2606.11169v1 Jun 09, 2026 cs.DC

Piper: A Programmable Distributed Training System

Yinxu Pan
Yinxu Pan
Citations: 74
h-index: 3
Megan Frisella
Megan Frisella
Citations: 14
h-index: 2
Shubham Tiwari
Shubham Tiwari
Citations: 1
h-index: 1
Andy Ruan
Andy Ruan
Citations: 0
h-index: 0
Parker Gustafson
Parker Gustafson
Citations: 0
h-index: 0
Mathews Jacob
Mathews Jacob
Citations: 0
h-index: 0
G. Bernstein
G. Bernstein
Citations: 86
h-index: 6
Stephanie Wang
Stephanie Wang
Citations: 7
h-index: 1

Large-scale model training increasingly relies on composing multiple parallelism strategies, such as data, pipeline, and expert parallelism, together with memory-saving optimizations like ZeRO. Deployed systems for foundation model pretraining often rely on human experts to manually design a high-level parallelism strategy then implement the corresponding low-level execution strategy, making it difficult to adapt the system to new strategies. Meanwhile, many general-purpose frameworks are more flexible but their implementations are still tied to a fixed set of common parallelism strategies, making it challenging to integrate state-of-the-art strategies. We present Piper, a user-controllable distributed training system that decouples the strategy from the runtime implementation. Piper allows users to declare a comprehensive distributed training strategy with a small set of model annotations and scheduling directives. Each directive applies a transformation on Piper's intermediate representation (IR), a unified global training DAG that represents all computation and communication. Using this IR, Piper compiles per-device execution plans and executes them with a distributed runtime agnostic to the strategy. We show that the combined system maintains performance parity on commonly available strategies such as ZeRO, while also enabling additional performance and memory efficiency gains through joint scheduling of compute and communication in composed parallelism strategies such as DeepSeek-V3's DualPipe.

0 Citations
0 Influential
3 Altmetric
15.0 Score
Original PDF

No Analysis Report Yet

This paper hasn't been analyzed by Gemini yet.

Log in to request an AI analysis.

댓글

댓글을 작성하려면 로그인하세요.

아직 댓글이 없습니다. 첫 번째 댓글을 남겨보세요!