AtlasPublicTopicHeader.png

HLT Tracking Public Results

Introduction

Approved plots that can be shown by ATLAS speakers at conferences and similar events. Please do not add figures on your own. Contact the responsible project leader in case of questions and/or suggestions. Follow the guidelines on the trigger public results page.

2018 Trigger Performance Plots

ATL-COM-DAQ-2018-061 The ATLAS Inner Detector Trigger performance in early 2018 data

The track finding efficiency of the Inner Detector (ID) trigger for tracks from offline medium quality muon candidates as a function of the offline muon pseudorapidity. The efficiency is evaluated with a 24 GeV muon trigger configured to select only on the track in the Muon Spectrometer. Offline muon candidates are required to have at least one pixel cluster, at least 4 SCT clusters, and no more than two missing hits from the silicon detectors where such hits would be expected. Also if expected, they should have at least one hit in the innermost pixel layer. Offline muons with transverse momentum (pT) greater than 4 GeV are used and as such the selection for muon candidates below the trigger threshold is biased towards candidates that appear to be higher in pT in the Muon Spectrometer. For the muon triggers the ID reconstruction first runs a Fast Track Finder algorithm followed by a Precision Tracking step. Statistical, Bayesian uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for tracks from offline medium quality muon candidates as a function of the offline muon transverse momentum (pT). The efficiency is evaluated with a 24 GeV muon trigger configured to select only on the track in the Muon Spectrometer. Offline muon candidates are required to have at least one pixel cluster, at least 4 SCT clusters, and no more than two missing hits from the silicon detectors where such hits would be expected. Also if expected, they should have at least one hit in the innermost pixel layer. The selection for offline muon candidates from below the trigger threshold is biased in favour of candidates that appear to be higher in pT in the Muon Spectrometer. For the muon triggers the ID reconstruction first runs a Fast Track Finder algorithm followed by a Precision Tracking step. Statistical, Bayesian uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for tracks from offline medium quality muon candidates as a function of the offline muon z position at the beamline. The efficiency is evaluated with a 24 GeV muon trigger configured to select only on the track in the Muon Spectrometer. Offline muon candidates are required to have at least one pixel cluster, at least 4 SCT clusters, and no more than two missing hits from the silicon detectors where such hits would be expected. Also if expected, they should have at least one hit in the innermost pixel layer. Offline muons with transverse momentum (pT) greater than 4 GeV are used and as such the selection for muon candidates below the trigger threshold is biased towards candidates that appear to be higher in pT in the Muon Spectrometer. For the muon triggers the ID reconstruction first runs a Fast Track Finder algorithm followed by a Precision Tracking step. Statistical, Bayesian uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for tracks from offline medium quality muon candidates as a function of the mean interaction multiplicity per bunch crossing. The efficiency is evaluated with a 24 GeV muon trigger configured to select only on the track in the Muon Spectrometer. Offline muon candidates are required to have at least one pixel cluster, at least 4 SCT clusters, and no more than two missing hits from the silicon detectors where such hits would be expected. Also if expected, they should have at least one hit in the innermost pixel layer. Offline muons with transverse momentum (pT) greater than 4 GeV are used and as such the selection for muon candidates below the trigger threshold is biased towards candidates that appear to be higher in pT in the Muon Spectrometer. For the muon triggers the ID reconstruction first runs a Fast Track Finder algorithm followed by a Precision Tracking step. Statistical, Bayesian uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for tracks from offline medium quality 1-prong tau candidates with offline transverse momentum (pT) greater than 20 GeV as a function of the offline track transverse momentum (pT). The efficiency is evaluated with a 25 GeV tau trigger configured to select only on the calorimeter cluster. Offline tracks are required to have no missing hits from those expected in the pixel detector, and a tight overall selection on the total number of silicon hits. For the tau triggers the ID reconstruction runs in two stages - the first stage running a Fast Track Finder algorithm in a narrow region around the calorimeter cluster direction, but extended along the entire luminous region at the beamline to identify the leading tracks and the z position of the interaction, and the second stage running in a wider region about this leading track, but very narrow in z, running a Fast Track Finder algorithm followed by the Precision Tracking. Statistical, Bayesian uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for tracks from offline medium quality 1-prong tau candidates with offline transverse momentum (pT) greater than 20 GeV as a function of the mean offline track psuedorapidity. The efficiency is evaluated with a 25 GeV tau trigger configured to select only on the calorimeter cluster. Offline tracks are required to have no missing hits from those expected in the pixel detector, and a tight overall selection on the total number of silicon hits. For the tau triggers the ID reconstruction runs in two stages - the first stage running a Fast Track Finder algorithm in a narrow region around the calorimeter cluster direction, but extended along the entire luminous region at the beamline to identify the leading tracks and the z position of the interaction, and the second stage running in a wider region about this leading track, but very narrow in z, running a Fast Track Finder algorithm followed by the Precision Tracking. Statistical, Bayesian uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for tracks from offline medium quality 1-prong tau candidates with offline transverse momentum (pT) greater than 20 GeV as a function of the mean interaction multiplicity per bunch crossing. The efficiency is evaluated with a 25 GeV tau trigger configured to select only on the calorimeter cluster. Offline tracks are required to have no missing hits from those expected in the pixel detector, and a tight overall selection on the total number of silicon hits. For the tau triggers the ID reconstruction runs in two stages - the first stage running a Fast Track Finder algorithm in a narrow region around the calorimeter cluster direction, but extended along the entire luminous region at the beamline to identify the leading tracks and the z position of the interaction, and the second stage running in a wider region about this leading track, but very narrow in z, running a Fast Track Finder algorithm followed by the Precision Tracking. Statistical, Bayesian uncertainties are shown.
png pdf

2017 Trigger Performance Plots

ATL-COM-DAQ-2017-107 The ATLAS ID Trigger performance in 2017 data

The track finding efficiency of the Inner Detector (ID) trigger for muons with pT > 4 GeV from medium quality offline muon candidates, shown as a function of the offline muon pT. The efficiency is evaluated for both the 10 GeV and 24 GeV muon triggers running in a mode where the trigger decision is made based on early muon candidates reconstructed from the Muon Spectrometer information only and so can contain candidates where the full offline reconstructed muons have a pT lower than the trigger threshold. The offline muons are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detector layers. The ID trigger first runs a Fast Track Finder stage followed by a detailed Precision Tracking stage to refine the track candidates identified in the first stage and improve their quality. Statistical, Bayesian uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for muons with pT > 4 GeV from medium quality offline muon candidates, shown as a function of the offline muon pseudorapidity, . The efficiency is evaluated for both the 10 GeV and 24 GeV muon triggers running in a mode where the trigger decision is made based on early muon candidates reconstructed from the Muon Spectrometer information only and so can contain candidates where the full offline reconstructed muons have a pT lower than the trigger threshold. The offline muons are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detector layers. The ID trigger first runs a Fast Track Finder stage followed by a detailed Precision Tracking stage to refine the track candidates identified in the first stage and improve their quality. Statistical, Bayesian uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for muons with pT > 4 GeV from medium quality offline muon candidates, shown as a function of the offline muon z0 position. The efficiency is evaluated for both the 10 GeV and 24 GeV muon triggers running in a mode where the trigger decision is made based on early muon candidates reconstructed from the Muon Spectrometer information only and so can contain candidates where the full offline reconstructed muons have a pT lower than the trigger threshold. The offline muons are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detector layers. The ID trigger first runs a Fast Track Finder stage followed by a detailed Precision Tracking stage to refine the track candidates identified in the first stage and improve their quality. Statistical, Bayesian uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for muons with pT > 4 GeV from medium quality offline muon candidates, shown as a function of the offline muon transverse impact parameter, d0. The efficiency is evaluated for both the 10 GeV and 24 GeV muon triggers running in a mode where the trigger decision is made based on early muon candidates reconstructed from the Muon Spectrometer information only and so can contain candidates where the full offline reconstructed muons have a pT lower than the trigger threshold. The offline muons are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detector layers. The apparent loss of efficiency at large |d0| is an artefact of the offline track reconstruction for poorly reconstructed offline tracks with no inner pixel layer hit when one might be expected. For these tracks d0 is less well reconstructed and the tracks thereby migrate to larger values reducing the purity of tracks with a genuine large impact parameter. The ID trigger first runs a Fast Track Finder stage followed by a detailed Precision Tracking stage to refine the track candidates identified in the first stage and improve their quality. Statistical, Bayesian uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for muons with pT > 4 GeV from medium quality offline muon candidates, shown as a function of the mean number of interactions per bunch crossing. The efficiency is evaluated for both the 10 GeV and 24 GeV muon triggers running in a mode where the trigger decision is made based on early muon candidates reconstructed from the Muon Spectrometer information only and so can contain candidates where the full offline reconstructed muons have a pT lower than the trigger threshold. The offline muons are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detector layers. The ID trigger first runs a Fast Track Finder stage followed by a detailed Precision Tracking stage to refine the track candidates identified in the first stage and improve their quality. Statistical, Bayesian uncertainties are shown.
png pdf
The trigger z-track resolution of the Inner Detector (ID) trigger for muons with pT > 4 GeV from medium quality offline muon candidates, shown as a function of the offline muon pseudorapidity. The resolution is evaluated for both the 10 GeV and 24 GeV muon triggers running in a mode where the trigger decision is made based on early muon candidates reconstructed from the Muon Spectrometer information only and so can contain candidates where the full offline reconstructed muons have a pT lower than the trigger threshold. Offline muons are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detectors. The ID trigger first runs a Fast Track Finder stage followed by a detailed Precision Tracking stage to refine the track candidates identified in the first stage and improve their quality. Statistical uncertainties are shown.
png pdf
The trigger z-track resolution of the Inner Detector (ID) trigger for muons with pT > 4 GeV from medium quality offline muon candidates, shown as a function of the offline muon pT. The resolution is evaluated for both the 10 GeV and 24 GeV muon triggers running in a mode where the trigger decision is made based on early muon candidates reconstructed from the Muon Spectrometer information only and so can contain candidates where the full offline reconstructed muons have a pT lower than the trigger threshold. Offline muons are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detectors. The ID trigger first runs a Fast Track Finder stage followed by a detailed Precision Tracking stage to refine the track candidates identified in the first stage and improve their quality. Statistical uncertainties are shown.
png pdf
The trigger track transverse impact parameter resolution of the Inner Detector (ID) trigger for muons with pT > 4 GeV from medium quality offline muon candidates, shown as a function of the offline muon pseudorapidity. The resolution is evaluated for both the 10 GeV and 24 GeV muon triggers running in a mode where the trigger decision is made based on early muon candidates reconstructed from the Muon Spectrometer information only and so can contain candidates where the full offline reconstructed muons have a pT lower than the trigger threshold. Offline muons are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detectors. The ID trigger first runs a Fast Track Finder stage followed by a detailed Precision Tracking stage to refine the track candidates identified in the first stage and improve their quality. Statistical uncertainties are shown.
png pdf
The trigger track transverse impact parameter resolution of the Inner Detector (ID) trigger for muons with pT > 4 GeV from medium quality offline muon candidates, shown as a function of the offline muon pT. The resolution is evaluated for both the 10 GeV and 24 GeV muon triggers running in a mode where the trigger decision is made based on early muon candidates reconstructed from the Muon Spectrometer information only and so can contain candidates where the full offline reconstructed muons have a pT lower than the trigger threshold. Offline muons are required to have at least one pixel clusters and at least 4 SCT clusters, with no more than two holes in the silicon detectors.The ID trigger first runs a Fast Track Finder stage followed by a detailed Precision Tracking stage to refine the track candidates identified in the first stage and improve their quality. Statistical uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for tracks with pT > 1 GeV within jets shown as a function of the offline track pT. The efficiency is evaluated the 55 GeV jet trigger. Offline tracks are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detectors and, if expected, should have at least one hit in the innermost pixel layer. For the jet and Bjet triggers the reconstruction in the ID trigger runs in three stages - the first stage runs a Fast Track Finder algorithm in very narrow regions about the directions of any jet above 30 GeV, where the regions are fully extended along the beam line, to identify tracks above 5 GeV that are then used to identify the primary vertex. The z-position of the vertex is then used to restrict the reconstruction of tracks in each jet to a narrow region in z about the primary vertex, but wider ( +- 0.4 ) in both eta and phi for the second and third stages consisting of Fast Track Finder, and Precision Tracking stages. Statistical, Bayesian uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for tracks with pT > 1 GeV within jets shown as a function of the offline track pseudorapidity. The efficiency is evaluated the 55 GeV jet trigger. Offline tracks are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detectors and, if expected, should have at least one hit in the innermost pixel layer. For the jet and Bjet triggers the reconstruction in the ID trigger runs in three stages - the first stage runs a Fast Track Finder algorithm in very narrow regions about the directions of any jet above 30 GeV, where the regions are fully extended along the beam line, to identify tracks above 5 GeV that are then used to identify the primary vertex. The z-position of the vertex is then used to restrict the reconstruction of tracks in each jet to a narrow region in z about the primary vertex, but wider ( +- 0.4 ) in both eta and phi for the second and third stages consisting of Fast Track Finder, and Precision Tracking stages. Statistical, Bayesian uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for tracks with pT > 1 GeV within jets shown as a function of the offline track z0. The efficiency is evaluated the 55 GeV jet trigger. Offline tracks are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detectors and, if expected, should have at least one hit in the innermost pixel layer. For the jet and Bjet triggers the reconstruction in the ID trigger runs in three stages - the first stage runs a Fast Track Finder algorithm in very narrow regions about the directions of any jet above 30 GeV, where the regions are fully extended along the beam line, to identify tracks above 5 GeV that are then used to identify the primary vertex. The z-position of the vertex is then used to restrict the reconstruction of tracks in each jet to a narrow region in z about the primary vertex, but wider ( +- 0.4 ) in both eta and phi for the second and third stages consisting of Fast Track Finder, and Precision Tracking stages. Statistical, Bayesian uncertainties are shown.
png pdf
The track finding efficiency of the Inner Detector (ID) trigger for tracks with pT > 1 GeV within jets shown as a function of the multiplicity of the mean number of pileup interactions in the event. The efficiency is evaluated the 55 GeV jet trigger. Offline tracks are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detectors and, if expected, should have at least one hit in the innermost pixel layer. For the jet and Bjet triggers the reconstruction in the ID trigger runs in three stages the first stage runs a Fast Track Finder algorithm in very narrow regions about the directions of any jet above 30 GeV, where the regions are fully extended along the beam line, to identify tracks above 5 GeV that are then used to identify the primary vertex. The z-position of the vertex is then used to restrict the reconstruction of tracks in each jet to a narrow region in z about the primary vertex, but wider ( +- 0.4 ) in both eta and phi for the second and third stages consisting of Fast Track Finder, and Precision Tracking stages. Statistical, Bayesian uncertainties are shown.
png pdf
The track transverse impact parameter resolutions for the Inner Detector (ID) trigger for tracks with pT > 1 GeV within jets shown as a function of the offline track pseudorapidity. The resolution is evaluated the 55 GeV jet trigger. Offline tracks are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detectors and, if expected, should have at least one hit in the innermost pixel layer. For the jet and Bjet triggers the reconstruction in the ID trigger runs in three stages - the first stage runs a Fast Track Finder algorithm in very narrow regions about the directions of any jet above 30 GeV, where the regions are fully extended along the beam line, to identify tracks above 5 GeV that are then used to identify the primary vertex. The z-position of the vertex is then used to restrict the reconstruction of tracks in each jet to a narrow region in z about the primary vertex, but wider ( +- 0.4 ) in both eta and phi for the second and third stages consisting of Fast Track Finder, and Precision Tracking stages. Statistical uncertainties are shown.
png pdf
The track z0 resolutions for the Inner Detector (ID) trigger for tracks with pT > 1 GeV within jets shown as a function of the offline track pseudorapidity. The resolution is evaluated the 55 GeV jet trigger. Offline tracks are required to have at least one pixel cluster and at least 4 SCT clusters, with no more than two holes in the silicon detectors and, if expected, should have at least one hit in the innermost pixel layer. For the jet and Bjet triggers the reconstruction in the ID trigger runs in three stages - the first stage runs a Fast Track Finder algorithm in very narrow regions about the directions of any jet above 30 GeV, where the regions are fully extended along the beam line, to identify tracks above 5 GeV that are then used to identify the primary vertex. The z-position of the vertex is then used to restrict the reconstruction of tracks in each jet to a narrow region in z about the primary vertex, but wider ( +- 0.4 ) in both eta and phi for the second and third stages consisting of Fast Track Finder, and Precision Tracking stages. Statistical uncertainties are shown.
png pdf

