Characterizing Compute-Communication Overlap in GPU-Accelerated Distributed Deep Learning: Performance and Power Implications

Documenti analoghi

• MoETuner: Optimized Mixture of Expert Serving with Balanced Expert Placement and Token Routing
  di: Go, Seokjin, et al.
  Pubblicazione: (2025)
• Characterizing the Efficiency of Distributed Training: A Power, Performance, and Thermal Perspective
  di: Go, Seokjin, et al.
  Pubblicazione: (2025)
• ViBE: Co-Optimizing Workload Skew and Hardware Variability for MoE Serving
  di: Go, Seokjin, et al.
  Pubblicazione: (2026)
• Lagom: Unleashing the Power of Communication and Computation Overlapping for Distributed LLM Training
  di: Xu, Guanbin, et al.
  Pubblicazione: (2026)
• Workload-Aware Hardware Accelerator Mining for Distributed Deep Learning Training
  di: Adnan, Muhammad, et al.
  Pubblicazione: (2024)