HetCCL: Accelerating LLM Training with Heterogeneous GPUs
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866915762283741184 |
|---|---|
| author | Kim, Heehoon Lee, Jaehwan Kim, Taejeoung Park, Jongwon Kim, Jinpyo Suh, Pyongwon Choi, Ryan H. Lee, Sangwoo Lee, Jaejin |
| author_facet | Kim, Heehoon Lee, Jaehwan Kim, Taejeoung Park, Jongwon Kim, Jinpyo Suh, Pyongwon Choi, Ryan H. Lee, Sangwoo Lee, Jaejin |
| contents | The rapid growth of large language models is driving organizations to expand their GPU clusters, often with GPUs from multiple vendors. However, current deep learning frameworks lack support for collective communication across heterogeneous GPUs, leading to inefficiency and higher costs. We present HetCCL, a collective communication library that unifies vendor-specific backends and enables RDMA-based communication across GPUs without requiring driver modifications. HetCCL introduces two novel mechanisms that enable cross-vendor communication while leveraging optimized vendor libraries, NVIDIA NCCL and AMD RCCL. Evaluations on a multi-vendor GPU cluster show that HetCCL matches NCCL and RCCL performance in homogeneous setups while uniquely scaling in heterogeneous environments, enabling practical, high-performance training with both NVIDIA and AMD GPUs without changes to existing deep learning applications. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_22585 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | HetCCL: Accelerating LLM Training with Heterogeneous GPUs Kim, Heehoon Lee, Jaehwan Kim, Taejeoung Park, Jongwon Kim, Jinpyo Suh, Pyongwon Choi, Ryan H. Lee, Sangwoo Lee, Jaejin Distributed, Parallel, and Cluster Computing Machine Learning The rapid growth of large language models is driving organizations to expand their GPU clusters, often with GPUs from multiple vendors. However, current deep learning frameworks lack support for collective communication across heterogeneous GPUs, leading to inefficiency and higher costs. We present HetCCL, a collective communication library that unifies vendor-specific backends and enables RDMA-based communication across GPUs without requiring driver modifications. HetCCL introduces two novel mechanisms that enable cross-vendor communication while leveraging optimized vendor libraries, NVIDIA NCCL and AMD RCCL. Evaluations on a multi-vendor GPU cluster show that HetCCL matches NCCL and RCCL performance in homogeneous setups while uniquely scaling in heterogeneous environments, enabling practical, high-performance training with both NVIDIA and AMD GPUs without changes to existing deep learning applications. |
| title | HetCCL: Accelerating LLM Training with Heterogeneous GPUs |
| topic | Distributed, Parallel, and Cluster Computing Machine Learning |
| url | https://arxiv.org/abs/2601.22585 |