ATL-COM-DAQ-2016-121 Inner Detector trigger timing study

The trigger track reconstruction time for the beamspot trigger for 14 TeV t ̄t Monte Carlo simulated with 46, 69 and 138 interactions per bunch crossing, measured on a 2.4 GHz Intel Xeon CPU. The software version used corresponds to the 2016 online trigger system. Statistical uncertainties are shown. A second-order polynomial is fit to the points.
png pdf

Phase-I Upgrade public plots

Timing of the Inner Detector (ID) track seeding algorithm for the full detector. The red dots represent the standard HLT ID algorithm, running on a single CPU core, the blue dots show the same logic algorithm ported to GPU. The track seeding algorithm is the first part of the track finding algorithm which produces triplets of spacepoints known as track seeds. The timing is shown as a function of the number of spacepoints in an event.
png eps pdf
Timings of the Inner Detector (ID) full-detector data preparation implemented in CUDA® performed on a NVIDIA® Tesla® K40m Graphical Processor Unit (GPU) compared with the standard C++ implementation of the ID trigger full detector data preparation running on a single core of an Intel® Xeon® E5-2670 2.6 GHz CPU. The measurements were made with a Monte Carlo simulated data sample of 𝑡𝑡 ̅ events at a centre of mass energy of 14 TeV with a mean of 46 interactions per bunch crossing. The mean execution time per event, for this dataset, was a factor of 21 faster for the GPU implementation. The data preparation consists of bytestream decoding, hit clustering, and spacepoint formation in both the Pixel and SCT detectors. The input data volume is the combined size of Pixel and SCT bytestream data.
png pdf

2015 Trigger Performance Plots

ATL-COM-DAQ-2016-083 Run 2 HLT tracking performance from 25ns running

The electron track finding efficiency of the Inner Detector (ID) trigger, shown for offline tracks with pT > 20 GeV from tight offline electron candidates, shown as a function of the offline track pseudorapidity. The efficiency is evaluated for the 24 GeV electron trigger. Offline tracks are required to have at least two pixel clusters, and at least six SCT clusters. The reconstruction in the ID trigger runs in two stages - the first runs a Fast Track Finder algorithm for fast pattern recognition, and the second stage runs a more detailed track fit using the clusters identified in the Fast Track Finder and using a track fitting algorithm from the offline processing. This strategy is adopted to accommodate the high rate and pileup expected during LHC run 2. Statistical, Bayesian uncertainties are shown.
png pdf
The electron track finding efficiency of the Inner Detector (ID) trigger, shown for offline tracks with |η| < 2.4 from tight offline electron candidates, shown as a function of the offline track transverse momentum. The efficiency is evaluated for the 24 GeV electron trigger. Offline tracks are required to have at least two pixel clusters, and at least six SCT clusters. The reconstruction in the ID trigger runs in two stages - the first runs a Fast Track Finder algorithm for fast pattern recognition, and the second stage runs a more detailed track fit using the clusters identified in the Fast Track Finder and using a track fitting algorithm from the offline processing. This strategy is adopted to accommodate the high rate and pileup expected during LHC run 2. Statistical, Bayesian uncertainties are shown.
png pdf
The muon track finding efficiency of the Inner Detector (ID) trigger, shown for offline muon candidates with |η| < 2.4, shown as a function of the offline track transverse momentum. The efficiency is evaluated for the 6 GeV muon trigger. Offline muon candidates are required to have at least two pixel clusters, and at least six SCT clusters. The reconstruction in the ID trigger runs in two stages - the first runs a Fast Track Finder algorithm for fast pattern recognition, and the second stage runs a more detailed track fit using the clusters identified in the Fast Track Finder and using a track fitting algorithm from the offline processing. This strategy is adopted to accommodate the high rate and pileup expected during LHC run 2. Statistical, Bayesian uncertainties are shown.
png pdf
The resolution for the transverse impact parameter of tracks reconstructed with the Inner Detector (ID) trigger, shown for offline muon candidates with |η| < 2.4, shown as a function of the offline muon transverse momentum from muon candidates collected with the 6 GeV muon trigger. Offline muon candidates are required to have at least two pixel clusters, and at least six SCT clusters. The reconstruction in the ID trigger runs in two stages - the first runs a Fast Track Finder algorithm for fast pattern recognition, and the second stage runs a more detailed track fit using the clusters identified in the Fast Track Finder and using a track fitting algorithm from the offline processing. This strategy is adopted to accommodate the high rate and pileup expected during LHC run 2. Statistical, Bayesian uncertainties are shown.
png pdf
The vertex finding efficiency of the Inner Detector (ID) trigger from the b-jet trigger. Vertex candidates are found using tracks reconstructed in the trigger from within all Regions of Interest (RoI) identified by jets with ET > 30 GeV found in the HLT using the anti-kT algorithm. The efficiency is shown as a function of the multiplicity of offline tracks with pT >1 GeV and |η| < 2.4 that lie within all the jet RoI in the event. Offline tracks are required to have at least two pixel clusters, and at least six SCT clusters. For speed the tracking for the vertexing runs only a single stage fast track finding stage since the vertex position is primarily used in the trigger only to update the RoI position for the subsequent precision track reconstruction and b-tagging, and so the full precision tracking is not required. Statistical, Bayesian uncertainties are shown.
png pdf
The resolution of z-vertex position of the Inner Detector (ID) trigger from the b-jet trigger. Vertex candidates are found using tracks reconstructed in the trigger from within all Regions of Interest (RoI) identified by jets with ET > 30 GeV found in the HLT chains with using the anti-kT algorithm. The resolution is shown as a function of the multiplicity of offline tracks with pT >1 GeV and |η| < 2.4 that lie within all the jet RoI in the event. Offline tracks are required to have at least two pixel clusters, and at least six SCT clusters.. For speed the tracking for the vertexing runs only a single stage fast track finding stage since the vertex position is primarily used in the trigger only to update the RoI position for the subsequent precision track reconstruction and b-tagging, and so the full precision tracking is not required. Statistical, Bayesian uncertainties are shown.
png pdf

