• Home
• People
• Jobs
• Visit
• Thesis Topics
• Publications
• Conference Rankings
• Research
• Bcast
• Teaching
• Contact

SPONSORS

COLLABORATIONS

The documents distributed by this server have been provided by the contributing authors as a means to ensure timely dissemination of scholarly and technical work on a noncommercial basis. Copyright and all rights therein are maintained by the authors or by other copyright holders, notwithstanding that they have offered their works here electronically. It is understood that all persons copying this information will adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.

Publications of SPCL

M. Besta, L. Schmidt, L. Nonino, R. Gerstenberger, P. Pang, P. Okanovic, A. Kubicek, T. Chen, B. Lipshitz, T. Hoefler:
	Performance Foundations of Parallel & Distributed Reasoning Language Models (arXiv:2608.27046. Aug. 2026) Abstract Reinforcement Learning with Verifiable Rewards (RLVR) and other RL-style post-training paradigms have been used for aligning large language models (LLMs) with reasoning standards. The resulting recent Reasoning Language Models (RLMs) such as DeepSeek-R1, o3, and Kimi k1.5 show that such RL-style post-training ("RL-for-LLMs") can substantially improve chain-of-thought reasoning, long-horizon planning, and self-correction. However, the computational footprint of these systems is massive: state-of-the-art RLM training requires millions of GPU-hours and tightly coupled multi-model pipelines that stress modern hardware far beyond classical supervised LLM training. This makes RLM training as much a parallel and distributed systems problem as an algorithmic one. In this work, to facilitate developing RLMs that are simultaneously high-performance, scalable, and cost-effective, we first systematize the RL-for-LLM paradigm and provide a compute-centric analysis of prominent post-training algorithmic frameworks: Proximal Policy Optimization (PPO), Group Relative Policy Optimization (GRPO), as well as their variants. Second, we develop a taxonomy of intra- and inter-model parallelism strategies for RL-for-LLMs, covering both traditional techniques (data, tensor, pipeline, sequence, context, and expert parallelism) as well as novel forms of parallelism and optimization techniques for multi-model RLM training, for example disaggregated placement, stage fusion, hybrid parallelism, and asynchronous execution. We harness the work-depth model of parallel computing to make our taxonomy and its insights rigorous and portable. Finally, we analyze existing RLM frameworks and we distill practical guidelines and outline open research directions for building scalable, fast, and cost-effective RLMs. Documents download article: access preprint on arxiv:
	BibTeX
@article{besta2026performance,   author={Maciej Besta and Leonard Schmidt and Lara Nonino and Robert Gerstenberger and Pierre Pang and Patrik Okanovic and Ales Kubicek and Tiancheng Chen and Baraq Lipshitz and Torsten Hoefler},   title={{Performance Foundations of Parallel & Distributed Reasoning Language Models}},   journal={arXiv:2608.27046},   year={2026},   month={08},   doi={10.48550/arXiv.2608.27046}, }