Papers for

hardware acceleration developers

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Rounding alignment improves stability in large language model training

Where Rounding Happens Matters: Rounding-Location Alignment Stabilizes LLM Reinforcement Learning

Abstract: Reinforcement learning for large language models often uses separate systems for inference and training. Even with identical weights and BF16 tensors, implementation differences can make an intended on-policy update off-policy and destabilize training. Existing remedies align the entire inference and training stacks, change global precision, or modify the learning algorithm. We ask whether aligning selected parts of the execution stack can stabilize BF16 RL while other implementation differences remain. We align training-side rounding locations with inference in three regions---RoPE, SwiGLU, and residual RMSNorm---by retaining selected intermediates in FP32 until their BF16 outputs. Across four Qwen-family models, aligning all three regions prevents the collapse observed under the default BF16 baseline; RoPE-path alignment alone also suffices in the studied runs, despite remaining numerical discrepancies. The interventions keep the inference engine, matrix-multiplication and attention kernels, and learning objective unchanged.

Fri 23 OctMachine Learning
The gist
Training large language models with reinforcement learning can become unstable because small differences in how numbers are rounded during calculations make the updates behave unexpectedly. The authors found that carefully matching where rounding happens during training and inference helps keep the training stable, especially by focusing on three specific parts of the model’s calculations. This approach avoids changing big parts of the system or the learning rules and still stops the model training from collapsing.
Open → 2610.32951v1

TrainBridge speeds up training on NPUs using GPU optimization clues

TrainBridge: Agentic NPU Training Optimization Guided by GPU Compilation Evidence

Abstract: We study how to improve training efficiency on Ascend NPUs. In the NPU training programs we study, framework-level compilation optimizations such as operator fusion are only partially applied, leaving fragmented computation and avoidable execution overheads. Profiling-guided tuning can address these inefficiencies, but it often requires costly trial and error. To make training on NPUs more efficient, we propose TrainBridge, an agentic approach for transferring GPU optimization guidance to NPU training. TrainBridge uses the highly optimized execution produced by mature GPU compilation stacks as a task-specific reference for NPU tuning. It extracts optimization intents from the GPU training execution to guide NPU program rewrites and further refines the rewritten program through iterative profiling. Our GPU-to-NPU migration experiments show that TrainBridge achieves an average 1.7x speedup across four workloads covering LLM pretraining, image classification, and vision-language training, and an average speedup of nearly 2.0x across 27 Modded-NanoGPT speedrun records with highly optimized GPU implementations.

Wed 21 OctComputation and LanguageArtificial Intelligence
The gist
Training machine learning models on special chips called NPUs often runs slower because some software optimizations are missing. The authors found a way to make NPU training faster by learning from how GPUs, a different kind of chip, run the same tasks very efficiently. Their method, TrainBridge, copies the GPU’s smart tricks for optimizing programs and adjusts them for NPUs, boosting speed by about 1.7 to 2 times.
Open → 2610.29171v1