Salvato in:
| Autore principale: | |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2509.18114 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915506878939136 |
|---|---|
| author | Moye, Javed I. Khan an Henry Uwabor |
| author_facet | Moye, Javed I. Khan an Henry Uwabor |
| contents | Autoregressive inference in large transformer-based language models (LLMs) presents significant challenges for runtime efficiency, particularly during the decode phase where load imbalance across GPU shards can cause throughput degradation and latency spikes. A DPU-assisted framework leveraged by BlueField-3 Data Processing Units can enable real-time detection and mitigation of load imbalance in multi-node tensor-parallel inference. By offloading monitoring tasks to the DPU and analyzing GPU telemetry and inter-node communication patterns, the resulting system can provide actionable feedback to inference controllers and schedulers. The goal of this study is three-fold i) identify the reported skews/imbalances/pathological conditions that arise in muti-GPU execution of a) LLM tensor computing (both during training and inference), b) identify their impact on computational performance, and c) make a critical assessment if those can be tracked for potential mitigation from a DPU network. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_18114 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A Study of Skews, Imbalances, and Pathological Conditions in LLM Inference Deployment on GPU Clusters detectable from DPU Moye, Javed I. Khan an Henry Uwabor Machine Learning Autoregressive inference in large transformer-based language models (LLMs) presents significant challenges for runtime efficiency, particularly during the decode phase where load imbalance across GPU shards can cause throughput degradation and latency spikes. A DPU-assisted framework leveraged by BlueField-3 Data Processing Units can enable real-time detection and mitigation of load imbalance in multi-node tensor-parallel inference. By offloading monitoring tasks to the DPU and analyzing GPU telemetry and inter-node communication patterns, the resulting system can provide actionable feedback to inference controllers and schedulers. The goal of this study is three-fold i) identify the reported skews/imbalances/pathological conditions that arise in muti-GPU execution of a) LLM tensor computing (both during training and inference), b) identify their impact on computational performance, and c) make a critical assessment if those can be tracked for potential mitigation from a DPU network. |
| title | A Study of Skews, Imbalances, and Pathological Conditions in LLM Inference Deployment on GPU Clusters detectable from DPU |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2509.18114 |