Quantum high energy density physics

Abstract:

Sequential models, such as Recurrent Neural Networks and Neural Ordinary Differential Equations, have long suffered from slow training due to their inherent sequential nature. For many years this bottleneck has persisted, as many thought sequential models could not be parallelized. We challenge this long-held belief with our parallel algorithm that accelerates GPU evaluation of sequential models by up to 3 orders of magnitude faster without compromising output accuracy. The algorithm does not need any special structure in the sequential models' architecture, making it applicable to a wide range of architectures. Using our method, training sequential models can be more than 10 times faster than the common sequential method without any meaningful difference in the training results. Leveraging this accelerated training, we discovered the efficacy of the Gated Recurrent Unit in a long time series classification problem with 17k time samples. By overcoming the training bottleneck, our work serves as the first step to unlock the potential of non-linear sequential models for long sequence problems.

Abstract:

We present experimental observations of K_β emission from highly charged Mg ions at solid density, driven by intense x rays from a free electron laser. The presence of K_β emission indicates the n = 3 atomic shell is relocalized for high charge states, providing an upper constraint on the depression of the ionization potential. We explore the process of state relocalization in dense plasmas from first principles using finite-temperature density functional theory alongside a wave-function localization metric, and find excellent agreement with experimental results.

Abstract:

An extension to the wave packet description of quantum plasmas is presented, where the wave packet can be elongated in arbitrary directions. A generalised Ewald summation is constructed for the wave packet models accounting for long-range Coulomb interactions and fermionic effects are approximated by purpose-built Pauli potentials, self-consistent with the wave packets used. We demonstrate its numerical implementation with good parallel support and close to linear scaling in particle number, used for comparisons with the more common wave packet employing isotropic states. Ground state and thermal properties are compared between the models with differences occurring primarily in the electronic subsystem. Especially, the electrical conductivity of dense hydrogen is investigated where a 15% increase in DC conductivity can be seen in our wave packet model compared to other models.

Abstract:

We report on ion emission from plasma produced on thick targets irradiated with nanosecond and femtosecond pulses delivered by mid-ultraviolet and soft x-ray lasers, respectively. To distinguish between different ion acceleration mechanisms, the maximum kinetic energy of ions produced under different interaction conditions is plotted versus laser fluence. The transformation of the time-of-flight detector signal into ion charge density distance-of-flight spectra makes it possible to determine the mean kinetic energy of the fastest ion groups based on the influence of the acoustic velocity of ion expansion. This allows obtaining additional characteristics of the ion production. The final energy of the group of fast ions determined using the ion sound velocity model is an order of magnitude larger in the fs-XFEL interaction than in the ns-UV one. On the contrary, the ablation yield of ions in our experiment is seven orders of magnitude greater when applying ns-UV laser pulses, not only due to higher energies of UV laser pulses, but also due to a significant difference in interaction and ion formation mechanisms.