Haiyue Ma

According to our database, Haiyue Ma authored at least 5 papers between 2024 and 2026.

Collaborative distances:

• Dijkstra number of four.
• Erdős number of four.

2026

Optimizing Dropout in LLM Training: Performance Comparison of Fusion and Overlap.

Haiyue Ma, Jian Liu, Ronny Krashinsky

Proceedings of the Abstracts of the 2026 ACM SIGMETRICS International Conference on Measurement and Modeling of Computer Systems, 2026

2025

Efficient MoE Serving in the Memory-Bound Regime: Balance Activated Experts, Not Tokens.

Yanpeng Yu, Haiyue Ma, Krish Agarwal, Nicolai Oswald, Qijing Huang, Hugo Linsenmaier, Chunhui Mei, Ritchie Zhao, Ritika Borkar, Bita Darvish Rouhani, David Nellans, Ronny Krashinsky, Anurag Khandelwal

CoRR, December, 2025

MoE-GPS: Guidlines for Prediction Strategy for Dynamic Expert Duplication in MoE Load Balancing.

Haiyue Ma, Zhixu Du, Yiran Chen

CoRR, June, 2025

Poster:Value-Aware Scheduler for Energy Reduction.

Haiyue Ma, Kaifeng Xu, David Wentzlaff

Proceedings of the 34th International Conference on Parallel Architectures and Compilation Techniques, 2025

2024

Reducing the Cost of Dropout in Flash-Attention by Hiding RNG with GEMM.

Haiyue Ma, Jian Liu, Ronny Krashinsky

CoRR, 2024