Periodic Reporting for period 4 - InPairs (In Silico Pair Plasmas: from ultra intense lasers to relativistic astrophysics in the laboratory)
Berichtszeitraum: 2021-03-01 bis 2022-08-31
InPairs studies the properties of plasmas composed of matter and anti-matter, electron-positron pair plasmas, under extreme intensities, both in laboratory and astrophysical conditions. These conditions are relevant at the focus of most intense lasers and particle beams now being deployed or in compact astrophysical objects such as black holes and neutron stars. We determined how the collective plasma effects coupled with QED processes impact the dynamics in these systems, and their experimental and astronomical signatures.
The project took advantage of the unique capabilities for kinetic simulation of plasmas in extreme conditions developed in InPairs, as well as from the theoretical and computational breakthroughs the team achieved on QED PIC simulations, radiation reaction in colliding laser-beam and beam-beam configurations, and the theoretical developments on the plasma instabilities of electron-positron plasmas, including magnetic reconnection. We uncovered the physics and the signatures of QED in plasmas under intense fields. We have also developed the numerical tools and algorithms capable of taking advantage of the next generation of exascale supercomputers to de deployed in Europe.
The project addressed several long-standing scientific questions: how pulsars radiate coherently in the radio regime; what is the behavior of plasmas in the presence of ultra intense magnetic fields; what experiments with lasers and particles can be performed in the laboratory to mimic these scenarios, addressing them with some of the largest and more sophisticated particle-in-cell simulations ever performed.
The project took advantage of the unique capabilities for kinetic simulation of plasmas in extreme conditions developed in InPairs, as well as from the theoretical and computational breakthroughs the team achieved on QED PIC simulations, radiation reaction in colliding laser-beam and beam-beam configurations, and the theoretical developments on the plasma instabilities of electron-positron plasmas, including magnetic reconnection. We uncovered the physics and the signatures of QED in plasmas under intense fields. We have also developed the numerical tools and algorithms capable of taking advantage of the next generation of exascale supercomputers to de deployed in Europe.
The project addressed several long-standing scientific questions: how pulsars radiate coherently in the radio regime; what is the behavior of plasmas in the presence of ultra intense magnetic fields; what experiments with lasers and particles can be performed in the laboratory to mimic these scenarios, addressing them with some of the largest and more sophisticated particle-in-cell simulations ever performed.
We have made significant progresses on the scalability of our main numerical simulation tool, Osiris demonstrating superb scalability. Within this challenge we have successfully improved OSIRIS for deployment on petascale and pre-exascale systems:
(1) Deployment of Xeon Phi cluster. A 4 node Xeon Phi cluster has been deployed in our group with the use of additional funding procured for supporting the network infrastructure. This cluster has been actively used for code development, small simulation work, and simulation data analysis;
(2) OSIRIS improvements for current Petascale HPC systems. The OSIRIS code has been further developed to fully support the Intel Knights Landing architecture, x86 AVX-512 capable processors, and CUDA enabled GPU systems. We currently can deploy the code with full hardware support on 8 of the top 10 systems in the world (www.top500.org - November 2018 list);
(3) Advanced diagnostics for large scale simulations. The simulation infrastructure has been upgraded to handle very large simulations, specifically through the development of a new file format (ZDF) and parallel I/O strategies that have significantly improved performance at these levels, and the upgrade of diagnostic routines that perform data reduction while the code is running;
(4) New algorithms and physics models. The OSIRIS code has been extended with improvements to the Quasi-3D algorithm, that now supports ionization modeling and soon QED. We also implemented a ponderomotive guiding center algorithm, which is another reduced model, and that enables us to model very large scenarios at a fraction of the computational cost, making it ideal for parameter scans. The code has also been extended for handling the extreme intensities that are now becoming available on top-tier laser systems, in particular through the continued development of a QED module including the most relevant processes;
(5) Synthetic instruments for radiation modeling. The data post processing infrastructure has been updated to include characteristics of laboratory measuring devices in the data analysis process. Particular focus was given to the simulation / experiment of (virtual) detectors and the analysis of the short wavelength radiation produced. Regarding the latter, we developed a novel method that can be used to obtain the detailed time evolution of this radiation, in a new tool called RaDIO.