ATL-COM-DAQ-2015-148 Run 2 HLT tracking timing plots

The Run 2 HLT Inner Detector tracking trigger processing time for the Fast Track Finder stage for the tau signature. Shown are the times for the single-stage, and the two-stage tracking. In the single-stage tracking, the tracking is performed in a single, large Region of Interest (RoI) with Δη = 0.4, Δφ=0.4 and Δz = 225 mm with respect to the RoI direction and position z=0 along the beamline. In the two-stage tracking, the tracking is first performed in an RoI with Δη = 0.1, Δφ=0.1 and Δz = 225 mm with respect to the RoI direction, to identify the core tracks, and then a second tracking stage is performed in an updated RoI centred on the highest pT track with Δη = 0.4, Δφ=0.4 and Δz = 10 mm with respect to that track. The total mean time for the two-stage tracking is 44.5 ms corresponding to a fractional saving in processing time for the fast tracking with respect to the single-stage tracking of greater than 30%. The data were taken during collisions in August 2015 with the LHC colliding with a 25 ns bunch spacing. The mean number of interactions per bunch crossing was <μ> ~ 14.
png pdf
The Run 2 HLT Inner Detector tracking trigger processing time for the Precision Tracking stage for the tau trigger. Shown are the times for the single-stage, and the two-stage tracking. In the single-stage tracking, the tracking is performed in a single, large Region of Interest (RoI) with Δη = 0.4, Δφ=0.4 and Δz = 225 mm with respect to the RoI direction and z=0 along the beamline. In the two-stage tracking, the precision tracking is performed only in the second stage RoI with Δη = 0.4, Δφ=0.4 and Δz = 10 mm, centred on the highest pT track reconstructed in the first stage. The data were taken during collisions in August 2015 with the LHC colliding with a 25 ns bunch spacing. The mean number of interactions per bunch crossing was <μ> ~ 14.
png pdf
A figure illustrating the RoIs from the one-stage tracking (pink) and two-stage tracking (blue - first stage, green - second stage)
png pdf

2015 Trigger Performance Plots

ATL-COM-DAQ-2015-110 Run 2 HLT tracking performance plots for 50 ns running

The Run 2 HLT Inner Detector tracking efficiency in the Minimum Bias Trigger shown as a function of offline track PT for good quality offline tracks with at least 2 pixel clusters and 6 SCT clusters. The offline tracks are required to be in the region |ηoffline| < 2.5. The closest matching trigger track within a cone of ΔR < 0.05 of the offline track is chosen. The data were taken during the early collisions in June 2015 with a mean number of interactions per bunch crossing of <μ> ~ 10. For the Minimum Bias Trigger the tracking runs as a single Precision Tracking stage. Bayesian uncertainties are shown.
png pdf
The Run 2 HLT Inner Detector tracking efficiency in the Minimum Bias Trigger shown as a function of offline track pseudorapidity for good quality offline tracks with at least 2 pixel clusters and 6 SCT clusters. The offline tracks are required to be in the region |ηoffline| < 2.5 and pT > 1 GeV. The closest matching trigger track within a cone of ΔR < 0.05 of the offline tracks is chosen. The data were taken during the early collisions in June 2015 with a mean number of interactions per bunch crossing of <μ> ~ 10. For the Minimum Bias Trigger the tracking runs as a single Precision Tracking stage. Bayesian uncertainties are Data 13 TeV, minimum Bias Trigger minBias Precision Tracking minBias Precision Tracking shown.
png pdf
The Run 2 HLT Inner Detector tracking efficiency in the Minimum Bias Trigger shown as a function of offline track transverse impact parameter, d0, with respect to the beamline for good quality offline tracks with at least 2 pixel clusters and 6 SCT clusters. The offline tracks are required to be in the region |ηoffline| < 2.5 and pT > 1 GeV. The closest matching trigger track within a cone of ΔR < 0.05 of the offline tracks is chosen. The data were taken during the early collisions in June 2015 with a mean number of interactions per bunch crossing of <μ> ~ 10. For the Minimum Bias Trigger the tracking runs as a single Precision Tracking stage. Bayesian uncertainties are shown.
png pdf
The residual in pseudorapidity between the HLT Inner Detector trigger track and the offline track for good offline tracks from the Minimum Bias trigger where the offline tracks are required to have at least 2 pixel clusters and 6 SCT clusters. The offline tracks are required to be in the region |ηoffline| < 2.5 and pT > 1 GeV. The closest matching trigger track within a cone of ΔR < 0.05 of the offline track is chosen. The data were taken during the early collisions in June 2015 with a mean number of interactions per bunch crossing of <μ > ~ 10. For the Minimum Bias Trigger the tracking runs as a single Precision Tracking stage.
png pdf
The Run 2 HLT Inner Detector tracking efficiency in the 24 GeV Electron Trigger shown as a function of the PT of the Electron track for tight offline electron candidates and where the offline track is required to have at least 2 pixel clusters and 6 SCT clusters. The offline electron tracks are required to lie in the region |ηoffline| < 2.5 and pT > 20 GeV. The closest matching trigger track within a cone of ΔR < 0.05 of the offline track is chosen. The data were taken during early July 2015. The HLT tracking for the Electron trigger runs a Fast Track Finder stage followed by a Precision Tracking stage. Bayesian uncertainties are shown.
png pdf
The Run 2 HLT Inner Detector tracking efficiency in the 24 GeV Electron Trigger shown as a function of the pseudorapidity of the Electron track for tight offline electron candidates and where the offline track is required to have at least 2 pixel clusters and 6 SCT clusters. The offline electron tracks are required to be in the region |ηoffline| < 2.5 and pT>20 GeV. The closest matching trigger track within a cone of ΔR < 0.05 of the offline track is chosen. The data were taken during early July 2015. The HLT tracking for the Electron trigger runs a Fast Track Finder stage followed by a Precision Tracking stage. Bayesian uncertainties are shown.
png pdf
The residual in 1/pT between the HLT Inner Detector trigger track and the Electron track for tight offline electron candidates passing the 24 GeV medium Electron trigger where the offline track is required to have at least 2 pixel clusters and 6 SCT clusters. The offline electron tracks are required to be in the region |ηoffline| < 2.5 and pT>20 GeV. The closest matching trigger track within a cone of ΔR < 0.05 of the offline track is chosen. The data were taken during early July 2015. The HLT tracking for the Electron trigger runs a Fast Track Finder stage followed by a Precision Tracking stage.
png pdf
The Run 2 HLT Inner Detector tracking efficiency in the 10 GeV Muon Trigger shown as a function of the PT of the muon track for offline muon candidates where the offline muon track is required to have at least 2 pixel clusters and 6 SCT clusters. The offline muons tracks are required to be in the region |ηoffline| < 2.5 and pT > 10 GeV. The closest matching trigger track within a cone of ΔR < 0.05 of the offline track is chosen. The data were taken during early July 2015. The HLT tracking for the Muon trigger runs a Fast Track Finder stage followed by a Precision Tracking stage. Bayesian uncertainties are shown.
png pdf
The Run 2 HLT Inner Detector tracking efficiency in the 10 GeV Muon Trigger shown as a function of the pseudorapidity of the muon track for offline muon candidates where the offline muon track is required to have at least 2 pixel clusters and 6 SCT clusters. The offline muon tracks are required to be in the region |ηoffline| < 2.5 and pT > 10 GeV. The closest matching trigger track within a cone of ΔR < 0.05 of the offline track is chosen. The data were taken during early July 2015. The HLT tracking for the Muon trigger runs a Fast Track Finder stage followed by a Precision Tracking stage. Bayesian uncertainties are shown.
png pdf
The Run 2 HLT Inner Detector tracking resolution on the transverse impact parameter with respect to the beamline in the 10 GeV Muon Trigger shown as a function of the pseudorapidity of the muon track for offline muon candidates where the offline muon track is required to have at least 2 pixel clusters and 6 SCT clusters. The offline muons tracks are required to be in the region |ηoffline| < 2.5 and pT > 10 GeV. The closest matching trigger track within a cone of ΔR < 0.05 of the offline track is chosen. The data were taken during early July 2015. The HLT tracking for the Muon trigger runs a Fast Track Finder stage followed by a Precision Tracking stage.
png pdf
The Run 2 HLT Inner Detector tracking resolution on the track pseudorapidity in the 10 GeV Muon Trigger shown as a function of the pseudorapidity of the muon track for offline muon candidates where the offline muon track is required to have at least 2 pixel clusters and 6 SCT clusters. The offline muons tracks are required to be in the region |ηoffline| < 2.5 and pT > 10 GeV. The closest matching trigger track within a cone of ΔR < 0.05 of the offline track is chosen. The HLT tracking for the Muon trigger runs a Fast Track Finder stage followed by a Precision Tracking stage. The data were taken during early July 2015.
png pdf
The residual in 1/pT between the HLT Inner Detector trigger track and the Muon track for Muon candidates passing the 10 GeV Muon trigger where the offline Muon track is required to have at least 2 pixel clusters and 6 SCT clusters. The offline muon tracks are required to be in the region |ηoffline| < 2.5 and pT > 10 GeV. The closest matching trigger track within a cone of ΔR < 0.05 of the offline track is chosen. The data were taken during early July 2015.
png pdf

2015 Trigger development Pubnote

ATL-DAQ-PUB-2015-026 Inner Detector trigger timing studies for Run 2

The full trigger tau reconstruction time for 14 TeV ttbar Monte Carlo with a mean of 46 interactions per bunch crossing, measured on a 2.4 GHz Intel Xeon CPU. The timing is dominated by track finding - only 36 ms is spent in calorimetry. Shown are the times for two strategies; in the first, “One-step”, strategy, tracks are reconstructed within a tau candidate Region of Interest (RoI) of size ∆η × ∆φ = 0.4 × 0.4 pointing to a calorimeter cluster, and with a spread along the beamline of |z0| < 225 mm, corresponding to the size of the interaction region. There is a mean of approximately 3 tau candidates per event. In the second (“Two-step”) trigger strategy, tracks are first reconstructed within a small RoI of size ∆η × ∆φ = 0.1 × 0.1 with the same z0 extent. Further tracks are then reconstructed in a wider (in η and φ) RoI with ∆η × ∆φ = 0.4 × 0.4, centred on the leading (highest-pT ) track from the previous step. This RoI is constrained to |Δz0| < 10 mm with respect to the perigee of the leading track. Splitting the tracking into these two steps reduces the non-linear dependence of the tracking on occupancy with minimal effect on efficiency.
png pdf
The processing time per event for two tau trigger strategies, shown for different pileup interaction multiplicities per bunch crossing of 46, 69 and 138, measured with 14 TeV ttbar Monte Carlo running on a 2.4 GHz Intel Xeon CPU. The timing is dominated by tracking - only between 36 and 50 ms is spent in calorimetry. In the first, “One-step”, trigger strategy tracks are reconstructed within a tau candidate Region of Interest (RoI) of size ∆η × ∆φ = 0.4 × 0.4 pointing to a calorimeter cluster and with a spread along the beamline of |z0|< 225 mm corresponding to the size of the interaction region. In the second (“Two-step”) trigger strategy, tracks are first reconstructed within a smaller RoI of size ∆η × ∆φ = 0.1 × 0.1 with the same z0 extent. Additional tracks are then reconstructed in a wider RoI (in η and φ) with ∆η × ∆φ = 0.4 × 0.4, centred on the leading (highest-pT ) track from the previous step. This RoI is constrained to |Δz0| < 10 mm with respect to the perigee of the leading track.
png pdf

