ESAF Euso Simulation and Analysis Framework Simulation, Reconstruction, Visulation, Analysis Simulates process from shower generation to light production, transport, detector response, reconstruction and analysis A package of C++, C and F77 programs organized as shared object Libraries, based on ROOT using its libraries with high modularity Air showers: CORSIKA, CONEX, built-in Slast shoower generator

Fake trigger backgoround simulations

The triggering philosophy The JEM-EUSO trigger philosophy is at the core of the concept of the instrument. The goal of the trigger system is to detect the occurrence of a scientifically valuable signal among the background noise detected by the JEM-EUSO telescope. Since the total number of pixels in the array is very large (~ 2×105), a multi-level trigger scheme was developed. This trigger scheme relies on the partitioning of the Focal Surface in subsections, named PDM (Photo Detector Module), which are large enough to contain a substantial part of the imaged track under investigation (this depends on the energy of air shower and the zenith angle).

JEM-EUSO FOCAL SURFACE LAYOUT Typical air-shower seen by JEM-EUSO JEM-EUSO

Outline of noise reduction capability. The general JEM-EUSO trigger philosophy asks for a System Trigger organized into two main trigger-levels. The two levels of trigger work on the statistical properties of the incoming photon flux in order to detect the physical events hindered in the background, basing on their position and time correlation. The trigger is issued in accordance with two different stages: Outline of noise reduction capability. Level Rate of signals/triggers at PDM level Rate of signals/triggers at FS level 1st level trigger (PDM) Photon trigger ~9.2 × 108 Hz ~1.4 × 1011 Hz Counting trigger ~7.1 × 105 Hz ~1.1 × 108 Hz Persistency trigger ~7 Hz ~103 Hz 2nd level trigger (PDM cluster) ~6.7 × 10-4 Hz ~0.1 Hz Expected rate of cosmic ray events ~6.7 × 10-6 Hz ~10-3 Hz

The Table gives a synthetic idea of the expected rate of signals at each stage, and the expected rejection power. The numbers here reported give a first rough estimation of the requirements. The exact power rejection of each trigger level will be optimized in future. The experience with the balloon measurement would provide us very useful information for tuning these parameters. The last row gives also a reference number on the expected rate of cosmic ray events, which could fluctuate by around one order of magnitude depending on the effective threshold of the detector. The previous table shows how important is the capability to cope with the nightglow background to reduce the data rate. The most critical level is the 1st one where the power reduction is of 8 orders of magnitude.

JEM-EUSO DAQ – Data reduction block scheme 4*10-3 compression 9.6 GB/s (FS) 10-3 compression 297 kbps 3 Gbyte/day Storage on SSD will give factor 3, up to 10 Gbyte/day Return with Soyuz FS Control Board MPU Operation Control PDM Control Board FPGA PTT Trigger Cluster Control Board FPGA LTT Trigger 20CCB IDAQ FEE ASIC+ FPGA Count 9EC 8PDM PhotoDetector Modules 200kch 1,476 EC 160 Boards 20 Boards 2 Boards LVDS with SpaceWire (ECSS-E-50-12A) The trigger system of the JEM-EUSO Telescope, BERTAINA (Torino), CATALANO (Palermo) The data acquisition and handling system of the Jem-Euso experiment, M. CASOLINO (Roma2)

1st level PDM trigger 2nd level CCB trigger

The AFTL is based on the following assumptions: FIRST + SECOND LEVEL TRIGGER CONCEPT The AFTL is based on the following assumptions: PIXELS ABOVE <BACKGROUND> . For each Elementary Cell (EC) pixels, digitalized anode pulses (pe) are counted within a GTU(2.5 µs) and compared with a pre-set digital threshold N. At every GTU the counters C1, one for each pixel, are reset. For each C1, if the counts are greater than the pre-set threshold , the successive pulses are conveyed to a second counter C2, one for each pixel, and a signal L, one for each pixel, flags the pixel as active. All the L signals are OR-ed. ELEMENTARY CELL ACTIVITY CHECK . A counter C3 (persistency counter), only one per EC, is increased at each GTU if the output signal O of the OR-ed L signals is active else it is reset. SPACE-TIME CORRELATION OF PIXELS ABOVE THRESHOLD . The C3 counts are compared with a pre-set digital threshold P. If the C3 counts reach the P threshold a signal is issued to the adder A that holds the C2 counters 2x2 (or 3x3) grouped. The resulting addition is then compared with the pre-set value S corresponding to the total number of pe requested in the 2x2 (or 3x3) grouped pixels. If the condition is met, an EC trigger is then generated. Obviously read-out of data is based on “free running method”: pixels counts recorded on memories of suitable depth are reading out at the occurrence of a trigger.

Motivation Very high statistics of simulated background needed → 105 events →1014 GTU's Impossible to simulate by ESAF: → 103 slower then used code → cannot be computed parallely (mem. share) Fast and standalone code written in C++ developed by Francseco Fenu

Contribution/Activity: Computing facility – JEM-EUSO cluster 16 node Supermicro® SuperServer AS-1042G-MTF Configuration of node : - 4x Opteron 6134 (2,3GHz) - 16GB RAM - 600GB SATAII HDD (WD VelociRaptor) 2x master/disk server Supermicro® SuperServer AS- 1042G-MTF Configuration od server: - 1x Opteron 6134 (2,3GHz) - 4x 2TB SATAII HDD (WD RE4) All together: CPU: 64 + 2 @ 2.3GHz Cores: 512 + 16 RAM: 264 GB (4GB / CPU) Disk space: 16x600GB + 8x2TB = 25,2TB

JEM-EUSO Kosice cluster Actually used and available for also for collaboration 7*32 cores @ 2.3 Ghz; 25 TB upgrade in progress right now Fedora Core 14 1.2.5-2.fc14 kernel 2.6.35.13- 91.fc14.x86_64 gcc 4.5.1-4 NFS shared disk space (temporarily), RAID configuration in progress right now ROOT v32.00, ESAF trunk (less than 1 month), GEANT4 9.4

M64 config

Pattern Recognition Disentangle signal from background and noise Avoid fake reconstruction of arbitrarily distributed parasitic photons Analyse count distribution in space and time for causal relation Estimate the geometrical properties of the signal track  2 algorithms implemented for this task

Pattern Recognition Hough transformation or clustering to disentangle signal from background… …and to determine the geometrical parameters of the track. .