Related Experiment Videos
Efficient Coupled-Cluster Python Frameworks for Next-Generation GPUs: A Comparative Study of CuPy and PyTorch on the
Antonina Dobrowolska1, Julian Świerczyński1, Paweł Tecmer1
1Institute of Physics, Faculty of Physics, Astronomy, and Informatics, Nicolaus Copernicus University in Toruń, Grudziadzka 5, Toruń 87-100, Poland.
Journal of Chemical Theory and Computation
|July 2, 2026
Summary
New batching algorithms accelerate coupled-cluster singles and doubles (CCSD) calculations on GPUs. These Python-based methods significantly speed up computations, particularly for large molecular systems, by optimizing tensor contractions.
Area of Science:
- Computational Chemistry
- Quantum Chemistry
- High-Performance Computing
Background:
- Coupled-cluster singles and doubles (CCSD) is a high-level quantum chemistry method.
- Efficient implementation of CCSD is crucial for large-scale molecular simulations.
- Leveraging graphical processing units (GPUs) offers significant computational acceleration.
Purpose of the Study:
- To develop and benchmark new batching algorithms for CCSD implementations on GPUs.
- To optimize tensor contractions for improved performance in Python-based quantum chemistry.
- To evaluate the efficiency of CuPy and PyTorch libraries on NVIDIA architectures.
Main Methods:
- Developed asymmetric and dynamic splitting for particle-particle ladder bottleneck contraction.
- Implemented a generic tensor contraction protocol for GPU-exclusive operations.
- Benchmarked performance on NVIDIA Hopper (H100) and Grace Hopper (GH200) architectures using CuPy and PyTorch.
- Utilized Cholesky-decomposed electron repulsion integrals for molecular CCSD calculations.
Main Results:
- Achieved a 10-fold speedup compared to previous GPU implementations.
- Reported additional speedups of 3-16 for single CCSD iterations.
- PyTorch showed a ~20% performance advantage over CuPy on the H100 architecture.
- Both libraries performed comparably on the GH200 architecture.
Conclusions:
- The new batching algorithms and GPU-accelerated protocol significantly enhance CCSD computational efficiency.
- The study demonstrates the effectiveness of Python libraries like PyTorch and CuPy for high-performance quantum chemistry on modern GPUs.
- These advancements enable faster and more scalable electronic structure calculations for complex molecular systems.
Related Concept Videos
01:20Parallel Processing
The brain processes sensory information rapidly due to parallel processing, which involves sending data across multiple neural pathways at the same time. This method allows the brain to manage various sensory qualities, such as shapes, colors, movements, and locations, all concurrently. For instance, when observing a forest landscape, the brain simultaneously processes the movement of leaves, the shapes of trees, the depth between them, and the various shades of green. This enables a quick and...
01:17Ampere-Maxwell's Law: Problem-Solving
A parallel-plate capacitor with capacitance C, whose plates have area A and separation distance d, is connected to a resistor R and a battery of voltage V. The current starts to flow at t = 0. What is the displacement current between the capacitor plates at time t? From the properties of the capacitor, what is the corresponding real current?
To solve the problem, we can use the equations from the analysis of an RC circuit and Maxwell's version of Ampère's law.
For the first part of the problem,...
To solve the problem, we can use the equations from the analysis of an RC circuit and Maxwell's version of Ampère's law.
For the first part of the problem,...
01:16Efficiency of The Carnot Cycle
The hypothetical Carnot cycle consists of an ideal gas subjected to two isothermal and two adiabatic processes. Since the internal energy of an ideal gas depends only on its temperature, which is the same before and after the completion of the Carnot cycle, there is no change in its internal energy. Hence, using the first law of thermodynamics, the total heat exchanged by the ideal gas equals the total work done. Thus, we can quantify the efficiency of the Carnot cycle via the heat exchanged...
01:24Fast Decoupled and DC Powerflow
The fast decoupled power flow method addresses contingencies in power system operations, such as generator outages or transmission line failures. This method provides quick power flow solutions, essential for real-time system adjustments. Fast decoupled power flow algorithms simplify the Jacobian matrix by neglecting certain elements, leading to two sets of decoupled equations:
01:19Distributed Loads
Distributed loads are a common type of load that engineers and scientists encounter in various practical situations. Distributed loads often refer to a type of load spread over a surface or a structure and can be modeled as continuous force per unit area.
For example, consider a bookshelf filled with books stacked vertically adjacent to each other. The weight of the books is evenly distributed over the length of the shelf. As a result, the pressure at different locations on the surface of the...
For example, consider a bookshelf filled with books stacked vertically adjacent to each other. The weight of the books is evenly distributed over the length of the shelf. As a result, the pressure at different locations on the surface of the...
01:17Accelerating Fluids
When a fluid is in constant acceleration, the pressure and buoyant force equations are modified. Suppose a beaker is placed in an elevator accelerating upward with a constant acceleration, a. In the beaker, assume there is a thin cylinder of height h with an infinitesimal cross-sectional area, ΔS.
The motion of the liquid within this infinitesimal cylinder is considered to obtain the pressure difference. Three vertical forces act on this liquid:
The motion of the liquid within this infinitesimal cylinder is considered to obtain the pressure difference. Three vertical forces act on this liquid: