98Zr PSA benchmark
AGATA Benchmarks#
Different docker images are provided as a way to build and/or use the different softwares on any system.
PSA Benchmarks#
Doppler correction optimization on 98Zr (E680 data)#
Concept#
A dataset has been extracted from the E680 fission data to contain raw traces of the 98Zr nucleus, having the good idea to have an isolated and strong \(\gamma\)-ray transition at high energy (1.2 MeV).
In this experiments, the fission fragments are emitted with velocities of \(\beta \sim 10%\), making the resolution of such transition strongly dependant on the quality of the Doppler correction, and consequently on the PSA precision.
After the full analysis of the data, a timestamp list has been extracted corresponding to each 98Zr event. This timestamp list has been used for reprocessing the raw data from the traces level using the TSFilter actor of the AGAPRO package. As a result, we obtain as output ADF files at the producer level, but containing only the traces for 98Zr events.
An other file has been produced containing for each of these timestamps, the required information for applying the Doppler correction in post analysis (namely velovity and emission angle of the recoil).
The concept of this dataset is to reprocess these data, for any new PSA development. The Doppler corrected spectrum obtained will be used to conclude on the increasing or not of the PSA quality (namely looking at the 1.2 MeV peak resolution)
Prerequisite#
To perform the PSA replay of this dataset, it is mandatory to have the AGAPRO package (including femul) compiled with ROOT (see the AGAPRO page)
Recovering the data with iRODS#
Configure iRODS:
You then need to configure iRODS for the AGATA server:
Create the file: ~/.irods/irods_environment.json with the following content:
{
"irods_cwd": "/ccin2p3/home/agata/D4B",
"irods_home": "/ccin2p3/home/agata/D4B",
"irods_host": "ccirods.in2p3.fr",
"irods_port": 5570,
"irods_user_name": "benchmark",
"irods_zone_name": "ccin2p3"
}
To install iRODS icommands:
The data are stored on the AGATA iRODS server (working on the 4.x iRODS version).
docker pull gitlab-registry.in2p3.fr/ip2igamma/docker_images:irods-icommands
docker run -it --rm -u $(id -u):$(id -g) \
-v ${HOME}/.irods:/home/.irods \
gitlab-registry.in2p3.fr/ip2igamma/docker_images:irods-icommands \
bash
- Using apptainer/singulariry:
First build the apptainer image from the docker container
apptainer build irods-icommands.sif docker://gitlab-registry.in2p3.fr/ip2igamma/docker_images:irods-icommands
Then launch the apptainer image:
To initialize iRODS:
Please contact O. Stezowski or J. Dudouet for the password
Locate the data:
user:$ icd /ccin2p3/home/agata/D4B/98Zr_DC_Benchmark
user:$ ils
/ccin2p3/home/agata/D4B/98Zr_DC_Benchmark:
FullReplay.sh
C- /ccin2p3/home/agata/D4B/98Zr_DC_Benchmark/ADL
C- /ccin2p3/home/agata/D4B/98Zr_DC_Benchmark/benchmark
C- /ccin2p3/home/agata/D4B/98Zr_DC_Benchmark/runs_0042_to_0072
C- /ccin2p3/home/agata/D4B/98Zr_DC_Benchmark/runs_0073_to_0092
C- /ccin2p3/home/agata/D4B/98Zr_DC_Benchmark/runs_0093_to_0118
C- /ccin2p3/home/agata/D4B/98Zr_DC_Benchmark/runs_0119_to_0188
Go in your analysis directory and download the data (~190Go):
The download can be quite long...
PSA processing#
The repository where the data have been downloaded should be like
user:$ ls -l
drwxrwxr-x. 2 dudouet matnuc 32768 Aug 7 14:00 ADL
-rwxrw-r--. 1 dudouet matnuc 875 Aug 11 14:25 FullReplay.sh
drwxrwxr-x. 2 dudouet matnuc 32768 Aug 11 14:24 benchmark
drwxrwxr-x. 5 dudouet matnuc 32768 Aug 11 09:26 runs_0042_to_0072
drwxrwxr-x. 5 dudouet matnuc 512 Aug 11 09:26 runs_0073_to_0092
drwxrwxr-x. 5 dudouet matnuc 512 Aug 11 09:26 runs_0093_to_0118
drwxrwxr-x. 5 dudouet matnuc 512 Aug 11 09:26 runs_0119_to_0188
During this experiment, we had three reboots of the GTS, used for delivering the timestamps to each event. Due to these reboot, the analysis needs to be separeted in blocks of data for which no reboot occurs, in ordre to not mix the timestamps and the Doppler correction information. This is why the data are separated in four folders.
The ADL folder contains the ADL basis that are mandatory to perform the PSA on this dataset.
Launching the PSA:
To start the full replay, just launch the FullReplay.sh. It will launch in 4 separate threads the 4 different datasets.
If you want to reprocess only oen dataset, process as follow:
user:$ cd runs_0042_to_0072
user:$ python3 gen_conf.py # Mandatory to uptade the good paths for data reading and writing
user:$ femul Topology_PreproToTracking.conf |tee runs_0042_to_0072_replay.log
After the processing, in each subset of runs, a ROOT file has been created containing the tracking output in:
user:$ ls -l runs_0042_to_0072/Out/Analysis
-rw-r--r-- 1 dudouet staff 453605833 11 aoû 15:31 Tree_0000.root
These ROOT Trees will finally be read and analysed to produce the Doppler corrected \(\gamma\)-ray spectrum.
Data benchmark#
Once the replay has been processed, the folder benchmark contains the program to produce the oppler corrected \(\gamma\)-ray spectrum.
First, compile this program:
And execute it (you should have the following output):
user:$ ./benchmark
Analysis progress thread 4: 13.00 % (710684/5466897)
Analysis progress thread 3: 100.00 % (494210/494211)
Analysis progress thread 2: 54.00 % (706482/1308328)
Analysis progress thread 1: 42.00 % (666960/1588098)
This program will read the new tracking results, and using timestamp matching with the Doppler correction files, will produce the Doppler corrected spectrum.
The spectrum is stored in the file Benchmark.root, in the histogram named: trackEDC_0. It should be like on the following picture:
The file benchmark.cpp can be modified to change the input TTrees to be analysed for the Doppler coorection.
The spectrum obtained using the default PSA configuration is stored in the file named Benchmark_Reference.root