All Stories

  1. Accelerating MoE Model Inference with Expert Sharding
  2. vPIM: Processing-in-Memory Virtualization
  3. PathFS
  4. Stream Processing with Adaptive Edge-Enhanced Confidential Computing
  5. A new data loader for PyTorch that carefully overlaps the data preprocessing and the training.
  6. TONE
  7. SILK+ Preventing Latency Spikes in Log-Structured Merge Key-Value Stores Running Heterogeneous Workloads