AERIS

Argonne Earth Systems Model for Reliable and Skillful Predictions

Presentation at the 2025 ALCF Hands On HPC Workshop
Author
Affiliation
Published

October 8, 2025

Modified

October 6, 2025


AERIS: Argonne Earth Systems Model for Reliable and Skillful Predictions

Figure 1

High-Level Overview of AERIS

Figure 2: Rollout of AERIS model, specific humidity at 700m.
Table 1: Overview of AERIS model and training setup
Property Description
Domain Global
Resolution 0.25° & 1.4°
Training Data ERA5 (1979–2018)
Model Architecture Swin Transformer
Speedup O(10k–100k)

Model Overview

  • Dataset: ECMWF Reanalysis v5 (ERA5)
  • Variables: Surface and pressure levels
  • Usage: Medium-range weather forecasting
  • Partition:
    • Train: 1979–2018
    • Val: 2019
    • Test: 2020
  • Data Size: 100GB at 5.6° to 31TB at 0.25°
Table 2: Variables used in AERIS training and prediction
Variable Description
t2m 2m Temperature
X u(v) uu (vv) wind component @ Xm
q Specific Humidity
z Geopotential
msl Mean Sea Level Pressure
sst Sea Surface Temperature
lsm Land-sea mask

Windowed Self-Attention

  • Benefits for weather modeling:
    • Shifted windows capture both local patterns and long-range context
    • Constant scale, windowed self-attention provides high-resolution forecasts
    • Designed (currently) for fixed, 2D grids
  • Inspiration from SOTA LLMs:
    • RMSNorm, SwiGLU, 2D RoPE
Figure 3: Windowed Self-Attention

Model Architecture: Details

Figure 4: Model Architecture

Sequence-Window-Pipeline Parallelism SWiPe

  • SWiPe is a novel parallelism strategy for Swin-based Transformers
  • Hybrid 3D Parallelism strategy, combining:
    • Sequence parallelism (SP)
    • Window parallelism (WP)
    • Pipeline parallelism (PP)
Figure 5
Figure 6: SWiPe Communication Patterns

🌌 Aurora

Table 3: Aurora Specs
Racks 166
Nodes 10,624
CPUs 21,248
GPUs 63,744
NICs 84,992
HBM 8 PB
DDR5c 10 PB
Figure 7: Aurora: Fact Sheet.

🌎 AERIS: Scaling Results

Figure 8: AERIS: Scaling Results
  • 10 EFLOPs (sustained) @ 120,960 GPUs
  • See (Hatanpää et al. ()) for additional details
  • arXiv:2509.13523

References

Hatanpää, Väinö, Eugene Ku, Jason Stock, Murali Emani, Sam Foreman, Chunyong Jung, Sandeep Madireddy, et al. 2025. “AERIS: Argonne Earth Systems Model for Reliable and Skillful Predictions.” https://arxiv.org/abs/2509.13523.

Footnotes

  1. Relative to PDE-based models, e.g.: GFS↩︎

  2. 🏆 Aurora Supercomputer Ranks Fastest for AI↩︎

Citation

BibTeX citation:
@unpublished{foreman2025,
  author = {Foreman, Sam},
  title = {AERIS},
  date = {2025-10-08},
  url = {https://samforeman.me/talks/2025/10/08/slides.html},
  langid = {en}
}
For attribution, please cite this work as:
Foreman, Sam. 2025. “AERIS.” October 8. https://samforeman.me/talks/2025/10/08/slides.html.