Trace-based performance analysis provides essential insights for understanding and optimizing large-scale parallel applications. However, traces from applications running on tens of thousands of processes can easily exceed terabytes, far surpassing the memory capacity of typical computing nodes. Existing approaches either require expensive distributed analysis or cannot work on platforms with limited memory. To address these challenges, we present StreamTrace, a streaming trace analysis system that enables efficient analysis of massive traces on a single computing node with bounded memory consumption. StreamTrace introduces two key techniques: communication pattern-aware chunk partitioning that minimizes cross-chunk dependence, and dynamic priority-based chunk scheduling that reduces analysis waiting time by prioritizing frequently depended upon processes. Our evaluation on traces from applications with up to 8,192 processes demonstrates that StreamTrace achieves up to 3.48 × speedup on a single node over existing distributed systems. The ablation study shows the proposed technique can improve the performance over naive streaming implementations by up to 8.50 ×.
Yuyang Jin, Ji-Dong Zhai· Proceedings of the Internati...· 0 citations
A serving system that treats the diffusion block as a compilation unit that improves end-to-end execution time by up to 2.7 times over the strongest surviving baseline under the same 8-GPU placement and remains feasible at the largest batch sizes where multiple baselines run out of memory, while preserving task quality relative to the dense reference.
Jia-Nian Zhu, Hang Wu, Ying-Hui Li et al.· 0 citations
UniEP fuses the MoE communication and computation into MegaKernels, effectively transforming complex architectural tuning into a unified parameter search space for automated adaptability and incorporates a deterministic token ordering mechanism that guarantees numerical consistency with sequential execution, even under aggressive overlap schedules.
Size Zheng, Xuegui Zheng, Li-Wen Chang et al.· IEEE International Symposium...· 1 citation
Prism abstracts the highly dynamic diffusion workload into a predictable, static execution flow transparent to the compiler, and achieves this via three techniques: spatial regularization, temporal stabilization, and specialized kernels that selectively bypass padding data.
Jianian Zhu, Hang Wu, Yinghui Li et al.· International Conference on...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