2014 Trigger development Public plots

ATL-COM-DAQ-2014-088 ID Trigger timing and efficiency results

The Event Filter Inner Detector (EFID) trigger track reconstruction time for the ATLAS Muon trigger for 14 TeV Z to mu+mu- Monte Carlo, measured running 2.3-2.8 GHz Xeon processors. The ATLAS Trigger code for these measurements was built on April 18th 2013. Shown are the times versus the mean number of pileup interactions for three different pileup multiplicities; 46, 69 and 139 interactions per bunch crossing. Shown are the total EF ID tracking time, and the times for the two most time consuming tracking components - the main pattern recognition algorithm, and the Ambiguity Solver - which together constitute nearly 90% of the ID tracking time at the Event Filter. Also shown are fits to the data points using a quadratic function.
png pdf
The distribution of processing times per call for the pattern recognition stage of the Event Filter Inner Detector (EFID) trigger tracking measured using Z→e+e- Monte Carlo running on a 2.4 GHz Xeon processor. Shown are the times for the tracking strategy used during the ATLAS Run 1 data taking in 2012 but implemented in different versions of the ATLAS code; one built in April 2013 and a more recent build from August 2014, with mean processing times of 297 ms and 100 ms per call respectively. The code from 2014 is significantly faster than that from 2013 as the offline tracking algorithms used in the Event Filter tracking underwent significant optimisation prior to the later build. Besides the Z to e+e- signal interaction, the Monte Carlo sample has a mean number of additional pileup interactions per bunch crossing, , equal to 46.
png pdf
The distribution of processing times per call for the Ambiguity Solver stage of the Event Filter Inner Detector (EFID) tracking measured using Z to e+e- Monte Carlo running on a 2.4 GHz Xeon processor. Shown are the times for the tracking strategy used during ATLAS Run 1 data taking in 2012 but implemented in different versions of the ATLAS code; one built in April 2013 and a more recent build from August 2014, with mean processing times of 129 ms and 13.3 ms per call respectively. The code from 2014 is significantly faster than that from 2013 as the offline tracking algorithms used in the Event Filter tracking underwent significant optimisation prior to the later build. Besides the Z to e+e- signal i nteraction, the Monte Carlo sample has a mean number of additional pileup interactions per bunch crossing, equal to 46.
png pdf
The distribution of total trigger processing time for the complete processing for the 24 GeV isolated electron trigger, from the ATLAS High Level Trigger configured to run on a single node running on Z to e+e- Monte Carlo running on a 2.4 GHz Xeon processor. Shown are the results from the development version of the ATLAS Trigger code built in August 2014 for the full electron trigger, with two alternative "strategies" for the ID Trigger; The "Run 1 strategy" runs the same algorithms that were executed during Run 1 data taking, which consisted of a fast tracking stage followed by a modified version of the full offline precision tracking, running the pattern recognition and Ambiguity Solver stages. In contrast, the "Run 2 strategy" [1] also runs a fast tracking stage, but then performs the precision tracking by directly seeding the ambiguity solver stage from the output of this fast tracking stage rather than running the offline pattern recognition. The full trigger processing times per event are shown and include the time spent running the chain multiple times in events with more than one electron candidate. The times per candidate are approximately 40% faster than the per event times. It should be noted that the total trigger processing includes that for the calorimeter reconstruction and additional, non-tracking, algorithms, which are common to both strategies and contribute approximately 22 ms to the total event processing time in each case. Besides the Z to e+e- signal interaction, the Monte Carlo sample has a mean number additional pileup interactions per bunch crossing, , equal to 46. [1] ATL-DAQ-PUB-2013-002
png pdf
The distribution of processing time for the Ambiguity Solver from the ATLAS High Level Trigger running on Z to e+e- Monte Carlo running on a 2.4GHz Xeon processor. Shown are the results from a development version of the ATLAS Trigger code built in August 2014, with two alternative "strategies" for the ID Trigger; The "Run 1 strategy" r uns the same algorithms that were executed during Run 1 data taking, which consisted of a fast tracking stage followed by a modified version of the full offline precision tracking, running the pattern recognition and Ambiguity Solver stages. In contrast, the "Run 2 strategy" [1] also runs a fast tracking stage, but then performs the precision tracking by directly seeding the Ambiguity Solver stage from the output of this fast tracking stage rather than running the offline pattern recognition. Besides the Z to e+e- signal interaction, the Monte Carlo sample has a mean number additional pileup interactions per bunch crossing. equal to 46. [1] ATL-DAQ-PUB-2013-002
png pdf
The muon finding efficiency for the ATLAS Inner Detector trigger tracking with respect to the true muon pseudorapidity for muons from Z to mu+mu- Monte Carlo at 14 TeV with a mean number of pileup interactions per bunch crossing, equal to 40. The efficiency is shown for the "Run 2 strategy" [1] which runs a Fast Tracking stage, the output of which is used to directly seed a Precision Tracking stage. Shown are the efficiencies for both the Fast, and the Precision Tracking. The efficiency is defined for trigger tracks matched to within DeltaR < 0.05 of a true muon direction and where the true muon pT > 3 GeV. The simulated events contain hits from the new ATLAS Insertable B-layer. [1] ATL-DAQ-PUB-2013-002
png pdf
The muon finding efficiency for the ATLAS Inner Detector trigger tracking with respect to the true muon transverse momentum for muons from Z to mu+mu- Monte Carlo 14 TeV with a mean number of pileup interactions per bunch crossing, , equal to 40. The efficiency is shown for the "Run 2 strategy" [1] which consists of a Fast Tracking stage, the output of which is used to directly seed a Precision Tracking stage. Shown are the efficiencies for both the Fast, and the Precision Tracking. The efficiency is defined for trigger tracks matched to within DeltaR < 0.05 of a true muon direction and where the modulus of the true muon pseudorapidity |eta| < 2.5. The simulated events contain hits from the new ATLAS Insertable B-layer. [1] ATL-DAQ-PUB-2013-002
png pdf

2013 Trigger development Pubnote

ATL-DAQ-PUB-2013-002 Studies for the development of the Inner Detector trigger algorithms at ATLAS

A schematic of the planned redesigned software HLT Inner Detector trigger


png pdf

EF reconstructed d0 resolution as a function of eta. The hollow (red) points are without IBL, the solid points are with IBL.


png pdf

EF reconstructed d0 resolution as a function of signed pt. The hollow (red) points are without IBL, the solid points are with IBL.


png pdf

EF reconstructed eta resolution as a function of eta. The hollow (red) points are without IBL, the solid points are with IBL.


png pdf

EF reconstructed eta resolution as a function of pt. The hollow (red) points are without IBL, the solid points are with IBL.


png pdf

EF reconstructed zed resolution as a function of eta. The hollow (red) points are without IBL, the solid points are with IBL.


png pdf

EF reconstructed zed resolution as a function of pt. The hollow (red) points are without IBL, the solid points are with IBL.


png pdf

L2 reconstructed d0 resolution as a function of eta. The hollow (red) points are without IBL, the solid points are with IBL.


png pdf

L2 reconstructed d0 resolution as a function of pt. The hollow (red) points are without IBL, the solid points are with IBL.


png pdf

L2 reconstructed eta resolution as a function of eta. The hollow (red) points are without IBL, the solid points are with IBL.


png pdf

L2 reconstructed eta resolution as a function of pt. The hollow (red) points are without IBL, the solid points are with IBL.


png pdf

L2 reconstructed z resolution as a function of eta. The hollow (red) points are without IBL, the solid points are with IBL.


png pdf

L2 reconstructed z resolution as a function of pT. The hollow (red) points are without IBL, the solid points are with IBL.


png pdf

Relative time per event spent in Level 2 Strategy A algorithms. The timing values were taken from the processing of 1367 events collected in the 2012 running period.


png pdf

Number of CPU instruction fetches per event generated collected by the callgrind profiling tools Functions with the highest number of instruction fetches are illustrated. The software library containing each function is shown in parentheses.


png pdf

Total number of unhalted CPU cycles sampled in the ID Trigger function findZinternal. Stalled cycles are a subset of the total unhalted cycles and load latency, branch-stalled cycles are a subset of the total unhalted cycles and load latency, branch-misprediction and instruction latency are a subset of the total stalled cycles.


png pdf

Comparing execution time for two versions of the Z-Finder algorithm, before and after optimisation to reduce branch mis-prediction.


png pdf

Comparing execution time for three versions of the test method. Two versions have identical code, but auto-vectorisation has been enabled in one case. The third has had vectorisation explicitly introduced using SSE intrinsics (auto-vectorisation is also enabled).


png pdf

2012 Data @ 8 TeV

ATL-COM-DAQ-2013-064 HLT tracking performance in 2012

The L2 and EF tracking efficiencies measured using a Tag & Probe analysis for probe electron candidates with ET>15 GeV and that are located inside L2 electron Regions of Interest, shown as a function of the offline track pseudorapidity. The trigger selects electron candidates based on the reconstruction in the Calorimeter only. Errors are statistical only.


png pdf

The L2 and EF tracking efficiencies measured using a Tag & Probe analysis for probe electron candidates with ET>15 GeV and that are located inside L2 electron Regions of Interest, shown as a function of the offline electron track pT. The trigger selects electron candidates based on the reconstruction in the Calorimeter only. Errors are statistical only.


png pdf

The L2 and EF tracking efficiencies measured using a Tag & Probe analysis for probe electron candidates with ET>15 GeV and that are located inside L2 electron Regions of Interest, shown as a function of the mean number of interactions per bunch crossing. The trigger selects electron candidates based on the reconstruction in the Calorimeter only. Errors are statistical only.


png pdf

The L2 and EF tracking efficiencies measured using a Tag & Probe analysis for probe electron candidates with ET>15 GeV and that are located inside L2 electron Regions of Interest, shown as a function of the ratio of the offline electron track pT to the offline cluster ET. The trigger selects electron candidates based on the reconstruction in the Calorimeter only. Errors are statistical only.


png pdf

ATL-COM-DAQ-2012-061 HLT Tracking Performance in electron and muon triggers in 2012 Data

L2 and EF Tracking efficiency for offline electron tracks of |η|<2.5 and that are located inside L2 electron Regions of Interest, shown as a function of the transverse momentum of the offline electron. The trigger monitoring selects high ET electron objects with a threshold of 24 GeV. Entries below this threshold represent candidates that have undergone significant bremstrahlung or originate from background mimicking an electron candidate. Errors are statistical only.

electrons-2012-PeriodA4.FinalDecorations.FinalLabels.NoGrid.pT_eff.png
png eps

L2 and EF Tracking efficiency for offline electron tracks of |η|<2.5 and that are located inside L2 electron Regions of Interest, shown as a function of the pseudorapidity of the offline electron. The trigger monitoring selects high ET electron objects with a threshold of 24 GeV. Errors are statistical only.


png eps

L2 and EF Tracking efficiency for offline electron tracks of |η|<2.5 and that are located inside L2 electron Regions of Interest, shown as a function of the number of vertices per event. The trigger monitoring selects high ET electron objects with a threshold of 24 GeV. Errors are statistical only.


png eps

L2 and EF Tracking efficiency for offline electron tracks of |η|<2.5 and that are located inside L2 electron Regions of Interest, shown as a function of the number of offline tracks per event. The trigger monitoring selects high ET electron objects with a threshold of 24 GeV. Errors are statistical only.


png eps

L2 and EF Tracking efficiency for offline electron tracks of |η|<2.5 and that are located inside L2 electron Regions of Interest, shown as a function of the pseudorapidity of the offline electron. The trigger monitoring selects high ET electron objects with a threshold of 22 GeV (Data 2011) and 24 GeV (Data 2012). Errors are statistical only.


png eps

L2 and EF Tracking efficiency for offline electron tracks of |η|<2.5 and that are located inside L2 electron Regions of Interest, shown as a function of the number of offline tracks per event. The trigger monitoring selects high ET electron objects with a threshold of 22 GeV (Data 2011) and 24 GeV (Data 2012). Errors are statistical only.


png eps

2011 Data @ 7 TeV

ATLAS-COM-CONF-2012-054

The tracking efficiency is studied in an unbiased monitoring mode with events passing a L1 calorimetric threshold of transverse energy 29 GeV. Offline tracks found in the RoI region are used as a reference set of tracks for the efficiency calculation. The offline tracks are required to have at least seven clusters in the Si-tracker including two pixel hits, one of which is required to be in the innermost layer of the Pixel detector. A minimal track pT of 1.5 GeV is required, which corresponds to the track selection applied in the offline tau reconstruction.

