Metraipolis › archive › dc_7e052102a116
MemX: Memory-Aware Automatic Parallelism for Cross-Vendor Heterogeneous GPU Training — arXiv:2610.41525
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.41525
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-02T04:46:36.236Z
- SHA-256 of the archived text
c98d129a2f3f68d77401d3c195dd6163e9d926bfc21488d78588420d8b734071- Stable URL of this page
https://metraipolis.com/documento/dc_7e052102a116
How to cite
MemX: Memory-Aware Automatic Parallelism for Cross-Vendor Heterogeneous GPU Training — arXiv:2610.41525 — arXiv (en), https://arxiv.org/abs/2610.41525 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7e052102a116 (Metraipolis archive dc_7e052102a116, sha256 c98d129a2f3f68d77401d3c195dd6163e9d926bfc21488d78588420d8b734071)
BibTeX
@misc{dc7e052102a116,
title = {MemX: Memory-Aware Automatic Parallelism for Cross-Vendor Heterogeneous GPU Training — arXiv:2610.41525},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.41525},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7e052102a116; sha256 c98d129a2f3f68d77401d3c195dd6163e9d926bfc21488d78588420d8b734071},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41525. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Cross-vendor heterogeneous GPU clusters are increasingly used for large-scale model training. However, their GPUs differ substantially in compute power, memory capacity, and interconnect characteristics. Existing automatic parallelism training systems usually minimize iteration time while treating memory as a hard constraint, almost failing in resolving the compute--memory asymmetry of cross-vendor clusters regardless correlation between memory and throughput. In this paper, we present MemX, a new memory-aware automatic parallelism system that jointly optimizes iteration time, peak memory usage, and cross-GPU memory balance. MemX consists of a weighted-product objective function, heterogeneous recomputation strategies, a minimal-set profiler, interpolation-based data augmentation, a dual-output XGBoost predictor, and a hash-signature prediction cache, in order to accelerate search of large-scale auto-parallel best configurations at lower cost. We implement MemX on Megatron-LM and evaluate it against four automatic parallelism baselines on two heterogeneous clusters. MemX achieves the lowest prediction Mean Absolute Percentage Error (MAPE) of 12.17% and is the only method that identifies the real Top-1 configuration under both time-based and memory-balance rankings. In terms of search cost and search space, MemX searches 31,240 configurations in 26.5 seconds, while an end-to-end profiling baseline is extrapolated to require about 17 days at the same scale. The simulation study that extends the measured 20-GPU data to 40-GPU and 60-GPU clusters further illustrates the effectiveness of MemX at larger scales. Tema degli esploratori: calcolo e infrastrutture (elenco arXiv cs.DC, nuovi lavori del giorno). Categorie arXiv: Distributed, Parallel, and Cluster Computing (cs.DC). Pagina del lavoro: https://arxiv.org/abs/2610.41525 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-02T04:42:41.703Z dalla pagina https://arxiv.org/list/cs.DC/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).