We perform a series of calculations using simulated QPUs, accelerated by NVIDIA CUDA-Q platform, focusing on a molecular analog of an amine-functionalized metal-organic framework (MOF) — a promising class of material...
详细信息
This paper proposes a two-dimensional (2D) bidirectional long short-term memory generative adversarial network (GAN) to produce synthetic standard 12-lead ECGs corresponding to four types of signals—left ventricular ...
详细信息
To achieve long-pulse steady operation, the physical mechanisms of boundary turbulence need further investigation. We employ the two-fluid model with flute reduction on BOUT++ to simulate the boundary plasma in Tokama...
To achieve long-pulse steady operation, the physical mechanisms of boundary turbulence need further investigation. We employ the two-fluid model with flute reduction on BOUT++ to simulate the boundary plasma in Tokamaks. The space and time scales of turbulence reproduced by our simulations closely relate to the spatial mesh size and time step size, respectively. As an inherent time scale, the Alfven time is sufficient to resolve MHD instabilities. The spatial scale can be refined by increasing mesh resolutions, which necessitates larger scale parallel computing resources. We have conducted nonlinear simulations using more than 33 million spatial meshes with 16,384 CPU processors in parallel. The results indicate that while the decrease in parallel efficiency with an increase in core numbers does not necessarily lead to shorter runtimes, higher computational complexity improves parallel efficiency for the same number of cores. In addition, the mesh resolution required for convergence conditions differs between linear and nonlinear simulations, with nonlinear simulations demanding higher resolution. Besides finer structure obtained, the fluctuation characteristic of density similar to WCM, which is more consistent with the experimental observation, also shows the requirement for high-resolution meshes and large-scale computing in the future.
暂无评论