Prismatic

Prismatic simulates atomic-resolution image formation in scanning transmission electron microscopy (STEM) to accelerate quantitative modeling of electron scattering for microscopy-based structural analysis.

Key Features:

• Plane-Wave Reciprocal-Space Interpolated Scattering Matrix (PRISM) Method: Implements the PRISM method (Plane-Wave Reciprocal-Space Interpolated Scattering Matrix) for rapid STEM image formation simulation, enabling up to 1000× speedups relative to traditional multislice methods.
• Multislice Method: Supports the multislice method with reported speed improvements of up to 15× on a single machine equipped with four GPUs.
• Parallel Computing Capabilities: Distributes computation across multiple CUDA-enabled GPUs and multicore processors for parallel execution.

Scientific Applications:

• Atomic Electron Tomography: Computes STEM images at speeds suitable for incorporation into atomic electron tomography reconstruction pipelines.
• Materials Science and Nanotechnology Research: Reduces computational burden to enable larger-scale simulation studies and more extensive experimentation in materials science and nanotechnology.

Methodology:

Implements PRISM and multislice algorithms and distributes computational tasks across multiple processing units, leveraging CUDA-enabled GPUs and multicore CPUs.