The tracking efficiencies as a function of the track parameters pT (a) and η (b) of the offline track and as a function of the number of offline vertices found in the event (c), for 2011 data. The errors shown are statistical.


png eps

The tracking efficiency is studied in an unbiased monitoring mode with events passing a L1 calorimetric threshold of transverse energy 29 GeV. Offline tracks found in the RoI region are used as a reference set of tracks for the efficiency calculation. The offline tracks are required to have at least seven clusters in the Si-tracker including two pixel hits, one of which is required to be in the innermost layer of the Pixel detector. A minimal track pT of 1.5 GeV is required, which corresponds to the track selection applied in the offline tau reconstruction.

The tracking efficiencies as a function of the track parameters pT (a) and η (b) of the offline track and as a function of the number of offline vertices found in the event (c), for 2011 data. The errors shown are statistical.


png eps

The tracking efficiency is studied in an unbiased monitoring mode with events passing a L1 calorimetric threshold of transverse energy 29 GeV. Offline tracks found in the RoI region are used as a reference set of tracks for the efficiency calculation. The offline tracks are required to have at least seven clusters in the Si-tracker including two pixel hits, one of which is required to be in the innermost layer of the Pixel detector. A minimal track pT of 1.5 GeV is required, which corresponds to the track selection applied in the offline tau reconstruction.

The tracking efficiencies as a function of the track parameters pT (a) and η (b) of the offline track and as a function of the number of offline vertices found in the event (c), for 2011 data. The errors shown are statistical.


png eps

ATL-COM-DAQ-2012-038 HLT Tracking Performance for Electrons and Muons in 2011 Data

L2 and EF Tracking efficiency for offline electron tracks of |η|<2.5 and that are located inside L2 electron Regions of Interest, shown as a funcKon of the transverse momentum of the offline electron. Errors are staKsKcal only.


png eps

L2 and EF Tracking efficiency for offline electron tracks of |η|<2.5 and that are located inside L2 electron Regions of Interest, shown as a funcKon of the rapidity of the offline electron. Errors are staKsKcal only.


png eps

L2 and EF Tracking efficiency for offline electron tracks of |η|<2.5 and that are located inside L2 electron Regions of Interest, shown as a funcKon of the number of verKces per event. Errors are staKsKcal only.


png eps

L2 and EF Tracking efficiency for offline electron tracks of |η|<2.5 and that are located inside L2 electron Regions of Interest, shown as a funcKon of the number of offline tracks per event. Errors are staKsKcal only.


png eps

L2 and EF Tracking efficiency for offline muon tracks of |η|<2.5 and pT > 15 GeV and that are located inside L2 muon Regions of Interest, shown as a funcKon of the transverse momentum of the offline muon. Errors are staKsKcal only.


png eps

L2 and EF Tracking efficiency for offline muon tracks of |η|<2.5 and pT > 15 GeV and that are located inside L2 muon Regions of Interest, shown as a funcKon of the rapidity of the offline muon. Errors are staKsKcal only.


png eps

L2 and EF Tracking efficiency for offline muon tracks of |η|<2.5 and pT > 15 GeV and that are located inside L2 muon Regions of Interest, shown as a funcKon of the number of verKces per event. Errors are staKsKcal only.


png eps

L2 and EF Tracking efficiency for offline muon tracks of |η|<2.5 and pT > 15 GeV and that are located inside L2 muon Regions of Interest, shown as a funcKon of the number of offline tracks per event. Errors are staKsKcal only.


png eps

ATL-COM-DAQ-2011-085

L2 and EF Inner Detector tracking reconstruction efficiency wrt offline muon tracks that are located inside monitoring (unbiased) triggers with a transverse momentum threshold 20GeV shown as a function of the transverse momentum of the offline muon. Offline tracks are selected by cuts that reduce secondaries and choose tracks which have at least a minimum number of Si hits. (|η|<2.5, Npixel hits>0, NSCT clusters>5, |zofflineVertex|<200mm, with respect to the offline vertex: |a0|<1.5mm, |Δz sinθ|<1.5 mm.)


png eps
L2 and EF Inner Detector tracking reconstruction efficiency wrt offline electron candidates that are located inside monitoring (unbiased) triggers with a threshold 22GeV in transverse energy shown as a function of the transverse momentum of the offline electron track.

Offline tracks are selected by cuts that reduce secondaries and choose tracks which have at least a minimum number of Si hits. (|η|<2.5, Npixel hits>0, NSCT clusters>5, |zofflineVertex|<200mm, with respect to the offline vertex: |a0|<1.5mm, |Δz sinθ|<1.5 mm.)


png eps
L2 and EF Inner Detector tracking reconstruction efficiency wrt offline muon tracks that are located inside monitoring (unbiased) triggers with a threshold 20GeV in transverse momentum shown as a function of the number of vertices found by the offline reconstruction.

Small inefficiency in the L2 is related to misidentification of the primary interaction and it is addressed in algorithm tuning for higher pile-up

– Offline tracks are selected by cuts that reduce secondaries and choose tracks which have at least a minimum number of Si hits. (|η|<2.5, Npixel hits>0, NSCT clusters>5, |zofflineVertex|<200mm, with respect to the offline vertex: |a0|<1.5mm, |Δz sinθ| <1.5 mm.)


png eps
Residuals of the inverse of the track pT between Inner Detector trigger track and a matching offline Inner Detector track. The resolution is quoted as RMS95, the RMS deviation for the central 95% of the distribution. The trigger tracks come from the monitoring triggers in muon region of interest with a threshold of 20GeV.

Small differences between trigger and offline track parameters are expected as a consequence of a simplified material model and different pattern recognition (L2) and partial access to calibration (L2,EF)

Offline tracks are selected by cuts that reduce secondaries and choose tracks which have at least a minimum number of Si hits. (|η| <2.5, Npixel hits>0, NSCT clusters>5, |zofflineVertex|<200mm, with respect to the offline vertex: |a0|<1.5mm, |Δz sinθ|<1.5 mm.).


png eps

Efficiency to find a hit on the Event Filter Inner Detector trigger track in the innermost layer of the pixel detector when a hit is expected and found by the offline reconstruction. A high efficiency is important for signatures which perform rejection on the basis of a non-existent hit in the innermost pixel layer (e.g. tigher electron selection).

Whether an innermost layer hit is expected depends on the knowledge of the detector status which can be largely deduced from data and is further refined by the conditions information for the offline. The difference between offline and online calibration caused a small inefficiency in this particular run at the end of negative eta.


png eps

2010 Data @ 7 TeV

CERN-PH-EP-2011-078

Plots are on a separate page

ATL-COM-DAQ-2010-050

L2 tracking efficiency as function of η
Tracking efficiency for L2 as a function of the η of the offline tracks, measured on good offline tracks with various PT thresholds.
Trigger tracking has been run in the whole of the detector (ie. the region-of-interest is the whole of the detector). Considered tracks were identified in the Si detectors.
The threshold behaviour of the efficiency is expected and comes from using specialised algorithms optimised for L2 requirements on the track reconstruction & timing constraints.
"Good" offline tracks are selected by cuts that would reduce secondaries and choose tracks which have at least a minimum number of Si hits. (|η|<2.5, Npixel hits>0, NSCT clusters>5, |zofflineVertex|<200mm, with respect to the offline vertex: |a0|<1.5mm, |Δz sinθ|<1.5 mm.)
eff_eta_L2.png
png eps
EF tracking efficiency as function of η
Tracking efficiency for EF as a function of the η of the offline tracks, measured on good offline tracks with various PT thresholds.
Trigger tracking has been run in the whole of the detector (ie. the region-of-interest is the whole of the detector). Considered tracks were identified in the Si detectors.
The threshold behaviour of the efficiency is expected and comes from using a configuration tuning optimised for EF requirements on the track reconstruction & timing constraints.
"Good" offline tracks are selected by cuts that would reduce secondaries and choose tracks which have at least a minimum number of Si hits. (|η|<2.5, Npixel hits>0, NSCT clusters>5, |zofflineVertex|<200mm, with respect to the offline vertex: |a0|<1.5mm, |Δz sinθ|<1.5 mm.)

png eps
EF d0 residual with respect to offline
Δd0 (impact parameter) between good offline tracks of PT>1GeV and matching EF trigger tracks.
Trigger tracking has been run in the whole of the detector (ie. the region-of-interest is the whole of the detector). Considered tracks were identified in the Si detectors.
"Good" offline tracks are selected by cuts that would reduce secondaries and choose tracks which have at least a minimum number of Si hits. (|η|<2.5, Npixel hits>0, NSCT clusters>5, |zofflineVertex|<200mm, with respect to the offline vertex: |a0|<1.5mm, |Δz sinθ|<1.5 mm.)

png eps
Stability of HLT tracking efficiency over time

Efficiency for Si-based L2 and EF tracking for a run at √s=7TeV as a function of time, for good offline tracks of PT>1GeV.
Larger error bars in some bins & empty bins reflect partial statistics of the run used in this plot.
Trigger tracking has been run in the whole of the detector (ie. the region-of-interest is the whole of the detector). Considered tracks were identified in the Si detectors.
"Good" offline tracks are selected by cuts that would reduce secondaries and choose tracks which have at least a minimum number of Si hits. (|η|<2.5, Npixel hits>0, NSCT clusters>5, |zofflineVertex|<200mm, with respect to the offline vertex: |a0|<1.5mm, |Δz sinθ|<1.5 mm.)


png pdf
Mean track z position over time
Mean track z0 position obtained from L2 and EF tracks for a run at √s=7TeV as a function of time.
Larger error bars in some bins & empty bins reflect partial statistics of the run used in this plot.
Trigger tracking has been run in the whole of the detector (ie. the region-of-interest is the whole of the detector). Considered tracks were identified in the Si detectors.

png pdf
HLT tracking efficiency for muon tracks as function of η
L2 and EF Si-tracking efficiency for good offline muon tracks of PT>4GeV and that are located inside L2 muon regions-of-interest, shown as a function of the η of the offline muon.
"Good" offline tracks are selected by cuts that would reduce secondaries and choose tracks which have at least a minimum number of Si hits. (|η|<2.5, Npixel hits>0, NSCT clusters>5, |zofflineVertex|<200mm, with respect to the offline vertex: |a0|<1.5mm, |Δz sinθ|<1.5 mm.)

png pdf
HLT tracking efficiency for muon tracks as function of pT
L2 and EF Si-tracking efficiency for offline muon tracks of |η|<2.5 and that are located inside L2 muon regions-of-interest, shown as a function of the transverse momentum of the offline muon.
Offline tracks are selected by cuts that would reduce secondaries and choose tracks which have at least a minimum number of Si hits. (|η|<2.5, Npixel hits>0, NSCT clusters>5, |zofflineVertex|<200mm, with respect to the offline vertex: |a0|<1.5mm, |Δz sinθ|<1.5 mm.)

png pdf
Track parameter residuals wrt offline - Level 2 and Atlas.EventFilter
The resolution for HLT tracks with respect to the Offline tracks for the transverse impact parameter measured with respect to the offline vertex as a function of the offline track transverse momentum. The resolution is quoted as σ95%, the RMS deviation for the central 95% of the distribution.
Trigger tracking has been run in the whole of the detector (ie. the region-of-interest is the whole of the detector). Considered tracks were identified in the Si detectors.
The L2 tracking does not reproduce the offline tracks as well as the EF due to different pattern recognition algorithms and a simplified material model. Small differences in the EF wrt the offline are a consequence of a configuration optimized for trigger requirements and of partial access to the calibration in the online environment.
Offline tracks are selected by cuts that would reduce secondaries and choose tracks which have at least a minimum number of Si hits. (|η|<2.5, Npixel hits>0, NSCT clusters>5, |zofflineVertex|<200mm, with respect to the offline vertex: |a0|<1.5mm, |Δz sinθ|<1.5 mm.)

