面向分子动力学模拟的异构节点静电力计算装置
By employing a heterogeneous node architecture of CPU, FPGA, and GPU, task allocation and communication are optimized, solving the resource constraints and communication challenges of electrostatic force calculation in molecular dynamics simulations, and achieving efficient electrostatic force calculation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ZHEJIANG UNIV
- Filing Date
- 2026-04-24
- Publication Date
- 2026-07-17
AI Technical Summary
Existing high-performance computing platforms suffer from resource limitations and communication challenges in molecular dynamics simulations, especially in electrostatic force calculations. This prevents them from fully utilizing the parallel processing capabilities of GPUs and FPGAs, resulting in low computational efficiency.
It adopts a heterogeneous node architecture of CPU, FPGA and GPU, connected through PCIe bus. The CPU processes input files and builds topology, the FPGA performs short-range electrostatic force parallel calculation, and the GPU performs long-range electrostatic force calculation. The parallel execution optimizes task allocation and reduces communication overhead.
It improves the overall speed of electrostatic force calculation, optimizes task allocation, reduces energy consumption, and enhances computing efficiency and hardware resource utilization.
Smart Images
Figure CN122086636B_ABST