The paper introduces Tessera, a distributed runtime that enables equivariant message‑passing Transformers to simulate molecular dynamics at ab initio accuracy for tens of millions of atoms. By sorting atoms along a space‑filling curve and maintaining consistent workload shapes, Tessera achieves near‑linear scaling on 512 GPUs, sustaining 30.7 million atoms with high weak‑scaling efficiency and low energy drift in large‑scale simulations.
arXiv:2606. 03199v1 Announce Type: new Abstract: Organic crystal structure prediction (CSP) is a requirement for computational modelling of organic solids, but traditionally costs several CPU-years per molecule.
arXiv:2608. 13057v1 Announce Type: cross Abstract: In expert-parallel (EP) MoE serving, every layer synchronizes at the slowest GPU.
ASPIRE (Atomistic Saddle-Point Inference with Refinement for Events) is a new framework that predicts multiple saddle-point candidates from a single atomic environment and refines them using Dimer searches on the original interatomic potential. Its equivariant set predictor, Ev-Quiformer, generates proposals via geometry-conditioned scalar‑vector event slots and decodes them into full atomic displacement fields. The authors provide two new datasets—BCCFE4VACAV‑4000 and BCCFE‑1TO4VAC—and demonstrate that ASPIRE achieves 77.20 % reference‑event coverage, outperforming a conventional Dimer baseline while halving the number of force evaluations and reducing wall‑time per configuration from 478.8 s to 176.3 s.
arXiv:2610.02395v1 Announce Type: new Abstract: Streaming GPU solvers for entropic optimal transport (EOT), such as FlashSinkhorn, avoid storing the dense kernel but still evaluate all $n\times m$ po...
arXiv:2604.09107v2 Announce Type: replace-cross Abstract: Modern LLM reinforcement learning (RL) workloads require a high-performance weight transfer system to scale training across heterogeneous com...
The paper examines the commonly used structure‑disjoint split for evaluating molecular machine learning models, which relies on single‑linkage clustering with the Maximum Common Edge Subgraph (MCES) distance at a threshold of 10. The authors show that this criterion, originally calibrated on 29,000 molecules, has never been tested on the current public libraries of 227,690 molecules due to its NP‑hard complexity. They propose an optimized algorithm that reduces the number of required integer programs from $2.6 imes10^{10}$ to $1.15 imes10^{7}$, enabling lossless partitioning at scale, and reveal that the resulting graph percolates—over 95 % of the library lies in a single component, making held‑out compounds systematically larger and more complex than training ones. "whyItMatters":"The study highlights that large‑library benchmarks using the traditional split conflate structural novelty with extrapolation to larger, more complex chemistry, potentially misleading assessments of model generalization."
arXiv:2606. 01502v1 Announce Type: cross Abstract: Frontier LLMs increasingly decide what a query attends to with a sparse-attention indexer that picks a few KV-cache blocks per query: attention's unit is now a small, reusable chunk.
The paper introduces juniper, a discrete denoising diffusion model that performs compositional backmapping from coarse‑grained (CG) beads to detailed molecular graphs. By conditioning on the octanol–water partition free energy, the model generates valid and unique molecules that closely match the target free‑energy distribution for two‑bead systems. This approach enables the conversion of CG screening results into candidate atomistic molecules for further study or synthesis.
arXiv:2512. 20968v2 Announce Type: replace-cross Abstract: Distributed attention is essential for scaling large language models (LLMs) to long contexts, yet existing methods either have limited parallelism or incur high communication costs.
NeMo-DCR is a delta‑compressed refit method that sends only the changes in a trillion‑parameter policy model, achieving bit‑exact synchronization across AWS regions. By using fixed affine mappings, XOR masks, and a relay‑tree object storage, it avoids full checkpoint transfers and cross‑cluster collectives, reducing refit times from 87.5 min to 150 s for a 1 T model at 3% change rate. The approach scales to 30 B–1 T models, delivering 12–40× speedups over traditional full‑checkpoint refits.
uFlowCSP is a MeanFlow-based crystal structure prediction model that learns the average probability‑flow velocity, enabling it to generate complete crystal structures in one to five network evaluations. It achieves inference speeds 5×–58× faster than diffusion and flow‑matching methods while matching or surpassing their performance, with a chemistry‑ and symmetry‑aware Transformer that uses canonical atom ordering and per‑token chemistry embeddings. On the MP‑20 benchmark, uFlowCSP attains comparable or higher accuracy with far fewer evaluations and significantly lower wall‑clock time, demonstrating improved accuracy per network evaluation.
arXiv:2607. 25504v1 Announce Type: cross Abstract: Fine-grained weight pruning and activation sparsification have emerged as effective approaches for reducing the compute and memory cost of inference for Transformer models.