png pdf
Tracking efficiency for LVL2 tracks wrt offline using the jet instance as a function of η
Offline track selection of tracks with at least 1 space point in the pixel and 6 clusters in the SCT, pT>2GeV,|η|<2.5, |d0|<1.5mm, |z0|<200mm (d0 and z0 cuts are corrected with respect to offline vertices), geometrical one-to-one best matching with ΔR<0.05

png pdf
Tracking efficiency for LVL2 tracks wrt offline using the jet instance as a function of pT
Offline track selection of tracks with at least 1 space point in the pixel and 6 clusters in the SCT, |η|<2.5, |d0|<1.5mm, |z0|<200mm (d0 and z0 cuts are corrected with respect to offline vertices), geometrical one-to-one best matching with ΔR<0.05

png pdf
Tracking efficiency for LVL2 tracks wrt offline using the jet instance as a function of Ï•
Offline track selection of tracks with at least 1 space point in the pixel and 6 clusters in the SCT, pT>2GeV,|η|<2.5, |d0|<1.5mm, |z0|<200mm (d0 and z0 cuts are corrected with respect to offline vertices), geometrical one-to-one best matching with ΔR<0.05
A smaller efficiency at -2 reflects distribution of inactive sensors.

png pdf
Tracking efficiency for LVL2 tracks wrt offline using the tau instance as a function of η
Offline track selection of tracks with at least 1 space point in the pixel and 6 clusters in the SCT, pT>2GeV,|η|<2.5, |d0|<1.5mm, |z0|<200mm (d0 and z0 cuts are corrected with respect to offline vertices), geometrical one-to-one best matching with ΔR<0.05

png pdf
Tracking efficiency for LVL2 tracks wrt offline using the tau instance as a function of pT
Offline track selection of tracks with at least 1 space point in the pixel and 6 clusters in the SCT, |η|<2.5, |d0|<1.5mm, |z0|<200mm (d0 and z0 cuts are corrected with respect to offline vertices), geometrical one-to-one best matching with ΔR<0.05

png pdf
Tracking efficiency for LVL2 tracks wrt offline using the tau instance as a function of Ï•
Offline track selection of tracks with at least 1 space point in the pixel and 6 clusters in the SCT, pT>2GeV,|η|<2.5, |d0|<1.5mm, |z0|<200mm (d0 and z0 cuts are corrected with respect to offline vertices), geometrical one-to-one best matching with ΔR<0.05
A smaller efficiency at -2 reflects distribution of inactive sensors.

png pdf
Tracking efficiencies for the e/gamma trigger signature group
The figure shows the inidividual efficiencies of the L2 and EF tracking algorithms for medium quality electron candidates with a cluster ET of at least 5 GeV as function of the transverse momentum of the offline reconstructed electron track. Offline electron candidates from photon conversions have been excluded. Note, mainly due to brems-strahlung effects the pT of the track can be much lower than the cluster ET. The data were collected in a run on May 21th. 2010.

gif eps
Tracking efficiencies for the e/gamma trigger signature group
The figure shows efficiencies inidvidually for the L2 and EF tracking algorithms for medium quality electron candidates with a cluster ET of at least 5 GeV as function of the h direction of the offline reconstructed electron track. For this figure no additional track pT cut is applied. Offline electron candidates from photon conversions have been excluded. The data were collected in a run on May 21th. 2010.

gif eps
Number of pixel hits per EF track vs track η
Trigger tracking has been run in the whole of the detector (ie. the region-of-interest is the whole of the detector, “FullScan” instance of the algorithm). Minimum bias data and MC are compared
nPixEF.png
png eps
Number of holes per L2 track
A hole is defined as an ID layer in which an offline track has a hit but the matched L2 track does not. The matching requirement between trigger and offline tracks is ΔR<0.1. Only offline tracks passing the following criteria are considered: |η|<2.5, pT>1GeV, |z0sinθ|<1.5mm (w.r.t. primary vertex), |d0|<1.5mm (w.r.t. primary vertex), number of pixel hits > 0, number of SCT clusters > 5 Trigger tracking has been run in the whole of the detector (ie. the region-of-interest is the whole of the detector, “FullScan” instance of the algorithm). Minimum bias data and MC are compared

png eps
L2 and EF tracking efficiency vs matched offline track pT for tau instances
Match trigger tracks with all offline reconstructed tracks inside the tau RoI. The matching requirement between trigger and offline tracks is ΔR<0.1. Only offline tracks passing the following criteria are considered: |η|<2.5, |z0sinθ|<1.5mm (w.r.t. primary vertex), |d0|<1.5mm (w.r.t. primary vertex), number of b-layer hits >0, number of pixel hits > 1, number of total Si hits > 6, χ2 probability of track fit > 1%
eff-tau-pT-new.png
png pdf
L2 and EF tracking efficiency vs matched offline track η for tau instances
Match trigger tracks with all offline reconstructed tracks inside the tau RoI. The matching requirement between trigger and offline tracks is ΔR<0.1. Only offline tracks passing the following criteria are considered: |η|<2.5, pT>1GeV, |z0sinθ|<1.5mm (w.r.t. primary vertex), |d0|<1.5mm (w.r.t. primary vertex), number of b-layer hits >0, number of pixel hits > 1, number of total Si hits > 6, χ2 probability of track fit > 1%

pdf png
L2 and EF tracking efficiency vs matched offline track pT for jet instances
Match trigger tracks with all offline reconstructed tracks inside the jet RoI. The matching requirement between trigger and offline tracks is ΔR<0.1. Only offline tracks passing the following criteria are considered: |η|<2.5, |z0sinθ|<1.5mm (w.r.t. primary vertex), |d0|<1.5mm (w.r.t. primary vertex), number of b-layer hits >0, number of pixel hits > 1, number of total Si hits > 6, χ2 probability of track fit > 1%

pdf png
L2 and EF tracking efficiency vs matched offline track η for jet instances
Match trigger tracks with all offline reconstructed tracks inside the jet RoI. The matching requirement between trigger and offline tracks is ΔR<0.1. Only offline tracks passing the following criteria are considered: |η|<2.5, pT>2GeV, |z0sinθ|<1.5mm (w.r.t. primary vertex), |d0|<1.5mm (w.r.t. primary vertex), number of b-layer hits >0, number of pixel hits > 1, number of total Si hits > 6, χ2 probability of track fit > 1%

pdf png


Major updates:
-- JiriMasik - 24-Sept-2013 Tau plots from 2011 -- JiriMasik - 04-Jul-2011 -- JoergStelzer - 13-Jun-2011 Responsible: JiriMasik
Subject: public

