密度矩阵重规范化组方法的并行实施 在单个DGX-H100 GPU节点上实现四分之一的petaFLOPS性能
Andor Menczer1,2, Maarten van Damme3, Alan Rask3
1Strongly Correlated Systems Lendület Research Group, Wigner Research Centre for Physics, H-1525 Budapest, Hungary.
Journal of chemical theory and computation
|September 19, 2024
概括
我们在NVIDIA DGX-H100架构上使用旋转适应密度矩阵重规范化组 (DMRG) 方法的混合CPU-多GPU实现实现了尖端性能,用于复杂的分子模拟.
科学领域:
- 量子化学 是一个量子化学.
- 计算物理 计算物理
- 高性能计算 高性能计算
背景情况:
- 准确的电子结构计算对于理解酶机制至关重要.
- 密度矩阵重规范化组 (DMRG) 是量子化学的一个强大的方法.
- 在现代硬件上将DMRG扩展到大型系统仍然是一个挑战.
研究的目的:
- 在NVIDIA DGX-H100架构上报告旋转适应的DMRG实现的性能结果.
- 在混合CPU-多GPU系统上评估张量网络算法的效率.
- 评估使用先进硬件解决具有挑战性的量子化学问题的可行性.
主要方法:
- 一个单节点混合型CPU-多GPU实现了自旋适应的DMRG方法.
- 在NVIDIA DGX-H100架构上的性能评估.
- 对FeMoco和P450酶的活性位点进行的计算.
主要成果:
- 实现了 246 teraFLOPS 的持续性能.
- 与DGX-A100架构相比,表现出2.5倍的性能改进.
- 与128核心CPU实现相比,展示了80倍的加速.
结论:
- 张量网络算法可以有效地利用高性能多GPU硬件.
- 张量网络和GPU加速器的结合使得解决复杂的量子化学问题成为可能.
- 这项工作为计算化学和相关领域的进步铺平了道路.
相关概念视频
Parallel Processing
397
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...
397
Fast Decoupled and DC Powerflow
362
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:
362
Fermi Level Dynamics
418
The vacuum level denotes the energy threshold required for an electron to escape from a material surface. It is usually positioned above the conduction band of a semiconductor and acts as a benchmark for comparing electron energies within various materials.
Electron affinity in semiconductors refers to the energy gap between the minimum of its conduction band and the vacuum level and it is a critical parameter in determining how easily a semiconductor can accept additional electrons.
The work...
Electron affinity in semiconductors refers to the energy gap between the minimum of its conduction band and the vacuum level and it is a critical parameter in determining how easily a semiconductor can accept additional electrons.
The work...
418
Maxwell-Boltzmann Distribution: Problem Solving
1.9K
Individual molecules in a gas move in random directions, but a gas containing numerous molecules has a predictable distribution of molecular speeds, which is known as the Maxwell-Boltzmann distribution, f(v).
This distribution function f(v) is defined by saying that the expected number N (v1,v2) of particles with speeds between v1 and v2 is given by
This distribution function f(v) is defined by saying that the expected number N (v1,v2) of particles with speeds between v1 and v2 is given by
1.9K