(1) Deployment of Xeon Phi cluster. A 4 node Xeon Phi cluster has been deployed in our group with the use of additional funding procured for supporting the network infrastructure. This cluster has been actively used for code development, small simulation work, and simulation data analysis;
(2) OSIRIS improvements for current Petascale HPC systems. The OSIRIS code has been further developed to fully support the Intel Knights Landing architecture, x86 AVX-512 capable processors, and CUDA enabled GPU systems. We currently can deploy the code with full hardware support on 8 of the top 10 systems in the world (www.top500.org - November 2018 list);
(3) Advanced diagnostics for large scale simulations. The simulation infrastructure has been upgraded to handle very large simulations, specifically through the development of a new file format (ZDF) and parallel I/O strategies that have significantly improved performance at these levels, and the upgrade of diagnostic routines that perform data reduction while the code is running;
(4) New algorithms and physics models. The OSIRIS code has been extended with improvements to the Quasi-3D algorithm, that now supports ionization modeling and soon QED. We also implemented a ponderomotive guiding center algorithm, which is another reduced model, and that enables us to model very large scenarios at a fraction of the computational cost, making it ideal for parameter scans. The code has also been extended for handling the extreme intensities that are now becoming available on top-tier laser systems, in particular through the continued development of a QED module including the most relevant processes;
(5) Synthetic instruments for radiation modeling. The data post processing infrastructure has been updated to include characteristics of laboratory measuring devices in the data analysis process. Particular focus was given to the simulation / experiment of (virtual) detectors and the analysis of the short wavelength radiation produced. Regarding the latter, we developed a novel method that can be used to obtain the detailed time evolution of this radiation, in a new tool called RaDIO.
The InPairs project focuses on using self-consistent ab initio massively parallel simulations to study electron-positron pair plasmas under extreme intensities, both in laboratory and astrophysical conditions.
We first focused on acquisition and setup of an hybrid CPU/GPU machine, has the first step for the development and optimization, testing and profiling of the Osiris code at node-level parallelization. This was unique and well beyond the state-of-the-art in particle-in-cell codes and further established OSIRIS as the state-of-the-art particle-in-cell code. We have setup a new Virtual Reality system to explore novel ways for visualizing the resulting simulations with already very exciting results that is already unique.
We have developed a novel radiation diagnostic that has been incorporated into the main simulation tool, Osiris. This simulation tool is new, and is a development that goes clearly beyond the state-of-the-art because it captures the spatiotemporal profile of the radiation, because it can capture the radiation from nearly all the particles in the simulation, and because it runs at run time with the Osiris simulation.
Progress has also been made on exotic laser beams. So far most of the developments regarding the interaction of twisted light with the plasma still did not produce a fundamental difference on the structures produced in the plasma. However, our work was the first to show that the orbital angular momentum of light could be used to access previously unconsidered degrees of freedom on ultra-intense laser-matter interactions in particular on particle acceleration.
The dynamics of ultra relativistic (e-,e+) fireball beam with plasmas has also been explored in detail, by performing two-dimensional particle-in-cell simulations, and for the first time the competition between different plasma instabilities has been identified. We also have shown that the longitudinal energy spread, typical of plasma-based accelerated electron–positron fireball beams, play a minor role in the growth of kinetic plasma instabilities in these scenarios. This has been of paramount importance for a new experiment recently launched at CERN. We worked on a possible novel experimental setups to investigate the generation of Weibel magnetic fields in laser-driven plasmas. The task was complex because the small-scale Weibel fields are usually concealed by the large-scale Biermann magnetic field, which develop as result of the target heating and expansion. Performing ab initio first-principles simulations, we were able to demonstrate that with an opportune tuning of the laser parameters, it is possible to disentangle the two mechanisms and clearly observe the Weibel instability at work. Results of these works have triggered several experiment programs worldwide that now aim to observe these two regimes.
We first focused on acquisition and setup of an hybrid CPU/GPU machine, has the first step for the development and optimization, testing and profiling of the Osiris code at node-level parallelization. This was unique and well beyond the state-of-the-art in particle-in-cell codes and further established OSIRIS as the state-of-the-art particle-in-cell code. We have setup a new Virtual Reality system to explore novel ways for visualizing the resulting simulations with already very exciting results that is already unique.
We have developed a novel radiation diagnostic that has been incorporated into the main simulation tool, Osiris. This simulation tool is new, and is a development that goes clearly beyond the state-of-the-art because it captures the spatiotemporal profile of the radiation, because it can capture the radiation from nearly all the particles in the simulation, and because it runs at run time with the Osiris simulation.
Progress has also been made on exotic laser beams. So far most of the developments regarding the interaction of twisted light with the plasma still did not produce a fundamental difference on the structures produced in the plasma. However, our work was the first to show that the orbital angular momentum of light could be used to access previously unconsidered degrees of freedom on ultra-intense laser-matter interactions in particular on particle acceleration.
The dynamics of ultra relativistic (e-,e+) fireball beam with plasmas has also been explored in detail, by performing two-dimensional particle-in-cell simulations, and for the first time the competition between different plasma instabilities has been identified. We also have shown that the longitudinal energy spread, typical of plasma-based accelerated electron–positron fireball beams, play a minor role in the growth of kinetic plasma instabilities in these scenarios. This has been of paramount importance for a new experiment recently launched at CERN. We worked on a possible novel experimental setups to investigate the generation of Weibel magnetic fields in laser-driven plasmas. The task was complex because the small-scale Weibel fields are usually concealed by the large-scale Biermann magnetic field, which develop as result of the target heating and expansion. Performing ab initio first-principles simulations, we were able to demonstrate that with an opportune tuning of the laser parameters, it is possible to disentangle the two mechanisms and clearly observe the Weibel instability at work. Results of these works have triggered several experiment programs worldwide that now aim to observe these two regimes.