Papers by Stefano Sinigardi
3D electromagnetic Particle In Cell parallel code
Features: fusion blocks: FPN, PAN, ASFF, BiFPN network modules: ResNet, CPS, SPP, RFB network arc... more Features: fusion blocks: FPN, PAN, ASFF, BiFPN network modules: ResNet, CPS, SPP, RFB network architecture search: CSPResNext50, CSPDarknet53, SpineNet49, EfficientNetB0, MixNet-M activations: SWISH, MISH other features: weighted-[shortcut], Sigmoid scaling (Scale-sensitivity), Label smoothing, Optimal hyper parameters, Dynamic mini batch size for random shapes, Squeeze-and-excitation, Grouped convolution, MixConv (grouped [route]), Elastic-module data augmentation: MixUp, CutMix, Mosaic losses: MSE, GIoU, CIoU, DIoU detection layers: [yolo] (fixed iou_thresh), [Gaussian_yolo] detection on video (sequence of frames) - layers: [crnn] (convolutional-RNN), [conv_lstm] (Convolutional LSTM)
A high power laser named FLAME with an intensity up to 1021 W/cm2, a repetition rate of 10 Hz and... more A high power laser named FLAME with an intensity up to 1021 W/cm2, a repetition rate of 10 Hz and a contrast value (between main pulse and pre-pulse) of 1010 is being deployed at the LNF – INFN in Frascati and it is expected to be fully operative by the middle of 2012. In this frame an experiment of light ions ac-celeration through laser interaction with thin metal targets (LILIA) has been proposed and funded. The aim of LILIA experiment is to study, design and ver-ify a scheme which foresees the production, the char-acterization and the transport of a proton beam to-ward a stage of post acceleration (high frequency compact Linac). Now the maximum operating laser intensity is limited to 1019 W/cm2 due to the lack of a parabola with a focal length shorter that the current used. In this configuration, according to the interac-tion theory by short pulse laser and to performed nu-merical simulations, we expect a proton beam with maximum energy of a few MeV with a total dose up to 1010-1...
Laser Acceleration of Electrons, Protons, and Ions II; and Medical Applications of Laser-Generated Beams of Particles II; and Harnessing Relativistic Plasma Waves III, 2013
ABSTRACT We present a start-to-end 3D numerical simulation of a hybrid scheme for the acceleratio... more ABSTRACT We present a start-to-end 3D numerical simulation of a hybrid scheme for the acceleration of protons. The scheme is based on a first stage laser acceleration, followed by a transport line with a solenoid or a multiplet of quadrupoles, and then a post-acceleration section in a compact linac. Our simulations show that from a laser accelerated proton bunch with energy selection at ~ 30MeV, it is possible to obtain a high quality monochromatic beam of 60MeV with intensity at the threshold of interest for medical use. In the present day experiments using solid targets, the TNSA mechanism describes accelerated bunches with an exponential energy spectrum up to a cut-off value typically below ~ 60MeV and wide angular distribution. At the cut-off energy, the number of protons to be collimated and post-accelerated in a hybrid scheme are still too low. We investigate laser-plasma acceleration to improve the quality and number of the injected protons at ~ 30MeV in order to assure efficient post-acceleration in the hybrid scheme. The results are obtained with 3D PIC simulations using a code where optical acceleration with over-dense targets, transport and post-acceleration in a linac can all be investigated in an integrated framework. The high intensity experiments at Nara are taken as a reference benchmarks for our virtual laboratory. If experimentally confirmed, a hybrid scheme could be the core of a medium sized infrastructure for medical research, capable of producing protons for therapy and x-rays for diagnosis, which complements the development of all optical systems.
We present 'jasmine', an implementation of a fully relativistic, 3D, electromagnetic Particle-In-... more We present 'jasmine', an implementation of a fully relativistic, 3D, electromagnetic Particle-In-Cell (PIC) code, capable of running simulations in various laser plasma acceleration regimes on Graphics-Processing-Units (GPUs) HPC clusters. Standard energy/charge preserving FDTD-based algorithms have been implemented using double precision and quadratic (or arbitrary sized) shape functions for the particle weighting. When porting a PIC scheme to the GPU architecture (or, in general, a shared memory environment), the particle-to-grid operations (e.g. the evaluation of the current density) require special care to avoid memory inconsistencies and conflicts. Here we present a robust implementation of this operation that is efficient for any number of particles per cell and particle shape function order. Our algorithm exploits the exposed GPU memory hierarchy and avoids the use of atomic operations, which can hurt performance especially when many particles lay on the same cell. We show the code multi-GPU scalability results and present a dynamic load-balancing algorithm. The code is written using a python-based C++ meta-programming technique which translates in a high level of modularity and allows for easy performance tuning and simple extension of the core algorithms to various simulation schemes.
