Recolección de Estudios Relevantes

memories Derandomizers Readout buffers (ROBs)

EVENT FILTER

Bunch crossing

rate 40 MHz

< 75 (100) kHz

~ 1 kHz

~ 100 Hz

Interaction rate

~1 GHz

Regions of Interest

(RODs) Full-event buffers and processor sub-farms Data recording 25 ns 1 kHz 4.5 kHz 30 ms 50 ms 2 s 10 s ⇠1 kHz

170 36 200 000 150 PB 0 0 1 1 0 1 0 2 2 2 0 1 1 2 12 h 48 h

2000

⇡1.6 MB <2 GB

0 1

0

Figure 2:Object Diagram of the GAUDI Architecture

4.2 Transient data stores

The data objects needed by the algorithms are organized in several transient data stores, depending on the nature of the data itself and its lifetime. The Transient Event Store contains event data that are valid only for the time it takes to process one event. The Transient Detector Store contains data that describe various aspects of the behavior of the detector (e.g. alignment) and generally have a lifetime that corresponds to the processing of many events. The Transient Histogram Store contains statistical data, which typically have a lifetime corresponding to the data processed in a complete job. Although the stores behave slightly differently, particularly with respect to the data lifetime (e.g. the event data store is cleared for each event), their implementations have many things in common and are based on a common component.

A transient store helps to minimize coupling between algorithm objects and data objects. This approach was inspired by the work done in the BaBar experiment [3]. An algorithm can deposit some piece of data into the transient store, and these data can be picked up later by other algorithms for further processing without knowing how they were created. This conforms to the ”blackboard” architectural style, in which the transient store fulfils the role of the blackboard.

The transient data store also serves as an intermediate buffer for any type of data conver- sion to another representation of the data, in particular the conversion into persistent objects or graphical objects. Thus data can have one transient representation and zero or more persistent or graphical representations.

The organisation of the data within the transient data stores is ”tree-like”, similar to a Unix file system. This allows data items that are logically related, such as Monte Carlo ”truth” infor- mation, to be structured and grouped at run-time. Each node in the tree may either contain data members, or other nodes containing further groups of data members (Figure 4). As in a directory structure, each node is theownerof everything below it and will delete all these items when it gets deleted. In general, object-oriented data models do not map onto a tree structure. Thus, mesh-like object associations have been implemented using symbolic links (again inspired from the Unix file system) in which the node does not acquire ownership of the referenced item.

836 Eur. Phys. J. C (2010) 70: 823–874

Fig. 1 The flow of the ATLAS simulation software, from event gen- erators (top left) through reconstruction (top right). Algorithms are placed insquare-cornered boxesand persistent data objects are placed inrounded boxes. The optional pile-up portion of the chain, used only when events are overlaid, isdashed. Generators are used to produce

data in HepMC format. Monte Carlo truth is saved in addition to energy depositions in the detector (hits). This truth is merged into Simulated Data Objects (SDOs) during the digitization. Also, during the digitiza- tion stage, Read Out Driver (ROD) electronics are simulated

are included in the output files for all the stages of the event simulation. The metadata include all configuration informa- tion for the job. Athena has also adopted the POOL (Pool Of persistent Objects for LHC) file handling and persistency

framework [11–13].

2.1 ATLAS simulation overview

An overview of the ATLAS simulation data flow can be

seen in Fig.1. Algorithms and applications to be run are

placed in square-cornered boxes, and persistent data objects are placed in round-cornered boxes. The optional steps re-

quired for pile-up or event overlay (see Sect.6.2) are shown

with a dashed outline.

A generator produces events in standard HepMC for-

mat [14]. These events can be filtered at generation time so

that only events with a certain property (e.g. leptonic decay or missing energy above a certain value) are kept. The gen-

erator is responsible for any prompt decays (e.g.ZorW

bosons) but stores any “stable” particle expected to prop-

agate through a part of the detector (see Sect.3). Because

it only considers immediate decays, there is no need to con- sider detector geometry during the generation step, except in controlling what particles are considered stable. During this step, the run number for the simulated data set and event numbers for each event are established. Event numbers are generally ordered in a single job, though events may be omitted because of filtering at each step. Run numbers for

simulated data sets derive from the job options used to gen- erate the sample and mimic real run numbers used during data taking.

These generated events are then read into the simulation. A record of all particles produced by the generator is re-

tained in the simulation output file (see Sect.3.6), but cuts

can be applied to select only certain particles to process in the simulation. Each particle is propagated through the full

ATLAS detector by GEANT4. The configuration of the de-

tector, including misalignments and distortions, can be set at run time by the user. The energies deposited in the sensitive portions of the detector are recorded as “hits,” containing the total energy deposition, position, and time, and are written to a simulation output file, called a hit file.

In both event generation and detector simulation, infor- mation called “truth” is recorded for each event. In the gen- eration jobs, the truth is a history of the interactions from the generator, including incoming and outgoing particles. A record is kept for every particle, whether the particle is to be passed through the detector simulation or not. In the simulation jobs, truth tracks and decays for certain particles are stored. This truth contains, for example, the locations of the conversions of photons within the inner detector and the subsequent electron and positron tracks. In the digitization jobs, Simulated Data Objects (SDOs) are created from the truth. These SDOs are maps from the hits in the sensitive re- gions of the detector to the particles in the simulation truth record that deposited the hits’ energy. The truth information

! ! !

! ! !

⌧± b b b b gq _!gq gq _!gq gg!q q q q !q0q0 g !q q 1% b b Bs !J/ b b B B = _N N ·N

N N N b 1 _! N N > 2 B Bs !J/ 1 ! N Bs ! J/ (µ+µ ) (K+K ) ! ! Bs !J/ (µ+µ ) (K+K ) Bs !J/ 10 eV

Chapter 5. The ATLAS experiment: computing and software 84 Pythia Generate partonic event Partonic event PythiaB Event contains a beauty quark? No PythiaB Throw event away Yes PythiaB Clone event Cloned events Pythia

Hadronize cloned events Events containing

B-hadrons

Events containing final states

StoreGate

Event(s) written in HepMC format

Pythia

Decay all unstable particles: B-mesons according to user input, B-

bar mesons according to tables PythiaB Event(s) contains signal process? Yes No Persistency

100

! !

3

>82 % p >1 GeV

z z

332 Page 16 of 35 Eur. Phys. J. C (2017) 77 :332 m] µ Transverse resolution [ 10 100 ATLAS = 8 TeV s Data (SVM) MC (SVM)

MC (truth without B constraint) MC (truth with B constraint)

Data / MC0.70.8 0.91 1.1 1.2 1.3 Number of tracks 0 10 20 30 40 50 60 70 (a) m] µ Longitudinal resolution [ 10 100 ATLAS = 8 TeV s Data (SVM) MC (SVM)

MC (truth without B constraint) MC (truth with B constraint)

Data / MC0.70.8 0.91 1.1 1.2 1.3 Number of tracks 0 10 20 30 40 50 60 70 (b) Fig. 13 Resolution of the primary vertex position inaxandbzas

function of the number of fitted tracks, estimated using the split-vertex method (SVM) for minimum-bias data (black circles) and MC simu- lation (blue squares). Also shown is the resolution obtained from the difference between the generator-level information and reconstructed

primary vertex position in MC simulation (labeled “truth”), with and without the beam-spot constraint (pinkandred trianglesrespectively). Thebottom panelin each plot shows the ratio of the resolution found using the split-vertex method in data to that obtained using the MC generator-level information without the beam-spot constraint

method (SVM). In this method thentracks associated to a primary vertex are ordered in descending order of their trans- verse momenta. The tracks are then split into two groups, one with even-ranking tracks and one with odd-ranking tracks, such that both groups have, on average, the same number of tracks,n/2. The vertex fit is applied independently to each group. The spatial separation between two resulting vertices gives a measurement of the intrinsic resolution for a vertex withn/2 tracks. The two split vertices must be reconstructed independently and therefore no beam-spot constraint is used during the fit.

Figure13shows the resolution in data calculated with the split-vertex method as a function of the number of tracks per vertex.

The split-vertex method is also used to calculate the res- olution for the minimum-bias simulation sample. There is good agreement between the data and simulation distribu- tions, showing that the reconstructed track parameters used in the vertex reconstruction are well modelled in the simula- tion. Figure13also shows the primary vertex resolution cal- culated as the difference between the true and reconstructed vertex position in the MC simulation. The good agreement between the split-vertex method and the resolution calculated with the MC generator-level information gives confidence that the split-vertex method provides a reliable measurement of the primary vertex resolution. At very low track multi- plicity the result of the split-vertex method deviates slightly from the resolution obtained using the generator-level infor- mation. Here the resolution obtained from the generator-level information benefits from the perfect knowledge of vertex

position decreasing the resolution spread, compared to the resolution obtained from the two reconstructed vertices in the split-vertex method. When the beam-spot constraint is included the resolution improves considerably in the trans- verse direction, staying below 20µm for the full range ofµ

studied. The longitudinal resolution reaches 30µm at high track multiplicity. Figure13also shows the resolution calcu- lated using MC generator-level information with and without beam-spot constraint.

7 Performance in the high pile-up regime

In this section, the study of the primary vertex reconstruction performance at lowµis extended to the high pile-up regime.

A dedicated data sample of minimum-bias events collected with values ofµbetween 55 and 72 was used to study the

performance of the primary vertex reconstruction in the pres- ence of multiple vertices. The simulation samples spanned values ofµfrom 0 to 22, typical of the standard 2012 data-

taking conditions, and from 38 to 72 to emulate the highµ

data sample.

The efficiency of primary vertex reconstruction decreases with increasing pile-up. In addition to the inefficiencies affecting single vertex reconstruction described in Sect.6, effects related to the merging of adjacent primary vertices start to play a significant role as pile-up increases. Figure14a shows the average number of vertices lost due to merging and to other effects, such as track reconstruction and detector acceptance.

123

332 Page 16 of 35 Eur. Phys. J. C (2017) 77 :332

m] µ Transverse resolution [ 10 100 ATLAS = 8 TeV s Data (SVM) MC (SVM)

MC (truth without B constraint) MC (truth with B constraint)

Data / MC0.70.8 0.91 1.1 1.2 1.3 Number of tracks 0 10 20 30 40 50 60 70 (a) m] µ Longitudinal resolution [ 10 100 ATLAS = 8 TeV s Data (SVM) MC (SVM)

MC (truth without B constraint) MC (truth with B constraint)

Data / MC 0.70.8 0.91 1.1 1.2 1.3 Number of tracks 0 10 20 30 40 50 60 70 (b)

Fig. 13 Resolution of the primary vertex position inaxandbzas

function of the number of fitted tracks, estimated using the split-vertex

method (SVM) for minimum-bias data (black circles) and MC simu-

lation (blue squares). Also shown is the resolution obtained from the

difference between the generator-level information and reconstructed

primary vertex position in MC simulation (labeled “truth”), with and

without the beam-spot constraint (pinkandred trianglesrespectively).

Thebottom panelin each plot shows the ratio of the resolution found using the split-vertex method in data to that obtained using the MC generator-level information without the beam-spot constraint

method (SVM). In this method thentracks associated to a primary vertex are ordered in descending order of their trans- verse momenta. The tracks are then split into two groups, one with even-ranking tracks and one with odd-ranking tracks, such that both groups have, on average, the same number of tracks,n/2. The vertex fit is applied independently to each group. The spatial separation between two resulting vertices gives a measurement of the intrinsic resolution for a vertex withn/2 tracks. The two split vertices must be reconstructed independently and therefore no beam-spot constraint is used during the fit.

Figure13shows the resolution in data calculated with the split-vertex method as a function of the number of tracks per vertex.

The split-vertex method is also used to calculate the res- olution for the minimum-bias simulation sample. There is good agreement between the data and simulation distribu- tions, showing that the reconstructed track parameters used in the vertex reconstruction are well modelled in the simula- tion. Figure13also shows the primary vertex resolution cal- culated as the difference between the true and reconstructed vertex position in the MC simulation. The good agreement between the split-vertex method and the resolution calculated with the MC generator-level information gives confidence that the split-vertex method provides a reliable measurement of the primary vertex resolution. At very low track multi- plicity the result of the split-vertex method deviates slightly from the resolution obtained using the generator-level infor- mation. Here the resolution obtained from the generator-level information benefits from the perfect knowledge of vertex

position decreasing the resolution spread, compared to the resolution obtained from the two reconstructed vertices in the split-vertex method. When the beam-spot constraint is included the resolution improves considerably in the trans- verse direction, staying below 20µm for the full range ofµ studied. The longitudinal resolution reaches 30µm at high track multiplicity. Figure13also shows the resolution calcu- lated using MC generator-level information with and without beam-spot constraint.

7 Performance in the high pile-up regime

In this section, the study of the primary vertex reconstruction performance at lowµis extended to the high pile-up regime. A dedicated data sample of minimum-bias events collected with values ofµbetween 55 and 72 was used to study the performance of the primary vertex reconstruction in the pres- ence of multiple vertices. The simulation samples spanned values ofµfrom 0 to 22, typical of the standard 2012 data- taking conditions, and from 38 to 72 to emulate the highµ data sample.

The efficiency of primary vertex reconstruction decreases with increasing pile-up. In addition to the inefficiencies affecting single vertex reconstruction described in Sect.6, effects related to the merging of adjacent primary vertices start to play a significant role as pile-up increases. Figure14a shows the average number of vertices lost due to merging and to other effects, such as track reconstruction and detector acceptance.

123

CP