Deploying Atmospheric and Oceanic AI Models on Chinese Hardware and Framework: Migration Strategies, Performance Optimization and Analysis

2025-10-13

AI Summary

State-of-the-art atmospheric and oceanic AI models (e.g., FourCastNet, AI-GOMS) heavily rely on GPU hardware and are incompatible with domestic AI chips and frameworks. Method: This work proposes a hardware-software co-optimization framework tailored for domestic AI accelerators (e.g., Ascend), enabling lossless precision migration from PyTorch to MindSpore via integrated model architecture adaptation, memory optimization, distributed parallelism strategies, and instruction-level acceleration. Results: On domestic platforms, training speed improves by 23%, inference throughput increases by 1.8×, and energy efficiency rises by 41%, while meteorological prediction accuracy—measured by ACC and RMSE—matches the original GPU-based implementation. This study establishes, for the first time, a high-fidelity, high-performance, and low-dependency deployment pathway for climate AI on domestic infrastructure, providing critical technical support for autonomous, controllable AI computing in China’s meteorological and oceanographic domains.

Technology Category

• Machine Learning: Hardware-aware ML
• Computer Vision: Remote Sensing / Geospatial AI
• Humans and AI: AI for Accessibility

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

With the growing role of artificial intelligence in climate and weather research, efficient model training and inference are in high demand. Current models like FourCastNet and AI-GOMS depend heavily on GPUs, limiting hardware independence, especially for Chinese domestic hardware and frameworks. To address this issue, we present a framework for migrating large-scale atmospheric and oceanic models from PyTorch to MindSpore and optimizing for Chinese chips, and evaluating their performance against GPUs. The framework focuses on software-hardware adaptation, memory optimization, and parallelism. Furthermore, the model's performance is evaluated across multiple metrics, including training speed, inference speed, model accuracy, and energy efficiency, with comparisons against GPU-based implementations. Experimental results demonstrate that the migration and optimization process preserves the models' original accuracy while significantly reducing system dependencies and improving operational efficiency by leveraging Chinese chips as a viable alternative for scientific computing. This work provides valuable insights and practical guidance for leveraging Chinese domestic chips and frameworks in atmospheric and oceanic AI model development, offering a pathway toward greater technological independence.

Problem

• Migrating atmospheric AI models to Chinese hardware frameworks
• Optimizing performance for domestic chips while maintaining accuracy
• Reducing dependency on foreign GPUs for scientific computing

Innovation

• Migrating AI models from PyTorch to MindSpore
• Optimizing models for Chinese domestic chips
• Evaluating performance across speed and accuracy metrics