Physical Review Special Topics - Accelerators and Beams, 2013
Laser accelerated proton beams have a considerable potential for various applications including o... more Laser accelerated proton beams have a considerable potential for various applications including oncological therapy. However, the most consolidated target normal sheath acceleration regime based on irradiation of solid targets provides an exponential energy spectrum with a significant divergence. The low count number at the cutoff energy seriously limits at present its possible use. One realistic scenario for the near future is offered by hybrid schemes. The use of transport lines for collimation and energy selection has been considered. We present here a scheme based on a high field pulsed solenoid and collimators which allows one to select a beam suitable for injection at 30 MeV into a compact linac in order to double its energy while preserving a significant intensity. The results are based on a fully 3D simulation starting from laser acceleration.
The LILIA experiment is planned at the SPARCLAB facility of the Frascati INFN laboratories. We ha... more The LILIA experiment is planned at the SPARCLAB facility of the Frascati INFN laboratories. We have simulated the laser acceleration of protons, the transport and energy selection with collimators and a pulsed solenoid and the post-acceleration with a compact high field linac. For the highest achievable intensity corresponding to a= 30 over 10 protons at 30 MeV with a 3% spread are selected, and at least10 protons are post-accelerated up to 60 MeV. If a 10 Hz repetition rated can be achieved the delivered dose would be suitable for the treatment ...
In the paper we will present the experimental set-up and the first tests of diagnostic devices ba... more In the paper we will present the experimental set-up and the first tests of diagnostic devices based on radiochromic films, a Thomson parabola and solid-state diodes arrays. A scheme for the focusing and the transport of an emitted proton beam based on a pulsed solenoid feed by a custom designed power supply will be also presented.
The acceleration of dense targets driven by the radiation pressure of high-intensity lasers leads... more The acceleration of dense targets driven by the radiation pressure of high-intensity lasers leads to a Rayleigh-Taylor instability (RTI) with rippling of the interaction surface. Using a simple model it is shown that the self-consistent modulation of the radiation pressure caused by a sinusoidal rippling affects substantially the wavevector spectrum of the RTI depending on the laser polarization. The plasmonic enhancement of the local field when the rippling period is close to a laser wavelength sets the dominant RTI scale. The nonlinear evolution is investigated by three dimensional simulations, which show the formation of stable structures with "wallpaper" symmetry.
Electron plasma wakefield acceleration (PWFA) mechanism is a promising non conventional accelerat... more Electron plasma wakefield acceleration (PWFA) mechanism is a promising non conventional acceleration scheme. Nonetheless further investigation is still needed to fully uncover the instability mechanisms so to mitigate them and make PWFA an effective tool. This work focuses in this direction, we discuss the necessity to use well matched driver bunches to further mitigate witness instabilities. Specifically we propose to inject driver bunches with larger emittance than the matched one (overcompressed bunch) so to let the system reach the matching condition by itself. This preliminary results lead us to the following consideration: while a limited number of cases can be studied with a particle-in-cell code, we understand the necessity for fast systematic analysis: we briefly introduce the hybrid
ArXiv, 2015
We present a detailed strong and weak scaling analysis of PICCANTE, an open source, massively par... more We present a detailed strong and weak scaling analysis of PICCANTE, an open source, massively parallel, fully-relativistic Particle-In-Cell (PIC) code. PIC codes are widely used in plasma physics and astrophysics to study the cases where kinetic effects are relevant. PICCANTE is primarily developed to study laser-plasma interaction. Within a PRACE Preparatory Access Project, various revisions of different routines of the code have been analysed on the HPC systems JUQUEEN at Juelich Supercomputing Centre (JSC), Germany, and FERMI at CINECA, Italy, to improve scalability and I/O performance of the application. The diagnostic tool Scalasca is used to identify suboptimal routines. Different output strategies are discussed. The detailed strong and weak scaling behaviour of the improved code are presented in comparison with the original version of the code.
Uploads
Papers by Stefano Sinigardi