A new AI-accelerated iterative solver, DAIST, is introduced for 3D-IC thermal modeling. It decomposes a full-package FEM simulation into block-level subdomains, replaces each subdomain solver with a neural operator, and couples them through iterative exchanges of interfacial temperature and heat flux. This approach eliminates the need to retrain models when die count or topology changes and offers a controllable accuracy‑runtime tradeoff, achieving up to 178× speedup with minimal temperature error on tested systems.
arXiv:2608. 16080v1 Announce Type: new Abstract: Thermal-aware optimization of multi-die 3D integrated circuits evaluates many designs, each a costly heat-equation solve.
arXiv:2510. 15968v2 Announce Type: replace-cross Abstract: Thermal management in 3D ICs is increasingly challenging due to higher power densities.
arXiv:2602. 04940v2 Announce Type: replace Abstract: Deep learning has emerged as a transformative tool for the neural surrogate modeling of partial differential equations (PDEs), known as neural PDE solvers.
arXiv:2607. 08025v1 Announce Type: new Abstract: While neural PDE solvers have demonstrated significant potential for accelerating engineering simulations, existing architectures remain constrained by high memory consumption and the single node bottleneck, where the maximum processable mesh resolution is strictly limited by the VRAM of a single compute unit.
arXiv:2606. 07982v1 Announce Type: new Abstract: High-dimensional transient heat diffusion under noisy boundary conditions exposes a fundamental limitation of classical numerical methods: accuracy degrades catastrophically where physical noise is unavoidable.
While neural PDE solvers have demonstrated significant potential for accelerating engineering simulations, existing architectures remain constrained by high memory consumption and the single node bottleneck, where the maximum processable mesh resolution is strictly limited by the VRAM of a single compute unit. To address these challenges, we propose PGD-NO, a neural operator with Precomputed Geometry Decomposition, that relocates the computational overhead of geometric encoding to a deterministic pre-computation phase.
Cova-PINN is a multi-domain physics-informed neural network designed to model fluid–solid conjugate heat transfer in complex geometries. It aligns conservation support with thermal interaction paths, jointly optimizing cross-domain composite control‑volume balances at the local scale and paired‑wall closure at the global exchanger scale. In tests on four triply periodic minimal surface heat exchangers and a DualMS design, Cova‑PINN reduces average outlet‑temperature and device‑level closure errors by 37.7 % and 60.2 % compared to the closest baseline, while improving full‑field and heat‑duty accuracy.
EGGPlacer is an energy‑guided diffusion framework for macro placement that integrates analytical placement feedback and differentiable thermal feedback. By treating these signals as composable energies, the method injects them into reverse diffusion without retraining the generative model. Experiments on public benchmarks show that EGGPlacer achieves competitive mixed‑size HPWL and outperforms other methods in routed wirelength and peak temperature.
arXiv:2603. 29268v2 Announce Type: replace Abstract: High-density through-substrate vias (TSVs) enable 2.
arXiv:2510. 23472v2 Announce Type: replace-cross Abstract: Chip placement is a vital stage in modern chip design, and black-box optimization (BBO) has been applied to it for decades.
arXiv:2607. 18020v1 Announce Type: new Abstract: Physics-Informed Neural Networks (PINNs) solve PDEs by incorporating physical constraints into neural-network training, but large-scale problems are limited by automatic-differentiation memory overhead and inefficient execution of grid-based PDE operators.
arXiv:2607. 18020v2 Announce Type: replace Abstract: Physics-Informed Neural Networks (PINNs) solve PDEs by incorporating physical constraints into neural-network training, but large-scale problems are limited by automatic-differentiation memory overhead and inefficient execution of grid-based PDE operators.