Topic attachments
I Attachment History Action Size Date Who Comment
Unknown file formateps 2011K-ipt-vs-eta-mu10-pT10-v3.eps r1 manage 14.6 K 2012-06-29 - 14:53 JiriMasik  
PNGpng 2011K-ipt-vs-eta-mu10-pT10-v3.png r1 manage 12.5 K 2012-06-29 - 14:53 JiriMasik  
PDFpdf ATL-COM-DAQ-2013-064-plot1.pdf r1 manage 65.1 K 2013-09-23 - 14:09 JiriMasik  
PNGpng ATL-COM-DAQ-2013-064-plot1.png r1 manage 79.1 K 2013-09-23 - 14:07 JiriMasik  
PDFpdf ATL-COM-DAQ-2013-064-plot2.pdf r1 manage 40.8 K 2013-09-23 - 14:09 JiriMasik  
PNGpng ATL-COM-DAQ-2013-064-plot2.png r1 manage 80.6 K 2013-09-23 - 14:07 JiriMasik  
PDFpdf ATL-COM-DAQ-2013-064-plot3.pdf r1 manage 46.7 K 2013-09-23 - 14:09 JiriMasik  
PNGpng ATL-COM-DAQ-2013-064-plot3.png r1 manage 78.3 K 2013-09-23 - 14:09 JiriMasik  
PDFpdf ATL-COM-DAQ-2013-064-plot4.pdf r1 manage 47.4 K 2013-09-23 - 14:09 JiriMasik  
PNGpng ATL-COM-DAQ-2013-064-plot4.png r1 manage 84.7 K 2013-09-23 - 14:09 JiriMasik  
PDFpdf ATL-COM-DAQ-2014-088-plot1.pdf r1 manage 16.4 K 2014-08-23 - 15:08 MarkSutton  
PNGpng ATL-COM-DAQ-2014-088-plot1.png r1 manage 143.1 K 2014-08-23 - 15:09 MarkSutton  
PDFpdf ATL-COM-DAQ-2014-088-plot2.pdf r1 manage 14.9 K 2014-08-23 - 15:20 MarkSutton  
PNGpng ATL-COM-DAQ-2014-088-plot2.png r1 manage 103.0 K 2014-08-23 - 15:20 MarkSutton  
PDFpdf ATL-COM-DAQ-2014-088-plot3.pdf r1 manage 15.1 K 2014-08-23 - 15:20 MarkSutton  
PNGpng ATL-COM-DAQ-2014-088-plot3.png r1 manage 104.1 K 2014-08-23 - 15:09 MarkSutton  
PDFpdf ATL-COM-DAQ-2014-088-plot4.pdf r1 manage 15.4 K 2014-08-23 - 15:08 MarkSutton  
PNGpng ATL-COM-DAQ-2014-088-plot4.png r1 manage 113.4 K 2014-08-23 - 15:09 MarkSutton  
PDFpdf ATL-COM-DAQ-2014-088-plot5.pdf r1 manage 16.3 K 2014-08-23 - 15:08 MarkSutton  
PNGpng ATL-COM-DAQ-2014-088-plot5.png r1 manage 124.6 K 2014-08-23 - 15:09 MarkSutton  
PDFpdf ATL-COM-DAQ-2014-088-plot6.pdf r1 manage 16.0 K 2014-08-23 - 15:08 MarkSutton  
PNGpng ATL-COM-DAQ-2014-088-plot6.png r1 manage 90.9 K 2014-08-23 - 15:09 MarkSutton  
PDFpdf ATL-COM-DAQ-2014-088-plot7.pdf r1 manage 14.9 K 2014-08-23 - 15:08 MarkSutton  
PNGpng ATL-COM-DAQ-2014-088-plot7.png r1 manage 85.9 K 2014-08-23 - 15:09 MarkSutton  
PDFpdf ATL-COM-DAQ-2015-026-plot1.pdf r1 manage 57.0 K 2015-03-30 - 18:57 MarkSutton ID Trigger tau slice timing plots
PNGpng ATL-COM-DAQ-2015-026-plot1.png r1 manage 118.8 K 2015-03-30 - 18:57 MarkSutton ID Trigger tau slice timing plots
PDFpdf ATL-COM-DAQ-2015-026-plot2.pdf r1 manage 36.3 K 2015-03-30 - 18:57 MarkSutton ID Trigger tau slice timing plots
PNGpng ATL-COM-DAQ-2015-026-plot2.png r1 manage 135.9 K 2015-03-30 - 18:57 MarkSutton ID Trigger tau slice timing plots
PDFpdf ATL-COM-DAQ-2015-110-plot01.pdf r1 manage 14.9 K 2015-08-03 - 13:17 MarkSutton  
PNGpng ATL-COM-DAQ-2015-110-plot01.png r1 manage 17.1 K 2015-08-03 - 13:18 MarkSutton  
PDFpdf ATL-COM-DAQ-2015-110-plot02.pdf r1 manage 15.1 K 2015-08-03 - 13:17 MarkSutton  
PNGpng ATL-COM-DAQ-2015-110-plot02.png r1 manage 16.3 K 2015-08-03 - 13:18 MarkSutton  
PDFpdf ATL-COM-DAQ-2015-110-plot03.pdf r1 manage 15.5 K 2015-08-03 - 13:17 MarkSutton  
PNGpng ATL-COM-DAQ-2015-110-plot03.png r1 manage 17.6 K 2015-08-03 - 13:18 MarkSutton  
PDFpdf ATL-COM-DAQ-2015-110-plot04.pdf r1 manage 43.4 K 2015-08-03 - 13:17 MarkSutton  
PNGpng ATL-COM-DAQ-2015-110-plot04.png r1 manage 28.4 K 2015-08-03 - 13:18 MarkSutton  
PDFpdf ATL-COM-DAQ-2015-110-plot05.pdf r1 manage 14.8 K 2015-08-03 - 13:17 MarkSutton  
PNGpng ATL-COM-DAQ-2015-110-plot05.png r1 manage 20.9 K 2015-08-03 - 13:18 MarkSutton  
PDFpdf ATL-COM-DAQ-2015-110-plot06.pdf r1 manage 16.9 K 2015-08-03 - 13:17 MarkSutton  
PNGpng ATL-COM-DAQ-2015-110-plot06.png r1 manage 22.1 K 2015-08-03 - 13:18 MarkSutton  
PDFpdf ATL-COM-DAQ-2015-110-plot07.pdf r1 manage 22.3 K 2015-08-03 - 13:17 MarkSutton  
PNGpng ATL-COM-DAQ-2015-110-plot07.png r1 manage 35.8 K 2015-08-03 - 13:18 MarkSutton  
PDFpdf ATL-COM-DAQ-2015-110-plot08.pdf r1 manage 15.0 K 2015-08-03 - 13:17 MarkSutton  
PNGpng ATL-COM-DAQ-2015-110-plot08.png r1 manage 20.7 K 2015-08-03 - 13:18 MarkSutton  
PDFpdf ATL-COM-DAQ-2015-110-plot09.pdf r1 manage 16.2 K 2015-08-03 - 13:17 MarkSutton  
PNGpng ATL-COM-DAQ-2015-110-plot09.png r1 manage 21.2 K 2015-08-03 - 13:18 MarkSutton  
PDFpdf ATL-COM-DAQ-2015-110-plot10.pdf r1 manage 17.8 K 2015-08-03 - 13:17 MarkSutton  
PNGpng ATL-COM-DAQ-2015-110-plot10.png r1 manage 24.8 K 2015-08-03 - 13:18 MarkSutton  
PDFpdf ATL-COM-DAQ-2015-110-plot11.pdf r1 manage 16.7 K 2015-08-03 - 13:18 MarkSutton  
PNGpng ATL-COM-DAQ-2015-110-plot11.png r1 manage 25.5 K 2015-08-03 - 13:18 MarkSutton  
PDFpdf ATL-COM-DAQ-2015-110-plot12.pdf r1 manage 19.4 K 2015-08-03 - 13:18 MarkSutton  
PNGpng ATL-COM-DAQ-2015-110-plot12.png r1 manage 29.7 K 2015-08-03 - 13:18 MarkSutton  
PDFpdf ATL-COM-DAQ-2016-083-plot01.pdf r1 manage 177.4 K 2016-07-28 - 21:20 MarkSutton  
PNGpng ATL-COM-DAQ-2016-083-plot01.png r1 manage 107.3 K 2016-07-28 - 21:23 MarkSutton  
PDFpdf ATL-COM-DAQ-2016-083-plot02.pdf r1 manage 149.5 K 2016-07-28 - 21:20 MarkSutton  
PNGpng ATL-COM-DAQ-2016-083-plot02.png r1 manage 104.1 K 2016-07-28 - 21:23 MarkSutton  
PDFpdf ATL-COM-DAQ-2016-083-plot03.pdf r1 manage 177.5 K 2016-07-28 - 21:20 MarkSutton  
PNGpng ATL-COM-DAQ-2016-083-plot03.png r1 manage 99.6 K 2016-07-28 - 21:23 MarkSutton  
PDFpdf ATL-COM-DAQ-2016-083-plot04.pdf r1 manage 345.6 K 2016-07-28 - 21:20 MarkSutton  
PNGpng ATL-COM-DAQ-2016-083-plot04.png r1 manage 110.5 K 2016-07-28 - 21:23 MarkSutton  
PDFpdf ATL-COM-DAQ-2016-083-plot05.pdf r1 manage 515.9 K 2016-07-28 - 21:20 MarkSutton  
PNGpng ATL-COM-DAQ-2016-083-plot05.png r1 manage 126.7 K 2016-07-28 - 21:23 MarkSutton  
PDFpdf ATL-COM-DAQ-2016-083-plot06.pdf r1 manage 659.2 K 2016-07-28 - 21:20 MarkSutton  
PNGpng ATL-COM-DAQ-2016-083-plot06.png r1 manage 132.3 K 2016-07-28 - 21:23 MarkSutton  
PDFpdf ATL-COM-DAQ-2016-121.pdf r1 manage 15.6 K 2016-09-22 - 14:21 JiriMasik  
PNGpng ATL-COM-DAQ-2016-121.png r1 manage 92.4 K 2016-09-22 - 14:21 JiriMasik  
PDFpdf ATL-COM-DAQ-2017-107-plot00.pdf r1 manage 17.3 K 2017-09-22 - 21:34 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot00.png r1 manage 124.8 K 2017-09-22 - 21:34 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot01.pdf r1 manage 18.2 K 2017-09-22 - 21:34 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot01.png r1 manage 122.5 K 2017-09-22 - 21:34 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot02.pdf r1 manage 18.0 K 2017-09-22 - 21:34 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot02.png r1 manage 126.4 K 2017-09-22 - 21:34 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot03.pdf r1 manage 35.3 K 2017-09-22 - 21:34 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot03.png r1 manage 138.4 K 2017-09-22 - 21:34 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot04.pdf r1 manage 29.4 K 2017-09-22 - 21:35 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot04.png r1 manage 126.7 K 2017-09-22 - 21:35 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot05.pdf r1 manage 18.4 K 2017-09-22 - 21:35 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot05.png r1 manage 129.8 K 2017-09-22 - 21:35 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot06.pdf r1 manage 17.5 K 2017-09-22 - 21:35 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot06.png r1 manage 130.1 K 2017-09-22 - 21:35 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot07.pdf r1 manage 18.5 K 2017-09-22 - 21:35 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot07.png r1 manage 133.9 K 2017-09-22 - 21:35 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot08.pdf r1 manage 17.8 K 2017-09-22 - 21:35 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot08.png r1 manage 136.4 K 2017-09-22 - 21:36 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot09.pdf r1 manage 16.5 K 2017-09-22 - 21:36 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot09.png r1 manage 105.1 K 2017-09-22 - 21:36 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot10.pdf r1 manage 16.4 K 2017-09-22 - 21:36 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot10.png r1 manage 100.9 K 2017-09-22 - 21:36 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot11.pdf r1 manage 25.3 K 2017-09-22 - 21:36 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot11.png r1 manage 99.0 K 2017-09-22 - 21:36 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot12.pdf r1 manage 16.6 K 2017-09-22 - 21:36 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot12.png r1 manage 97.1 K 2017-09-22 - 21:37 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot13.pdf r1 manage 16.9 K 2017-09-22 - 21:37 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot13.png r1 manage 106.2 K 2017-09-22 - 21:37 MarkSutton  
PDFpdf ATL-COM-DAQ-2017-107-plot14.pdf r1 manage 16.7 K 2017-09-22 - 21:37 MarkSutton  
PNGpng ATL-COM-DAQ-2017-107-plot14.png r1 manage 101.0 K 2017-09-22 - 21:37 MarkSutton  
PDFpdf ATL-COM-DAQ-2018-061-fig1.pdf r1 manage 57.4 K 2018-06-15 - 10:58 MarkSutton 2018 performance plots for muons and taus
PNGpng ATL-COM-DAQ-2018-061-fig1.png r1 manage 100.2 K 2018-06-15 - 10:59 MarkSutton  
PDFpdf ATL-COM-DAQ-2018-061-fig2.pdf r1 manage 56.5 K 2018-06-15 - 10:58 MarkSutton 2018 performance plots for muons and taus
PNGpng ATL-COM-DAQ-2018-061-fig2.png r1 manage 104.1 K 2018-06-15 - 10:59 MarkSutton  
