC-based production and data management operations sys tems DTSTART;VALUE=DATE-TIME:20110906T135000Z DTEND;VALUE=DATE-TIME:20110906T141500Z DTSTAMP;VALUE=DATE-TIME:20130518T214731Z UID:indico-contribution-93877-47@cern.ch DESCRIPTION:Speakers: Dr. STAGNI\, Federico (Conseil Europeen Recherche Nu cl. (CERN))\, Dr. CHARPENTIER\, Philippe (Conseil Europeen Recherche Nucl. (CERN))\nThe LHCb computing model was designed in order to support the LH Cb physics program\, taking into account LHCb specificities (event sizes\, processing times etc...). Within this model several key activities are de fined\, the most important of which are real data processing (reconstructi on\, stripping and streaming\, group and user analysis)\, Monte-Carlo simu lation and data replication. In this contribution we detail how these acti vities are managed by the LHCbDIRAC Data Transformation System. The LHCbDI RAC Data Transformation System leverages the workload and data management capabilities provided by DIRAC\, a generic community grid solution\, to su pport data-driven workflows (or DAGs). The ability to combine workload and data tasks within a single DAG allows to create highly sophisticated work flows with the individual steps linked by the availability of data. This a pproach also provides the advantage of a single point at which all activit ies can be monitored and controlled. To highlight the versatility of the s ystem we present in more detail experience with real data of the 2010 and 2011 LHC run.\n\nWhile several interfaces are currently supported (includi ng python API and CLI)\, we will present the ability to create LHCb workfl ows through a secure web interface\, control their state in addition to cr eating and submitting jobs. To highlight the versatility of the system we present in more detail experience with real data of the 2010 and 2011 LHC run.\n\nhttp://indico.cern.ch/contributionDisplay.py?contribId=47&sessionI d=6&confId=93877 LOCATION: URL:http://indico.cern.ch/contributionDisplay.py?contribId=47&sessionId=6& confId=93877 END:VEVENT END:VCALENDAR