Multi-plane HyperX achieves significantly smaller network diameter and superior cost-effectiveness versus multi-plane Fat-Tree, Dragonfly, and Dragonfly+ for large AI/HPC systems.
Technology-driven, highly-scalable dragonfly topology
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
years
2026 3verdicts
UNVERDICTED 3representative citing papers
Piper introduces resource modeling and pipelined hybrid parallelism for MoE training, delivering 2-3.5X higher MFU than prior frameworks and 1.2-9X better all-to-all bandwidth.
Aurora reached 1.01 EF/s FP64 HPL and 11.64 EF/s HPL-MxP through locality-aware mapping, CPU-GPU pipelining, mixed-precision orchestration, and hybrid resilience on a large Intel GPU-based system.
citing papers explorer
-
Multi-Plane HyperX: A Low-Latency and Cost-Effective Network for Large-Scale AI and HPC Systems
Multi-plane HyperX achieves significantly smaller network diameter and superior cost-effectiveness versus multi-plane Fat-Tree, Dragonfly, and Dragonfly+ for large AI/HPC systems.
-
Piper: Efficient Large-Scale MoE Training via Resource Modeling and Pipelined Hybrid Parallelism
Piper introduces resource modeling and pipelined hybrid parallelism for MoE training, delivering 2-3.5X higher MFU than prior frameworks and 1.2-9X better all-to-all bandwidth.
-
Sustaining Exascale Performance: Lessons from HPL and HPL-MxP on Aurora
Aurora reached 1.01 EF/s FP64 HPL and 11.64 EF/s HPL-MxP through locality-aware mapping, CPU-GPU pipelining, mixed-precision orchestration, and hybrid resilience on a large Intel GPU-based system.