PDFpdf ATL-COM-DAQ-2018-061-fig3.pdf r1 manage 64.0 K 2018-06-15 - 10:58 MarkSutton 2018 performance plots for muons and taus
PNGpng ATL-COM-DAQ-2018-061-fig3.png r1 manage 100.8 K 2018-06-15 - 10:59 MarkSutton  
PDFpdf ATL-COM-DAQ-2018-061-fig4.pdf r1 manage 109.2 K 2018-06-15 - 10:58 MarkSutton 2018 performance plots for muons and taus
PNGpng ATL-COM-DAQ-2018-061-fig4.png r1 manage 102.4 K 2018-06-15 - 10:59 MarkSutton  
PDFpdf ATL-COM-DAQ-2018-061-fig5.pdf r1 manage 82.2 K 2018-06-15 - 10:58 MarkSutton 2018 performance plots for muons and taus
PNGpng ATL-COM-DAQ-2018-061-fig5.png r1 manage 113.1 K 2018-06-15 - 10:59 MarkSutton  
PDFpdf ATL-COM-DAQ-2018-061-fig6.pdf r1 manage 58.1 K 2018-06-15 - 10:58 MarkSutton 2018 performance plots for muons and taus
PNGpng ATL-COM-DAQ-2018-061-fig6.png r1 manage 113.9 K 2018-06-15 - 10:59 MarkSutton  
PDFpdf ATL-COM-DAQ-2018-061-fig7.pdf r1 manage 70.2 K 2018-06-15 - 10:58 MarkSutton 2018 performance plots for muons and taus
PNGpng ATL-COM-DAQ-2018-061-fig7.png r1 manage 121.1 K 2018-06-15 - 10:59 MarkSutton  
PDFpdf CPUIr-prelim.pdf r1 manage 175.2 K 2013-10-03 - 15:23 MarkSutton  
PNGpng CPUIr-prelim.png r1 manage 114.7 K 2013-10-03 - 15:23 MarkSutton  
PDFpdf EF_rd0_vs_eta_sigma.pdf r1 manage 57.6 K 2013-10-03 - 15:23 MarkSutton  
PNGpng EF_rd0_vs_eta_sigma.png r1 manage 77.2 K 2013-10-03 - 15:23 MarkSutton  
PDFpdf EF_rd0_vs_signed_pt_sigma.pdf r1 manage 42.8 K 2013-10-03 - 15:23 MarkSutton  
PNGpng EF_rd0_vs_signed_pt_sigma.png r1 manage 86.5 K 2013-10-03 - 15:23 MarkSutton  
PDFpdf EF_reta_vs_eta_sigma.pdf r1 manage 57.0 K 2013-10-03 - 15:23 MarkSutton  
PNGpng EF_reta_vs_eta_sigma.png r1 manage 71.9 K 2013-10-03 - 15:23 MarkSutton  
PDFpdf EF_reta_vs_pt_sigma.pdf r1 manage 57.4 K 2013-10-03 - 15:23 MarkSutton  
PNGpng EF_reta_vs_pt_sigma.png r1 manage 74.0 K 2013-10-03 - 15:23 MarkSutton  
PDFpdf EF_rzed_vs_eta_sigma.pdf r1 manage 57.4 K 2013-10-03 - 15:43 MarkSutton  
PNGpng EF_rzed_vs_eta_sigma.png r1 manage 69.1 K 2013-10-03 - 15:43 MarkSutton  
PDFpdf EF_rzed_vs_signed_pt_sigma.pdf r1 manage 56.8 K 2013-10-03 - 15:43 MarkSutton  
PNGpng EF_rzed_vs_signed_pt_sigma.png r1 manage 76.3 K 2013-10-03 - 15:43 MarkSutton  
Unknown file formateps Electrons-SuperImp-eff_vs_ntracks.eps r1 manage 10.8 K 2012-06-29 - 10:12 JiriMasik  
PNGpng Electrons-SuperImp-eff_vs_ntracks.png r1 manage 9.1 K 2012-06-29 - 10:12 JiriMasik  
Unknown file formateps Electrons-SuperImp-eta_eff.eps r1 manage 14.0 K 2012-06-29 - 10:12 JiriMasik  
PNGpng Electrons-SuperImp-eta_eff.png r1 manage 9.7 K 2012-06-29 - 10:12 JiriMasik  
PDFpdf GooDA-findZ-prelim.pdf r1 manage 101.3 K 2013-10-03 - 15:43 MarkSutton  
PNGpng GooDA-findZ-prelim.png r1 manage 105.6 K 2013-10-03 - 15:43 MarkSutton  
PDFpdf IDdataprepGPU.pdf r1 manage 28.3 K 2015-04-08 - 17:43 JohnTMBaines ATL-COM-DAQ-2015-036
PNGpng IDdataprepGPU.png r1 manage 219.9 K 2015-04-08 - 17:43 JohnTMBaines ATL-COM-DAQ-2015-036
PDFpdf L2StratA-rel.pdf r1 manage 97.4 K 2013-10-03 - 15:33 MarkSutton  
PNGpng L2StratA-rel.png r1 manage 363.5 K 2013-10-03 - 15:33 MarkSutton  
PDFpdf L2_rd0_vs_eta_sigma.pdf r1 manage 57.4 K 2013-10-03 - 15:45 MarkSutton  
PNGpng L2_rd0_vs_eta_sigma.png r1 manage 77.8 K 2013-10-03 - 15:45 MarkSutton  
PDFpdf L2_rd0_vs_signed_pt_sigma.pdf r1 manage 42.7 K 2013-10-03 - 15:45 MarkSutton  
PNGpng L2_rd0_vs_signed_pt_sigma.png r1 manage 86.3 K 2013-10-03 - 15:45 MarkSutton  
PDFpdf L2_reta_vs_eta_sigma.pdf r1 manage 56.9 K 2013-10-03 - 15:45 MarkSutton  
PNGpng L2_reta_vs_eta_sigma.png r1 manage 72.3 K 2013-10-03 - 15:45 MarkSutton  
PDFpdf L2_reta_vs_pt_sigma.pdf r1 manage 57.2 K 2013-10-03 - 15:45 MarkSutton  
PNGpng L2_reta_vs_pt_sigma.png r1 manage 74.6 K 2013-10-03 - 15:45 MarkSutton  
PDFpdf L2_rzed_vs_eta_sigma.pdf r1 manage 57.6 K 2013-10-03 - 15:49 MarkSutton  
PNGpng L2_rzed_vs_eta_sigma.png r1 manage 69.4 K 2013-10-03 - 15:49 MarkSutton  
PDFpdf L2_rzed_vs_signed_pt_sigma.pdf r1 manage 56.6 K 2013-10-03 - 15:49 MarkSutton  
PNGpng L2_rzed_vs_signed_pt_sigma.png r1 manage 76.3 K 2013-10-03 - 15:49 MarkSutton  
PDFpdf OptimisationTiming-prelim.pdf r1 manage 256.5 K 2013-10-03 - 15:44 MarkSutton  
PNGpng OptimisationTiming-prelim.png r1 manage 154.8 K 2013-10-03 - 15:44 MarkSutton  
PDFpdf VectorTiming-prelim.pdf r1 manage 261.0 K 2013-10-03 - 15:44 MarkSutton  
PNGpng VectorTiming-prelim.png r1 manage 202.9 K 2013-10-03 - 15:44 MarkSutton  
Unknown file formateps electrons-2011-PeriodLM.FinalDecorations.eff_vs_ntracks.eps r1 manage 10.2 K 2012-06-29 - 11:55 JiriMasik  
PNGpng electrons-2011-PeriodLM.FinalDecorations.eff_vs_ntracks.png r1 manage 8.7 K 2012-06-29 - 11:55 JiriMasik  
Unknown file formateps electrons-2011-PeriodLM.FinalDecorations.eff_vs_nvtx.eps r1 manage 11.3 K 2012-06-29 - 11:55 JiriMasik  
PNGpng electrons-2011-PeriodLM.FinalDecorations.eff_vs_nvtx.png r1 manage 9.0 K 2012-06-29 - 11:55 JiriMasik  
Unknown file formateps electrons-2011-PeriodLM.FinalDecorations.eta_eff.eps r1 manage 12.7 K 2012-06-29 - 11:55 JiriMasik  
PNGpng electrons-2011-PeriodLM.FinalDecorations.eta_eff.png r1 manage 8.7 K 2012-06-29 - 11:55 JiriMasik  
Unknown file formateps electrons-2011-PeriodLM.FinalDecorations.pT_eff.eps r1 manage 13.6 K 2012-06-29 - 11:55 JiriMasik  
PNGpng electrons-2011-PeriodLM.FinalDecorations.pT_eff.png r1 manage 9.4 K 2012-06-29 - 11:55 JiriMasik  
Unknown file formateps electrons-2012-PeriodA4.FinalDecorations.FinalLabels.NoGrid.eff_vs_ntracks.eps r1 manage 9.4 K 2012-06-29 - 07:49 JiriMasik  
PNGpng electrons-2012-PeriodA4.FinalDecorations.FinalLabels.NoGrid.eff_vs_ntracks.png r1 manage 8.5 K 2012-06-29 - 07:51 JiriMasik  
Unknown file formateps electrons-2012-PeriodA4.FinalDecorations.FinalLabels.NoGrid.eta_eff.eps r1 manage 11.4 K 2012-06-29 - 07:52 JiriMasik  
PNGpng electrons-2012-PeriodA4.FinalDecorations.FinalLabels.NoGrid.eta_eff.png r1 manage 8.3 K 2012-06-29 - 07:52 JiriMasik  
Unknown file formateps electrons-2012-PeriodA4.FinalDecorations.FinalLabels.NoGrid.pT_eff.eps r1 manage 10.7 K 2012-06-29 - 07:53 JiriMasik  
PNGpng electrons-2012-PeriodA4.FinalDecorations.FinalLabels.NoGrid.pT_eff.png r1 manage 8.7 K 2012-06-29 - 07:53 JiriMasik  
Unknown file formateps electrons-2012-PeriodA4.FinalDecorations.FinalLabels.eff_vs_nvtx.eps r1 manage 11.4 K 2012-06-29 - 07:47 JiriMasik  
PNGpng electrons-2012-PeriodA4.FinalDecorations.FinalLabels.eff_vs_nvtx.png r1 manage 9.7 K 2012-06-29 - 07:49 JiriMasik  
Unknown file formateps muons-2011-PeriodLM.FinalDecorations.eff_vs_ntracks.eps r1 manage 10.1 K 2012-06-29 - 11:55 JiriMasik  
PNGpng muons-2011-PeriodLM.FinalDecorations.eff_vs_ntracks.png r1 manage 8.6 K 2012-06-29 - 11:55 JiriMasik  
Unknown file formateps muons-2011-PeriodLM.FinalDecorations.eff_vs_nvtx.eps r1 manage 11.2 K 2012-06-29 - 11:57 JiriMasik  
PNGpng muons-2011-PeriodLM.FinalDecorations.eff_vs_nvtx.png r1 manage 8.8 K 2012-06-29 - 11:57 JiriMasik  
Unknown file formateps muons-2011-PeriodLM.FinalDecorations.eta_eff.eps r1 manage 12.1 K 2012-06-29 - 11:57 JiriMasik  
PNGpng muons-2011-PeriodLM.FinalDecorations.eta_eff.png r1 manage 8.5 K 2012-06-29 - 11:57 JiriMasik  
Unknown file formateps muons-2011-PeriodLM.FinalDecorations.pT_eff.eps r1 manage 10.7 K 2012-06-29 - 11:57 JiriMasik  
PNGpng muons-2011-PeriodLM.FinalDecorations.pT_eff.png r1 manage 8.7 K 2012-06-29 - 11:57 JiriMasik  
Unknown file formateps periodK-blayer-vs-eta-pt6-v3.eps r1 manage 9.2 K 2012-06-29 - 14:54 JiriMasik  
PNGpng periodK-blayer-vs-eta-pt6-v3.png r1 manage 8.5 K 2012-06-29 - 14:54 JiriMasik  
Unknown file formateps periodK-offline-el-vs-pT-pt10-v3.eps r1 manage 10.0 K 2012-06-29 - 14:53 JiriMasik  
PNGpng periodK-offline-el-vs-pT-pt10-v3.png r1 manage 10.8 K 2012-06-29 - 14:53 JiriMasik  
Unknown file formateps periodK-staco-trk-vs-nvtx-mu10-v3.eps r1 manage 9.9 K 2012-06-29 - 14:53 JiriMasik  
PNGpng periodK-staco-trk-vs-nvtx-mu10-v3.png r1 manage 9.3 K 2012-06-29 - 14:53 JiriMasik  
PNGpng periodK-staco-trk-vs-pT-mu10-pt6-v3.png r1 manage 10.7 K 2012-06-29 - 14:53 JiriMasik  
Unknown file formateps periodL-tau29-trk-vs-eta.eps r1 manage 11.4 K 2013-09-24 - 16:09 JiriMasik  
PNGpng periodL-tau29-trk-vs-eta.png r1 manage 13.3 K 2013-09-24 - 16:09 JiriMasik  
Unknown file formateps periodL-tau29-trk-vs-nvtx.eps r1 manage 14.2 K 2013-09-24 - 16:09 JiriMasik  
PNGpng periodL-tau29-trk-vs-nvtx.png r1 manage 15.0 K 2013-09-24 - 16:09 JiriMasik  
Unknown file formateps periodL-tau29-trk-vs-pT.eps r1 manage 12.9 K 2013-09-24 - 16:09 JiriMasik  
PNGpng periodL-tau29-trk-vs-pT.png r1 manage 12.9 K 2013-09-24 - 16:09 JiriMasik  
PDFpdf plan.pdf r1 manage 446.8 K 2013-10-03 - 15:33 MarkSutton  
PNGpng plan.png r1 manage 201.5 K 2013-10-03 - 15:33 MarkSutton  
Unknown file formateps track_seeding_time_plot.eps r1 manage 9.8 K 2015-10-29 - 14:52 AdemarDelgado  
PDFpdf track_seeding_time_plot.pdf r1 manage 15.7 K 2015-10-29 - 14:52 AdemarDelgado  
PNGpng track_seeding_time_plot.png r2 r1 manage 10.4 K 2015-10-29 - 15:03 AdemarDelgado  
Edit | Attach | Watch | Print version | History: r28 < r27 < r26 < r25 < r24 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r28 - 2018-06-15 - MarkSutton
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Atlas All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright &© 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback